GB2142730A

GB2142730A - Measurement of chamber volume

Info

Publication number: GB2142730A
Application number: GB08418790A
Authority: GB
Inventors: Harry W Fisher; Eugene B Perrine; William W Ho; William F Hall
Original assignee: Rockwell International Corp
Current assignee: Boeing North American Inc
Priority date: 1980-03-28
Filing date: 1984-07-24
Publication date: 1985-01-23
Also published as: FR2479459B1; AU548537B2; NL8101534A; GB2072854A; CA1157681A; DE3112227A1; GB2072854B; AU6767681A; GB2142730B; GB8418790D0; FR2479459A1

Abstract

The invention relates to a method and apparatus for measuring the volume of a chamber, for example a variable volume chamber 12 of regular geometry defined by a movable element, particularly the chamber of a meter prover. Electromagnetic waves are generated in the chamber, the frequency is varied to establish a resonant wave standing condition in the chamber; the mode of generation being selected so that at said resonant condition, the electric and magnetic components of the field generated are related to the dimensions of the chamber. By determining the capacity of the chamber at first and second positions of the piston or other movable element a displacement volume is determined with great accuracy and this is compared with the high resolution signal to determine the incremental volume of fluid displaced by incremental measured movement of said element, e.g. relative to a single pulse of the train. <IMAGE>

Description

12 GB 2 142 730A 12 translator and an inverter circuit in the form of a

line driver while the data input line is processed to invert the signal before application to the SDC circuit 1 92a. The clock circuit 193 applies a signal via a logic circuit 195 comprised of logic translators to the clock inputs of the SDC circuit 1 92a.

The signal conditioning circuit 138 as generally shown in Fig. 4E is more completely shown 5 in the schematic drawing of Fig. 4M. The PI/O circuit 200 is coupled to the CPU 120 by the I/D data bus 140 and those direct connections at the top and bottom so indicated. The outputs as variously taken from the PI/O circuit 200 are coupled by logic circuit 1 98a to variously energize the lights as shown on the system control and display unit 111. The logic circuit 1 98a is comprised of a logic translator for each output of the PI/O circuit 200 and a plurality of drivers to energize the corresponding lights. In addition, signals from the switch mechanisms of reset and stop pushlites 111 f and 111 g are applied via the logic circuit 1 98a and in particular to a set of NAND gates as shown in Fig. 4M whose outputs are applied via inverters to the PI/O circuit 200. A second PI/O circuit 196 is coupled by the I/D data bus 140 and direct connections to the CPU 120. A set of its outputs are applied via a logic circuit 1 94a to control various functions including direction and velocity of the servomotor 20; the logic circuit 1 94a receives seven inputs that are coupled via a set of NOR gates and series connected inverters and logic translators to a corresponding plurality of drivers, whose outputs serve to control the direction and velocity of the servomotor 20. Further, two outputs of the PI/O circuit 196 are applied via a logic circuit 1 94b comprised of a series connected inverter and a logic translator to 20 a driver before being applied to control the pneumatic solenoids 66 and 64 associated with the valves 34 and 36. A set of five inputs are derived from the logic circuit 1 94c which processes inputs from the proximity detectors 50, 52, and 54; the logic circuit 1 94c comprises a circuit of resistors and diodes as coupled via inverters to corresponding inputs of the PI/O circuit 196.

The logic circuit 194d is coupled to the logic circuit 1 70a shown in Fig. 4C and comprises a 25 series of NOR gates, the input signal comprising a one-second clock signal to control the sampling of the various pressure and temperature signals.

Referring now to Fig. 5, there is shown a high level flow diagram of the various steps of the program as stored within the PROM 124 and executed under the control of the CPU 120 using data as entered into the RAM 126. Initially, the power is applied to the computer system in step 30 210. Typically, the power supply for the computer system as shown in Fig. 3 can take the form of that supply as manufactured by Power Mate under their designations Power Mate EMA 18/24B and EMA 12/51); Analog Devices-925; Datel-MPS 5/12, MPS 5/3, and MPD 12/ 3; and Practical Automation- PS6 28. Thereafter, step 214 zeroes or erases the storage locations within RAM 126, before entering an initializing subroutine 300, whereby the various 35 portions of the computer system are initialized as will be explained later in detail with respect to Figs. 6A and 6B. It is noted that at various points during the course of the program, a return is made via entrance point 212 to step 214 to restart the operation of the program. As shown in Fig. 413, there is a switch 139 to be set to indicate whether it is desired to calibrate parts of the prover system 10 or to run a meter test. If the switch 139 is disposed to its calibrate mode, the 40 decision step 216 moves to step 400 wherein a subroutine is executed to calibrate the various analog inputs such as derived from the temperature and pressure measuring devices as shown in Figs. 1 and 2A, and the corresponding A/D converters to which these signals are applied, as will be explained in more detail with respect to Fig. 7. As the system moves to step 500, the operator can recall data from the various input measuring devices such as temperature measuring devices 42, 44, 48, and 57; the pressure measuring devices 51 and 46; and the output of the linear encoder 26. In addition, the operator may also initiate various of the meter self-tests. This subroutine will be described in more detail with respect to Figs. BA to P. After gathering the appropriate data', the program moves to step or routine 900 wherein a test or a series of tests of a meter 38 is carried out by the meter prover system 10 and the results thereof 50 displayed or recorded upon the CRT or printer. The routine 900 will be explained in more detail with respect to Figs. 9A to Q.

Referring now to Figs. 6A and 613, there is shown the initializing routine 300 wherein in step 302, a command is sent via the I/D bus 140 to cause the logic circuit 198a, as shown in Fig. 4E, to energize the standby lite 11 le. In step 304, a scaling factor corresponding to a one cubic foot test is transferred from the RAM 126 to the programmable counter 1 74a within the interval timer 174 as shown in 4C, to appropriately scale the output of the meter encoder counter 1 79a whereby upon counting an appropriate number of pulses, e.g., 40,000, the logic circuit 177 outputs a pulse indicative that one cubic foot of fluid has been drawn through the meter 38. Next, in step 306, the interrupts associated with the CPU 120 are enabled to permit 60 at any time later in the program the interrupts to be executed if the operator, for example, depresses the reset pushbutton 111 f or the stop pushbutton 111 g. Up to this point in the initialization subroutine 300, the interrupts associated with the pushbuttons 111 f and 111 g could not be enabled, but after execution of step 306, these interrupts are available to be serviced. Next in step 308, the circuitry shown in Fig. 4C as associated with the rotary meter 65 11 GB 2 142 7 30A 11 shown in the functional block diagram of Fig. 4C. The PI/0 circuit 184 is shown as being coupled by the I/D bus 140 to the CPU 120 and being coupled to the counter 182 comprised of a pair of counters 182a, 182b. in turn, the signal conditioner and logic circuit 1 70b is shown as comprised of a series of NOR gates 169 and a flip flop 169e whose output is applied via a NAND gate 180 and an inverter to an input of the counter 182a. The output of the linear 5 encoder 26 is applied via corresponding logic translators and inverters to the aforementioned NOR gates 169 of the signal conditioner and logic circuit 1 70b. The rotary meter encoder 40 is applied to the signal conditioner and logic circuit 1 70a that is similar to the signal conditioner and logic circuit 1 70b to provide a composite signal to the interval timer 174 and a conditioned signal corresponding to the A signal to the interval timer 176. It is understood that the [/C circuits designated 174 and 176 of Fig. 4K, also include the programmable counters 174a and 1 76a, respectively. The output of the interval timer 174 is applied via the logic circuit 177 comprised of a NAND gate and a pair of NOR gates as shown in Fig. 4K, to the Interrupt 2 input of CPU 120. The output of interval timer 176 is applied to the control board of the servomotor 20.

As shown in Fig. 4K, the signal conditioner and logic circuit 1 70a is comprised of first and second inputs receiving respectively the output signals A and A as developed by the rotary meter encoder 40. As shown respectively in Figs. 11 A and 11 P, the A 1 signal lags the A signal to provide, as will be explained, an indication of the direction of rotation in which the meter encoder 40 is moving. It is understood that the meter encoder 40 is designed for this particular 20 system to rotate in a clockwise direction, and if jitter or mechanical vibration is imposed upon the meter encoder 40, at least momentarily, the A signal may appear to be leading the A signal; Fig. 11 C shows the output signal A as it would appear as if it leads the A signal by W, this condition being undesired, indicating an erroneous signal condition. The signal conditioner and logic circuit 170a is designed to eliminate such conditions as will now be explained. The A and A signal are each applied through a level shifter and inverter circuits to NOR gates 181a and 181 b, respectively. The output of the NOR gate 181 a is coupled to an input of the NOR gate 181 b and the output of the NOR gate 181 b is coupled via an inverter to an input of a NOR gate 181 c. As shown in Fig. 4K, an inverted signal, i.e., 180 out of phase with an input to the NOR gate 181 a, is supplied to a NOR gate 181 d, whose output is applied to the other input of 30 the NOR gate 181 c. The effective output of the signal conditioner and logic circuit 1 70a is derived from the output of the NOR gate 181 c as shown in Fig. 11 D, assuming that the meter encoder 40 is rotated in a clockwise or desired direction, and is applied to the interval timer 174 to be counted as explained above. However, if even on a relatively short time basis, the signal appears to be leading the A signal, a DC (or logic zero signal) output signal will be derived from the NOR gate 181 c indicating the presence of jitter or some other erroneous signal. In similar fashion, the signal conditioner and logic circuit 1 70b receives the A and signal as derived from the linear encoder 26, these signals being also illustrated by Figs. 11 A and 11 B respectively. In similar fashion, the A and A signals are applied to a similar set of NOR gates 1 69a, b, c, and d. The output of NOR gate 1 69c is applied to a latch as comprised of a 40 pair of NOR gates interconnected as shown in Fig. 4K. In similar fashion, if the A signal as derived from the linear encoder 26 is lagging its A signal, the output as shown in Fig. 11 C will be applied via the latch 1 69e to the interval timer 176 to be counted by its counter 1 76a. A further set of NOR gates is also included in the signal conditioner and logic circuit 1 70b to provide an output signal as applied to the input 22 of the interval timer 176 to indicate the occurrence of the A signal leading the A signal as derived from the linear encoder 26, indicating that the piston 14 is being driven in a reverse condition, i.e., is being driven by the servomotor in a downward direction toward its park position.

In Fig. 4L, there is shown a schematic diagram of the signal conditioner and logic circuit 136 as generally shown in Fig. 4D. Data is transferred between the parallel data controller (PDC) 50 192b and the CPU 120 via the I/D bus 140 and direct connections to the CPU 120. The output of the PDC circuit 1 92b is coupled to a logic circuit 1 90b and by a plurality of lines as shown on the right-hand side of the PDC 1 92b. The logic circuit 1 90b is primarily comprised of a logic translator connected to each of its outputs. A NAND gate is incorporated into the logic circuit 1 90b to provide a reset signal to the printer 111 a. As shown in Fig. 4L, a strobe signal is 55 applied to the printer 111 a which in turn applies an acknowledge signal (ACKO) to the PDC circuit 19 2b, whereby data may be transmitted to be printed by the printer 111 a under the control of a set of signals so marked. In addition, a busy signal may be developed by the printer 111 a to inhibit the transmission of data from the PDC circuit 1 92b to the printer 111 a. Further, the SDC circuit 192a is coupled via the I/D bus 140 and direct connections to the CPU 120; 60 its output as taken from the right-hand side of the SDC circuit 1 92a is applied via the logic circuit 1 90a to provide data into and from the CRT terminal 112 under the control of preselected signals as provided by the circuitry as shown on the left- hand side and designated control; briefly, the control signals provide fixed signals to determine the mode of operation of the CRT terminal 112. The logic circuit 1 90a coupled to the data out signal comprises a logic 65 GB 2 142 730A 10 removed and reapplied. Upon reapplication of power, the meter prover 10 will return to the standby mode. During the course of the execution of the program, appropriate signals are generated and applied via the I/D bus 140, the PI/O circuit 200, and the logic and driver circuit 1 98a to energize the appropriate indicator lights 111 b to 111 e.

The signal conditioning circuits as shown in the functional block diagrams of Figs. 4A to 4E 5 are shown in more detail in the schematic diagrams of Figs. 4G to M. The signal conditioning circuit 130 as generally shown in Fig. 4A is more specifically shown in the schematic diagrams of Figs. 4G, 4H, and 41.

In Fig. 4G, there is shown the multiplexer 149 as comprising a plurality of relays having mercury wetted relay contacts, which reduce resistance presented thereby and are coupled to 10 the channels connected to the amplifier modules for providing to its relay of the multiplexer 149, a voltage corresponding to the temperature as measured by one of the prover temperature transducers 57 or 48, the meter temperature transducers 44 or 42, or one of the room temperature transducers RO to R3. A selected channel is applied by the energized relay of the multiplexer 149 via the second multiplexer 154, as shown in Fig. 4G, and the amplifier 152 to 15 the A/D converter 158, as shown in Fig. 4H. In Fig. 4H, there is shown the PI/O circuit 160 as being coupled by the I/D bus 140 to the CPU 120. In addition, an output is derived from the PI/O circuit 160 to be applied to a HEX to one decoder 153, as shown in Fig. 4G, which in turn energizes one of a plurality of drivers 155 to close the corresponding relay of the 20 multiplexer 149; the PI/O circuit output is also applied via the logic circuit 156 as comprised of 20 an AND gate 1 56a, an inverter 1 56b, and a logic translator 1 56c to the multiplexer 154. Further, with respect to Fig. 4H, when the output of a selected one of the temperature measuring modules is applied to the A/D converter 158, a convert signal is applied to the A/D converter 158 from the PI/O circuit 160 via the expander circuit 153. In response to this convert signal, the A/D converter 158 converts the inputted analog temperature signal to a corresponding digital signal to be transmitted via the PI/O circuit 160 to the I/D bus 140, and transmits an end of conversion status signal via the conductor 147 to the PI /0 160.

Referring now to Fig. 41, there is shown a schematic diagram of an amplifier module to which each of the temperature measuring devices may be applied and amplified to provide a voltage output signal to be applied via a corresponding channel to the multiplexer 149. Illustratively, 30 the temperature measuring devices may comprise a resistance temperature device as manufac tured by Senso-Metrics under their designation No. 601222. The resistance temperature device (RTD) is coupled as one arm of a resistance bridge 201 comprised of the RTD, and resistors R1, R2, and R3. The excitation voltage, as applied to the a and b terminals of the bridge 201, as well as the output voltage as derived from the terminals c and d thereof are coupled to a 35 conditioning circuit 203 as illustratively made by Analog Devices under their designation Model 21331.. Basically, the conditioning circuit 203 includes an operational amplifier 205 to which is applied the output of the bridge 201 to be amplified before being applied to a Bessel filter 207 whereby selected frequencies may be removed before being further amplified by an operational amplifier 209 to be applied to a corresponding channel of the multiplexer 149. An extremely 40 stable voltage aupply serves to energize the circuit 203 and may illustratively comprise a voltage supply as manufactured by Analog Devices under their designation AD584. As indicated in Fig.

41, the gain of the operational amplifier 205 is controlled by the resistance R4 placed between terminals 10 and 11 of the circuit 203, while the output offset is adjusted by setting the potentiometer R5 couplep!o the terminal 29 of the circuit 203. In addition, the voltage and 45 current as applied to energize the bridge 201 are respectively controlled by adjusting the potentiometers R7 and R6.

The signal conditioning circuit 164 as generally shown in Fig. 4B is shown in more detail by the schematic diagram of Fig. 4J. The PI/O circuit 168 is shown as being coupled to the I/D bus 140 and to the CPU 120 to provide transmission of data therebetween. A further input is 50 made from the A/D converter 166 via corresponding set of inverters 167 to the PI/O circuit 168. The pressure transducers 51, 46, and PB are connected respectively via the amplifiers 16 1 a, b, and c to the multiplexer 163. As indicated in Fig. 4J, the multiplexer 163 is made up of a corresponding plurality of relays which are energized to apply a selected output as derived from one of the pressure transducers to the A/D converter 166. The PI/O circuit 168 determines which of the relays of the multiplexer 163 is to be energized by applying control signals via the bus 169 to the logic circuit 165 comprised of a binary coded decimal to decimal converter and decoder 1 65a, which in effect selects which of the relays of the multiplexer 163 to be energized and applies a high going signal via a corresponding output via a set 1 65b of logic translators to a corresponding set 1 65c of power drivers, whereby a corresponding relay of 60 the multiplexer 163 is energized to apply the corresponding temperature output to the A/D converter 166. Next, the PI/O 168 applies a convert signal via conductor 159 to the A/D converter 166, which converts the inputted analog signal to a corresponding digital signal, and transmits a status signal to the PI/O circuit 168.

In Fig. 4K, there is shown a schematic diagram of the conditioning circuit 134 as generally 65 9 GB 2142 730A 9 the I/D bus 140 via a first logic circuit 190b, and a parallel data controller (PDC) 192b, which may illustratively take the form of the PDC as manufactured by the assignee of this invention under their designation number 10453 and provides a two-way controlled access between the I/D bus 140 and the printer 111 a. Thus, on command through the PDC 192b, a signal is developed by the logic circuit 1 90b whereby the printer 111 a is strobed and an appropriate acknowledging signal (ACKO) is transferred via the logic circuit 190b and the PDC 192b to indicate that the printer 111 a is available for printing. If the printer 111 a is busy, an appropriate busy signal will be transmitted back to the I/D bus 140. If a command has been issued to print data, the control portion of the signal is transmitted via the PDC 1 92b and the logic circuit 1 90b to control the printer 111 a to print that data which appears upon the data 10 channel derived from the logic circuit 190b.

Further, the operator may enter appropriate commands upon the keyboard of the CRT terminal 112 that is interconnected via the logic circuit I 90a and a serial data controller (SDC) 192a to the computer system via the I/D bus 140. The SDC 1 92a may illustratively take the form of that SDC as manufactured by the assignee of this invention under their model No. 15 10930. The SDC 192a is capable of receiving the serially oriented data as derived from the CRT terminal 112 including instructions entered by the operator upon the terminal's keyboard.

The SDC 192a converts these serial signals inputted from the logic circuit 1 90a at an appropriate baud rate set by the clock 193,- and transmits a set of digital signals via the I/D bus 140 to the CPU 120. In turn, data to be displayed upon the CRT terminal 112 is transmitted by 20 the I/D bus 140 via the SDC 192a and by the logic circuit 190a to be displayed upon the terminal's CRT.

The signal conditioning and interface circuit 138, as shown in Fig. 4E, interfaces between the I/D bus 140 and the first or inlet valve 34 and the second or exhaust valve 36, as well as to provide signals to and from the motor control. Motor control signals in terms of speed and direction are applied via the I/D bus 140 to be received and transmitted via the PI/O circuit 196 to a logic circuit 1 94a whereby these digital signals are applied to the motor control to effect a corresponding action of the servomotor 20. Similarly, at the appropriate time under the control of the executed program, signals are developed to close or to open the valves 34 and 36 by actuating their corresponding solenoids 66 and 64, respectively; these valve control signals 30 are applied via the PI/O circuit 196 and the logic circuit 1 94b to a pair of pneumatic valves disposed within the control room 105 as seen in Fig. 2C, whereby a 50 psi supply of air is selectively applied to each of the valve solenoids 66 and 64, respectively, to open and close these valves upon command. In this manner, the heat generated by the valve solenoids is removed from the temperature controlled prover room 106. In addition, each of the solenoids 35 66 and 64 includes a proximity detector to determine whether the valve is open or closed. The output signals developed by the proximity detectors 50, 52, and 54 to determine the approximate position of the piston 14 are applied via a logic circuit 1 94c and the PI /0 196 to the I/D bus 140. A one second sampling clock is developed from the encoder board (as explained above) and is applied via the logic circuit 194d to the PI/O circuit 196. Similarly, the 40 operator may actuate a switch to determine whether the meter fluid flow is to be obtained from the proximity detector 27 or from the encoder 26, as shown in Fig. 4C; this command signal is applied via the PI/O circuit 196 to the I/D bus 140. In order to achieve proper control over the servomotor 20, the motor status in terms of its speed, direction, and measured torque is applied via the logic circuit 198b and the PI/O circuit 200 to the I/D bus 140.

The display unit 111 includes a front panel as shown in Fig. 2D and is provided to display a series of lights and backlighted pushbuttons variously indicating the condition of the system. As shown in Fig. 2D, the display panel includes the printer 111 a for providing printouts of the flow rate and the percentage of the fluid through the meter 38. In addition, there is included a plurality of lights 111 b to 111 e. The light 111 b is energized to indicate that a self-test of the 50 meter prover 10 is being run, as will be explained. The standby light 111 e indicates that power has been applied to the meter prover 10 and that an initialization process has been started to place the prover 10 in its standby mode. While in this standby mode, a series of keyboard responses are required of the operator and upon completion of entry of the data via the keyboard of the CRT terminal 112, the meter prover 10 will automatically go into the test in progress mode as indicated by the energization of the test in progress light 111 c; in this mode, the meter 38 is actually being tested. Upon completion of a meter test, the test completed indicator light 111 d is energized. At this time, the final percent accuracy is calculated and is printed out on the hard copy printer 111 a. In addition, there is included a backlit stop pushbutton 111 g and a back-lit reset pushbutton 111 f. If during any phase of the operation, 60 the restart pUshbutton 111 f is depressed, the meter prover 10 will respond as if power is initially applied, as will be more fully explained with respect to Fig. 9H. The stop pushbutton 111 g is depressed only when an emergency situation occurs that may cause damage to the prover 10. Upon depressing the stop pushutton 111 g, the servomotor 20 is quickly decelerated to a halt and the prover 10 is locked up in its stop mode until primary power is 65 8 GB 2 142 730A 8 characteristics of the rotary meter encoder 40 in terms of the number of pulses it generates per revolution. In an illustrative embodiment of this invention, a count of 40,000 is placed in the programmable counter 1 74a corresponding to a volume of one cubic foot to be drawn through the meter 40. Assuming that the characteristics of the meter 38 and the encoder 40 remain the same for varying volumes, illustrative counts of 20,000 and 10, 000 may be stored in the programmable counter 1 74a if it is desired, respectively, to draw one-half and one-fourth cubic foot of fluid through the meter 38. By entering a count based upon the characteristics of the meter 38 into the the programmable counter 1 74a which is counted down by pulses derived from the rotary meter encoder 40, a more accurate test of the meter is assured in that the beginning and ending of meter test will be effected at the same point in the cycle of the rotation 10 of the meter and its rotary encoder 40, as explained above.

As indicated in Fig. 4C, a clock A is derived from the system clock 122 via the CPU 120 and the I/D bus 140 is applied to the interval timer 176. A selected factor is placed into down counter 1 76a to provide an output from the interval timer 176 corresponding to a sampling pulse of one pulse per second. In an illustrative embodiment of this invention, the system clock as derived from the clock 122 is in the order of 200 KHz, and the factor placed in the counter 1 76a is such to provide the desired one pulse per second to the logic circuit 178 and therefrom to the motor control board. As will be explained later, this sampling pulse is used to time the sampling of the measurements of pressure and temperature.

Further, the output of the linear encoder 26 is a pair of signals A and A, 90 out of phase with each other, which are applied to a signal and conditioner circuit 1 70b. The circuit 1 70b is similar to circuit 1 70a in that it processes the inputted signal A and A to shape and condition these input signals eliminating jitter that might otherwise indicate a false output from the linear encoder 26. In addition, the circuit 1 70b is able to detect the direction in which the piston 14 is moving from the inputted signals A and A and if the outputted signals A and A do not indicate that the piston 14 is moving in the desired direction, no signals are outputted from the signal and conditioner circuit 1 70b. The circuit 1 70b provides a train of conditioned pulses corresponding to the linear encoder output to the counter 182, which after initiation counts and accumulates the output of linear encoder 26. The accumulated output of the counter 182 is applied to PI/O circuit 184 and upon command is transferred via the I/D bus 140 to the remaining portions of the computer system.

In one embodiment of this invention, a rotary encoder transducer 40 is coupled to the meter 38 and in particular includes an optical encoder rotatively coupled to the domestic meter tangent arm of the meter 38 to detect the rotation of the tangent arm as gas flows therethrough to provide a plurality of output signals A and A as explained above. In an alternative embodiment of this invention, the proximity detector 27 may be used to detect the mechanical rotation of the domestic meter tangent arm of the meter 38 by a mechanism that will be explained later to provide an output signal to a signal conditioner and logic circuit 1 70c, which is in turn connected to the interval timer 176 and to the interval timer 174. Due to the arrangement of the mechanical mechanism coupled to the tangent arm of the meter 38, the proximity detector 27 produces a signal of lesser resolution than that produced by the optical encoder 26, as it detects the rotation of the meter tangent arm; the particular advantage of the proximity detector arrangement is that of the relative simplicity of its mechanical and electrical structure. The choice of whether to use the proximity detector 27 or the rotary encoder transducer 40 is made by the operator by throwing a switch 19 1, as shown in Fig. 4K. When 45 the operator determines to use the proximity detector 27, the programmable counter 1 74a is encoded with numbers of 8, 4, and 2 corresponding to measured volume flows of one cubic foot, one-half cubic foot, and one-fourth cubic foot. The operator initiates the entering of the appropriate factors whether for the proximity detector 27 or for the rotary encoder 40, by first throwing the switch 191 to the appropriate position and entering the test volume via the 50 keyboard of the CRT terminal 112. In Fig. 4F, there is shown a perspective view of a typical meter 38, which

measures the fluid flow by the use of two diaphragms, only one of which is shown as 1202; the flow meter as shown in Fig. 4F is more fully explained in U. S. Patent 2,544,665 dated March 31, 1951. As shown, a flag rod 1203 senses the flexing of the diaphragm 1202 to cause the arm 1204 to 55 oscillate. A second diaphragm (not shown) and an associated flag rod (not shown) cause arm 1206 to oscillate in an alternate cycle. As explained in the noted patent, the combination of the arms 1206 and 1204, and the arms 1208 and 1210 cause the tangent arm 1214 to rotate as described in said patent. A metallic target 1212 at the point of intersection of the arms 1208 and 1210, is rotated past the proximity detector 27, whereby an output is provided to the logic 60 circuit 1 70c to be processed as explained above.

The conditioning and interface circuit 136 is more fully shown with respect to 4D, wherein communication is made between the printer 111 a, the display unit 111 for printing out desired parameters as measured by the meter prover system 10 including the measured flow rate(s) and the percentage(s) of error for the tests performed. In particular, the printer 111 a is coupled to 65 7 GB 2 142 7 30A 7 same to a digital signal to be transmitted to the RAM 126 as explained above.

The interface and conditioning circuit 134 is more fully explained with respect to Fig. 4C. The rotary meter encoder 40 is coupled to the meter 38 to provide first and second signals A and A, 90 out of phase with each other, to the signal conditioning.9nd logic circuit 1 70a. In particular, the circuit 1 70a processes the input signals A to A to eliminate possible problems due to jitter of the signals as may be imposed by mechanical vibration upon the rotary meter encoder 40. The signal conditioner and logic circuit 1 70a generates a composite pulse signal corresponding to each set of input pulses of the signals A and A, and applies same to an interval timer 174 including a programmable counter 174a into which is loaded a factor dependent upon the selected volume of fluid to be drawn through the meter 38, in a manner 10 that will be explained. In particular that factor is placed in the programmable counter 1 74a and upon counting down to zero from that factor, a pulse is generated by the interval timer 174 and applied to a logic circuit 177 whose output is applied to initiate a CPU Interrupt 2 subroutine, whereby the testing of the meter 38 is terminated, as will be explained later in detail with respect to Fig. 9J. The logic circuit 1 70a is responsive to the input signal A from the rotary meter encoder 40 to apply a corresponding, conditioned pulse to an interval timer 176, which performs the operation of recognising the leading edge of the signal A to initiate timing or counting of the programmable counter 174a as well as of the counter 176a and of the linear encoder counter 182. More specifically, the CPU 120 controls the meter prover 10 and senses that the piston 14 has been accelerated from its park position to its start-test position, as indicated by the presence of an output signal from the proximity detector 52. Upon the sensing of the output of proximity detector 52, the program as executed by the CPU 120 periodically, e.g., approximately every 40 microseconds, accesses the interval timer 176 to see whether it has received an input signal from the signal conditioner and logic circuit 1 70a indicative of the leading edge of the input signal A of the rotary encoder transducer 40. Upon the detection of 25 the first leading edge of the output of the logic circuit 1 70a, after the piston 14 has passed the proximity detector 52, an initiate signal is applied to the PI/O circuit 184, which applies an initiate count signal to the counter 182 and also en initiate signal via the I/D bus 140 to the programmable counter 174a. In this manner, each of the counters 174a, 176a, and 182 are activated to start counting at the same time. In this illustrative example, the programmable counter 1 74a counts down in response to output signals of the rotary encoder transducer 40.

The initiating and terminating of the meter test, i.e., the counting by the programmable counter 1 74a and the counter 182, are made responsive to the output of the rotary encoder 40 of the meter 38 whose accuracy is to be measured. More specifically, the rotary encoder 40 is coupled to the fluid or gas meter 38 as will be more fully described with respect to Fig. 4F and 35 upon rotation of its rotatively mounted rod, the rotary meter encoder 40 as coupled thereto will produce a train of pulses corresponding to the rotation this tangent arm and the cycling of the meter's diaphragm. As shown in Figs. 4F, the rotary member is connected by coupling arms to the meter diaphragm and its rotation is not linear so that the output of the encoder 40 is in a sense frequency modulated. Therefore, in order to obtain an accurate measurement of the rotary 40 meter encoder 40, it is desired to count the pulses as derived from the encoder 40 so that the counting begins and terminates at approximately the same point in the rotation of the meter's rotary member. This is accomplished by initiating the counting in response to the rotary meter encoder 40. In particular, a meter test is conducted by accelerating the piston 14 from its park position to a steady state velocity so that upon its passing the proximity detector 52, disposed at 45 the start-test position, an output is provided therefrom to enable, as will be explained later, the detection of the leading edge of the next output signal from the logic circuit 1 70a corresponding to the leading edge of the next output signal A of the rotary meter encoder 40. The interval timer 176 responds to the leading edge to effect the simultaneous initiation of the counting of the programmable counter 174a and the counter 182. Upon the occurrence of the programma- 50 ble counter 1 74a being counted down from its selected factor dependent upon the desired volume to be drawn through the meter 38, the interval timiar 174 provides its output to the logic circuit 177 to enable Interrupt 2 of the CPU 120, which in turn terminates the counting of the counters 1 74a and 182 and transfers the respective counts to corresponding locations within the RAM 126. It is noted that the termination of counting could be implemented by software, but would involve an additional number of steps thus unduly complicating the programming of this system as well as adding to the time required to carry out the timing operation as described above. In addition, the effecting of the initiating of counting in response to the output of the rotary meter encoder 40 insures a more accurate test and calibration of the meter 38 under test.

As explained above, the output train of pulses as derived from the rotary encoder 40 is applied to count down the count initially placed in the programmable counter 1 74a. Significantly, the count as placed in the programmable counter 1 74a is variable dependent upon the volume desired to be drawn through the meter 38 and into the chamber 28. The count is based upon the structural dimensions of and characteristics of the meter 38, as well as the 6 GB 2 142 730A 6 indicating the status of the servomotor 20 as well as an input signal from a test mode switch to indicate whether the rotary encoder 40 or proximity detector 27 are to be used to measure meter flow, is entered via the logic circuit 198 and the interface circuit 200 to the I/D bus 140.

The signal conditioning and interface circuit 130 is shown in more detail in Fig. 4A, as including two line inputs derived from each of the temperature transducers 57, 48, 42 and 44 indicative respectively of the prover temperatures TP1 and TP2, and the meter temperatures TM3 and TM4, and is connected to a multiplexer 149. In addition, the four temperature transducers RO, R1, R2, and R3 disposed about the prover rom 106 in which the meter prover system 10 is housed, are applied through the next four inputs to the multiplexer 149. The aforementioned temperature transducers are connected into amplifier modules that serve to develop voltage outputs proportional to the temperature sensed and to apply these outputs to the corresponding inputs of the multiplexer 149. In this manner, each temperature transducer is associated with its own amplifier module so that its output to the multiplexer 149 may be adjusted to insure a substantially uniform output in terms of voltage amplitude and off-set for each of the temperature transducers connected to the multiplexer 149. The details of the amplifier modules for each of the transducers shown in Fig. 4A will be explained below with respect to Fig. 41 along with a detailed description of the signal conditioning and logic circuit as generally shown in Fig. 4A. The multiplexer 149 serves to time multiplex the inputs at each of its eight inputs and to scale the temperature signal outputs to be applied one at a time via a multiplexer 154 to an amplifier 152 taking the illustrative form of that amplifier as 20 manufactured by Analog Devices under their designation AD522. The second multiplexer 154 is normally set to apply one of the eight temperature input signals via the operational amplifier 152 to the analog digital (A/D) converter circuit 158, which may illustratively take the form of the A/D converter as manufactured by Burr Brown under their designation ADC 80. In a calibrate mode, the multiplexer 154 is actuated to apply a precision, calibrating voltage to the A/D converter 158. As is well known in the art, the DC voltage of the analog signals is adjusted to a level which may be readily accepted by the A/D converter 158, which in turn converts these analog signals to digital outputs which are applied via its 12 output lines to a parallel input-output (PI/0) device 160 which may illustratively take the form of that device manufac tured by the assignee of this invention under their designation P/N 11696. The PI/O circuit 30 permits input commands to be transferred via the I/D bus 140 to the multiplexer 149 to control which of the inputs is to be sampled at a particular time, as well as to the logic circuit 156 to enable the multiplexer 154 to apply one of the outputs of the multiplexer 149 or the voltage calibration input signal to the A/D converter circuit 158. In operation, the CPU 120 places a call signal via the I/D bus 140 to the PI/O circuit 160 which responds thereto by enabling a call for information to be read out and converted to digital data to be applied to the I/D bus 140. In addition, a command is derived from the PI/O circuit 160 to time the conversion of the analog signals to digital signals by the A/D converter circuit 158, and a signal indicative of the status of the A/D converter circuit is applied via the PI/O circuit 160 to the I/D bus 140. The voltage calibration signal permits the zero and span of the operational amplifier 152 to be adjusted so that the full amplitude of each input signal may appear at the A/D converter circuit 158.

The conditioning and interface circuit 132 is more fully shown with respect to Fig. 413 wherein there is shown that the outputs of the transducer 51 and 46 indicating respectively the prover or piston pressure PP1 and the meter pressure MP2 are applied via operational amplifiers 45 1 61a and 1 61b to a multiplexer 163. In addition, the output from the pressure transducer 109 for measuring the barometric or ambient pressure of the prover rom 106 is applied via operational amplifier 161 c to the multiplexer 16 3. 1 nitially, the CPU 120 transmits a command to the multiplexer 163 to select which of the outputs of the pressure transducers 51, 46 or 109 is to be read out via the I/D bus 140 and the parallel input-output (PI/0) circuit 168. In response thereto, the PI/O circuit 168 applies control signals via the four-line bus 169 to the multiplexer 163, to select one of the three pressure indicating signals or a signal indicative of the voltage calibration input signal to be applied to an A/D converter 166, which converts the input analog signal to a corresponding digital signal to be applied to the PI /0 circuit 168 to be in turn transmitted via the I/D bus 140. Next, upon command of the CPU 120, the PI/O circuit 55 168 commands via line 159 the A/D circuit 166 to convert the selected analog pressure output signal to a corresponding digital signal to be transmitted via the I/D bus 140. The conversion of the input analog data to digital data requires a discrete time period for the A to D conversion to take place and in addition, for the digital data appearing upon the 12 output data lines of the A/D circuit 166 to stabilize before they are read by the PI/O circuit 168. When signal stabilization has occurred on the 12 data lines output from the A/D circuit 166, a status signal is generated by the A/D circuit 166. In response to the status signal, the PI/O circuit 168 reads the data appearing on the output lines of the A/D circuit 166 and applies these signals via the I/D bus 140 to the RAM 126 as shown in Fig. 3. After this process is completed, the 65 system is able to select another pressure output as derived from another transducer, converting 65 GB 2 142 730A 5 the display unit 111 to display the various states of operation of the meter prover system 10, while permitting operator input through the keyboard of the CRT terminal 112, of selected meter test functions. The computer system includes a central processing unit (CPU) 120 of the type manufactured by the assignee of this invention under their designation PPS-8 Microcom- puter, from which address signals are applied from CPU 120 via an address bus 128 to a programmable read only memory (PROM) 124 and to a random access memory (RAM) 126. The RAM 126 may take the form of the 256 X 8 RAM as manufactured by the assignee of this invention, whereas the PROM 124 may take the form of that PROM manufactured by Intel Corporation under their designation 2708. A system clock 122 provides system clock signals (e.g., 200 KHz) to the CPU 120 and may illustratively take the form of the clock generator circuit P/N 10706 as manufactured by the assignee of this invention. As shown in Fig. 3, each of the clock 122, the CPU 120, the PROM 124, and the RAM 126 are interconnected by an instruction-data (I/D) bus 140, which may illustratively take the form of a 14-line bus not only interconnecting the forementioned elements but also connected to each of the signal condition ing and interface circuits 130, 132, 134, 136, and 138.

The circuit 130 conditions and interface the signals indicative of the prover temperature signals TPI and TP2 as derived respectively from the temperature measuring devices 57 and 48. Further, the meter temperature signals TM3 and TM4 as derived from the devices 42 and 44 are also applied to the circuit 130. As shown in Fig. 2C, four additional room temperature measuring devices RO, R1, R2, and R3 are provided about the prover room 106 in which the 20 meter prover system 10 as shown in Figs. 1 and 2A is disposed; in this regard it is understood that the ambient conditions about the meter prover system 10 are well regulated in order to maintain stable as possible the ambient temperature of the meter prover system 10. It is normal practice to store the meters 38 to be tested in this environment for a time to permit them to reach the same ambient conditions at which the meter prover system 10 is disposed. As shown 25 in Fig. 3, the temperature signals are applied to a signal conditioner and logic circuit 1 50'and from there via an interface circuit 151 to the I/D bus 140.

In similar fashion, the pressure signals are applied to the conditioning and interface circuit 132, which comprise a signal conditioner and logic circuit 162 and an interface circuit 164. In particular the outputs of the differential pressure measuring device 51 indicating the differential 30 pressure PP 1 of the prover 10 and of the differential pressure measuring device 46 indicating the meter pressure MP2, and the barometric Pressure measuring device 109 indicating the ambient or atmospheric pressure P13 of the prover room 106 are applied to the circuit 132.

The output of the meter encoder 40 and the linear optical encoder 26 are applied to the signal conditioning and interface circuit 134. In particular, the output of the meter encoder 40 35 is applied to a signal conditioning and logic circuit 1 70a, whose outputs are applied in turn to interface circuits 171 and 173. A clock circuit 175 applies a signal to the interface circuit 173.

In an alternative embodiment of this invention, a proximity detector 27 is used for detecting the rotation of the meter encoder 40 and the output of the proximity detector 27 is applied to the signal conditioner and logic circuit 1 70a. This is illustrated in Fig. 3 by the input signals 40 designated as rotary encoder pulses and meter test proximity detector; it is understood that only one of these inputs is made at a time to the circuit 1 70a. The output of the linear optical encoder 26 is applied via the signal conditioning and logic circuit 1 70b and interface 179 to the I/D bus 140.

In order to provide an indication of the measured parameters, such as temperature, pressure, 45 as well as the fluid volumes drawn by the prover 10 and as measured by the meter 38, outputs are applied from the I /D bus 140 via the circuit 136 to the display unit 111 which includes a hard copy data printer as manufactured by Practical Automation, Inc. under their designation No. DMTP-3. In particular, the circuit 136 includes an interface circuit 1 92a for applying parameter output signals via a signal conditioner and logic circuit 1 90a to the hard copy printer. 50 Further, the circuit 136 includes an interface circuit 1 92b providing the parameter output signals via a signal conditioner and logic circuit 1 90b to the CRT data terminal 112. In addition, operator input command signals as input on the terminal's keyboard are transferred via logic circuit 1 90b and the interface circuit 1 92b to the I/D bus 140. A clock circuit 193 controls the baud rate at which signals may be transferred between the CRT data terminal 112 and the computer system. The CRT data terminal 112 may illustratively take the form of a CRT display as manufactured by Hazeltine Corporation under their designation 1500. Such a terminal permits input commands via the alphanumeric keys upon its keyboard, as well as to display the commands being entered, and the operator accessed parametric data.

Finally, there is shown a signal conditioning and interface circuit 138 for interconnecting the 60 I/D bus 140 and the inlet valve 34 and the exhaust valve 36, as well as to apply the control signals to the servomotor 20. In addition, the servomotor 20 is associated with a motor control such as the DC Servocontroller as manufactured by Control Systems Research, Inc. under their designation NC 10 1 whereby feedback signals indicative of the speed of the servomotor 20 are applied via the logic circuit 194 to in turn effect the servomotor 20 control. In addition, signals 65 4 GB 2 142 730A 4 changes in the volume of the chamber 28. As will be explained, the techniques of establishing electromagnetic waves in the microwave range permits an accuracy to at least one part in 106 of the volume of the chamber 28. In addition, in order to test the accuracy of the temperature sensors 48 and 57, a high precision temperature measuring device 68 is also inserted within the conduit 30. The temperature transducers 48 and 57 may illustratively take the form of a 5 RTD Model No. 601222 temperature transducer as manufactured by Senso- Metrics, Incorpor ated, whereas the high precision temperature transducer 68 may be of the type as manufac tured by Hewlett Packard under their Model No. 1811 5A; this transducer 68 must be used with a Hewlett Packard 2804A quartz thermometer. The proximity switches 50, 52, and 54 may illustratively take the form of a proximity switch as manufactured by Microswitch under their 10 designation FMSA5, whereas the limit switches 49 and 53 may be of a type as manufactured by Microswitch under their designations BA-2R-A2 and ADA3721 R. The servomotor 20 may illustratively take the form of that manufactured by Control Systems Research, Inc. under their designation NC100 DC Servo Motor/Tachometer. The optical encoder 26 and scale 24 may take the form of a Pos-Econ-5 linear encoder as manufactured by Heindenhain Corporation. As 15 will be explained later in detail, the measurements of pressure and temperature within the meter 38 and within the prover 10 are used to calculate a factor by which the volume as drawn into the prover 10 is adjusted for these variables. In particular, the temperature transducers 48 and 57 measure the temperature at the bottom and topmost portions of the chamber 28 to obtain a spatially averaged measurement of the temperature across the entire volume within the chamber 20 28 of the prover 10. In similar fashion, the temperature transducers 42 and 44 provide indications of the temperatures of the gas at the inlet and outlet of the meter 38 whereby a spatial averaging of the temperature of the gas flowing through the meter 38 may be obtained.

The temperature transducers as selectively incorporated into the meter prover 10 are highly stable, low-thermal-mass platinum resistance thermometers and provide accurate readings of 25 these variables whereby the volume as drawn into the meter prover 10 may be accurately determined.

As shown in Figs. 2A and B, the proximity devices 50, 52, and 54, as well as the limit switches 49 and 53 are supported upon the upright member 90, which is in turn supported from one of the upstanding struts 88. One end of the linear scale 24 is supported from the 30 member 94 and extends downward substantially parallel to the direction of the movement of the piston 14, having its lower end supported upon an upper portion of the interior of the cylinder 12. Only a limited number of the some 40,000 markings 102 are shown in Fig. 2A. In the lowermost portion of Fig. 2A, the inlet valve 34 and the exit valve 36 are actuated, respectively, from open to closed positions by pneumatic actuators 66 and 64, respectively.

In Fig. 2A, there is shown a sealing means indicated generally at 78 for the piston 14 to prevent the fluid as drawn through the meter 38 and into chamber 28 from leaking about the edges of the piston 14 as it is moved rectilinearly within the cylinder 12. The details of this piston seal 78 are disclosed in the U.S. patent application entitled "Piston Seal", filed concurrently herewith and assigned to the assignee hereof.

Referring now to Fig. 2C, the meter prover 10 is shown as being disposed within a controlled environment formed by an enclosure 104 comprising a prover room 106 for receiving the meter prover 10 and a control room 105, wherein the control's console is disposed and includes a display unit 111 containing a printer and a series of lights for indicating the various states of the meter prover, a CRT terninal 112 including a keyboard whereby various commands may be 45 input by the operator, and a variety of heat generating equipment including power supplies, motor control, amplifiers, etc. The control of the ambient conditions about the meter prover 10 is insured by disposing the meter prover 10 within the prover room 106 remote from the heat generating display unit 111 and terminal 112. As shown in Fig. 2C, the meter 38 to be tested is also disposed within the prover room 106 and is coupled to the meter prover 10 by the 50 conduit 32. The temperature in the prover room 106 is measured by the four temperature sensitive devices RO, R1, R2, and R3 disposed about the prover room 106 and upon a strut 76 of the prover 10. A barometric pressure transducer 109 and a barometer are also disposed within the prover room 106 to measure the ambient pressure. Electrical connections are made to the various temperature measuring devices as shown in Fig. 2C, as well as those temperature 55 and pressure measuring devices as shown in Figs. 1 and 2A and are directed through a pair of -troughs 110 to the control console disposed within the control room 105. In this manner, the ambient conditions under which the meter 38 to be tested and the meter prover 10 operate can be accurately controlled to insure the integrity of the measurements being made upon the meter 38 and the meter prover 10.

Referring now to Fig. 3, there is shown a functional block diagram of the computer architecture of the computer system implementing the various functions including processing of the temperature and pressure measurements, of the linear encoder and meter encoder output signals, and to appropriately close the valves 34 and 36. In addition, outputs are provided to the CRT terminal 112 to indicate the measured parameters as well as to the system control and65 3 GB 2 142 730A 3 switch 53 and the lower failsafe switch 49 are used to prevent physical damage to the meter prover 10 if for reason of failure, the servomotor 20 should continue to drive the piston 14 to either extremity. If the abutment 92 should engage either of the failsafe switches 49 or 53, the servomotor 20 will be deenergized and the piston 14 brought to an abrupt halt. Also in a volume self-test mode of operation, as will be explained, the proximity detectors 52 and 54 are used to detect the movement of the piston 14 between designated locations.

In general, the servomotor 20 accelerates the piston 14 to a given speed and the output of the linear encoder 26 is gated by the output of the proximity detector 52 to permit its pulses to be accumulated and counted. The counting of the pulses derived from the linear encoder 26 is gated off, in the volume self-test mode, by the occurrence of an output from the proximity detector 54 indicating the passage of the abutment 92 there past. Thereafter, the servomotor is decelerated to a stop. By contrast, in the meter test mode, the piston 14 is driven upward by the servomotor 20 and when the abutment 92 passes the proximity detector 52, an enable signal is generated thereby whereby on the occurrence of the next or leading edge of the next output pulse. from the meter encoder 40, the counting of the output pulses of the linear encoder 15 may then begin. In the meter test mode, the counting of the linear encoder pulses is terminated when the counting of the meter encoder pulses has reached a predeterpined count correspond ing to a volume of fluid drawn through the meter.

In order to facilitate an understanding of-this invention, a brief summary of its operation will now be given, while a more detailed discussion of the operation of the meter prover 10 will be 20 provided below. A first or initialization mode determines whether the piston 14 is in its park position as by determining whether the proximity detector 50 detects the presence of the abutment 92 as explained above; if not, the servomotor 20 is energized to drive the piston 14 to its park position. If the abutment 92 is in a position to be detected or before the piston 14 is returned to its park position, the second or exit valve 36 is opened to permit the exit of fluid driven from the chamber 28 through the conduit 30, and then the inlet meter valve 34 is'closed to prevent the fluid from being driven therethrough and possibly injuring the meter 38. Upon command of the operator that a meter 38 is to be tested, first meter valve 34 is opened and then the second valve 36 is closed to permit the flow of fluid through the meter 38, the conduit 32, the open valve 34 and the conduit 30 into the chamber 28, as the piston 14 is being driven in an upward direction by the servomotor 20. The piston 14 is gradually accelerated to a given steady state velocity and is maintained at that selected velocity during the course of the fluid volume test measurement, while the output pulses of the encoders 26 and 40 are accumulated by an arithmetic unit including registers to accumulate counts indicative of a precise volume as measured by the linear encoder 26 and of the volume as measured by the 35 meter 38, respectively. The meter test is initiated by the passage of the piston 14 past the start test proximity detector 52 that enables upon the occurrenceof the next output pulse or more precisely its leading edge from the meter encoder 40, the counting or accumulation of the meter encoder pulses as well as the linear encoder pulses. Dependent upon the desired volume to be drawn through the meter 38 under test, the meter test will terminate upon the counting of a number of meter encoder pulses. In particular, the register for accumulating the meter encoder pulses 38 upon counting the predetermined number dependent upon the fluid volume, provides an output applied to the linear encoder system terminating its counting of the input pulses derived from the meter encoder 40. The stored counts indicative of the fluid volume as measured by the meter encoder 40 and the linear encoder 26 are compared, i.e., a ratio 45 therebetween is obtained to provide an indication of the meter registration.

In addition, measurements of temperature and pressure are taken in order that the measured volumes may be adjusted for these conditions.. In particular a pair of temperature measuring devices 42 and 44 are respectively disposed at the entrance and exit ports of the meter 38. A differential pressure transducer 46 is disposed to measure the difference between the pressure 50 established by the fluid in the conduit 32 and ambient pressure. In addition, temperature measuring devices 48 and 57 are disposed respectively at the conduit 30 coupled to the chamber 28 and upon the piston 14 to provide indications of the temperature of the fluid within the chamber 28. In addition, a second differential pressure transducer 51 is disposed within the piston 14 to provide an indication of the differential pressure between the ambient pressure and 55 that established within the chamber 28. The temperature outputs TM3 and TM4 derived from the temperature measuring devices 42 and 44, respectively, are averaged to provide an average meter temperature AMT, whereas the outputs TP1 and TP2 of the temperature measuring devices 57 and 48, respectively, are averaged to provide an indication of the average prover temperature APT. As will be explained in greater detail later, these input parameters are used to 60 provide an adjustment of the measured volumes as derived from encoders 26 and 40 dependent upon the measured conditions of pressure and temperature.

As shown in Fig. 2A, a microwave antenna 70 is disposed in the head 60 to generate microwaves within the chamber 28, whereby its volume may be accurately determined. It is contemplated that this measurement may be made periodically to detect even the minutest 2 GB 2 142 730A 2 dimensions of the cavity and frequencies of excitation; and Figure 14 is a cavity response curve showing reflected power Pr as a function of the excitation frequency.

Detailed Description of Invention

Referring now to the drawings, and in particular to Figs. 1 and 2A, there is shown the meter prover system 10 of this invention as coupled to a meter 38 to be tested. The meter prover system 10 includes a cylinder 12 through which a piston 14 is driven in rectilinear fashion, by a programmable, variable speed motor 20 such as a servomotor.

The cylinder 12 is supported in upright position by a series of struts 76 (only one of which 10 are shown) secured to a collar 77 which in turn is secured around the exterior of cylinder 12. The upper end of cylinder 12 is closed by a head 86 from which a series of struts 88 (only one of which is shown) extend upwardly. A support plate 94 is fixed to the upper end of struts 88 and servomotor 20 is mounted on the top of plate 94. The upper end of a lead screw 18 is journalled for rotation in plate 94 by means of bearing 98 and is drive connected to the drive 15 shaft of the servomotor 20 by means of a coupling 100. A lead nut 22 fixed within a housing 23 is threadedly received on lead screw 18. The lead screw 18 is telescoped within sleeve 17, the upper end of which is secured to the housing 23. The lower end of sleeve 17 projects through and is slidingly received in bushing 96 in head 86. The upper end of piston shaft 16 is secured to the lower end of sleeve 17.

An intermediate cylinder head 91 separates the upper portion of the interior of cylinder 12 from the lower portion in which the piston 14 is contained. Piston shaft 16 projects through and is slidably received in bushing 93 in head 9 1, the lower end of shaft 16 being connected to the piston 14.

Thus, as the servomotor 20 rotates, the lead screw 18 rotates in nut 22 causing the housing 25 23, sleeve 17, and shaft 16 to move vertically in either direction depending on the direction of servomotor rotation.

The bottom of cylinder 12 is closed by a head 60 and the cylinder 12 therefore encloses and defines between the piston 14 and head 60 a variable volume chamber 28. An opening 62 in head 60 places chamber 28 in communication with conduit 30, conduit 32, and meter 38. A 30 first inlet valve 34 is disposed between the cylinder 12 and the meter 38 to control the flow of fluid, e.g., a gas, therebetween. A second, exit valve 36 is connected to the conduit 30 in order to permit the exit of fluid from the cylinder 12 when the valve 36 has been opened.

The precise position of the piston drive shaft 16 and therefore the piston 14 is provided by a high precision, linear optical encoder 26 that is coupled to the drive shaft 16 to move therewith. 35 More specifically, the encoder 26 illustratively includes first and second sets of light sources and photocletectors disposed on either side of a linear scale 24 having a high number of scale marks 102. In one illustrative embodiment of this invention, the linear scale 24 is disposed in a fixed position with respect to movable encoder 26 and includes 40,000 scale marks 102 (2500 marks per inch); of course, only a limited number of such a high number of scale marks could 40 be illustrated in the drawings. Thus, as the encoder 26 is moved rectilinearly along the length of the linear scale 24, first and second sets of pulse trains A and A are developed, 90 out of phase with respect to each other, as the light beams generated by the first and second light sources are intercepted by the scale marks 102. The-outputs A and A from the optical encoder 26 indicate precisely the,position of the piston 14 and likewise the volume of fluid has been 45 drawn through the meter 38. As will be explained, the volume of the chamber 28 within the cylinder 12 is precisely measured, and each output pulse derived from the encoder 26 provides a precise indication of an incremental volume as drawn into the chamber 28 within the cylinder 12 as the piston 14 is withdrawn, i.e., is directed upward by the servomotor 20. Before the piston 14 is begun to be raised thus creating a vacuum within the chamber 28, the exit valve 50 36 is closed and the meter valve 34 is opened to permit a flow of fluid through the meter 38 via the conduit 32, the open valve 34 and a pair of conduits 30 and 32 into the housing 28.

During a meter test, the meter 38 provides via its encoder 40 an output train of pulses indicative of the flow of fluid therethrough. The train of pulses as derived from the meter encoder 40, is compared to the train of pulses derived from the linear encoder 26 to provide an 55 indication of the meter accuracy in terms of the meter registration, a ratio corresponding to the volume measured by the meter encoder 40 to the volume measured by the linear encoder 26.

Further, the programmed movement of the piston 14 uses a plurality of proximity sensors 50, 52, and 54, as well as a pair of limit switches 49 and 53. As will be explained, the servomotor 20 drives the piston 14 rectilinearly within the housing 28. As shown in Figs. 1 and 2A, the 60 piston 14 is in its uppermost position wherein an abutment 92 on housing 23 26, contacts and closes the upper limit switch 53 thereby deactuating the servomotor 20 when the piston 14 is driven upward and thereby halting the movement of the piston 14. When the servomotor 20 drives the piston in a downward direction as shown in Fig. 1, the abutment 92 may then engage the lower limit switch 49 again bringing the piston 14 to a halt. The upper failsafe 1 GB 2 142 730A 1 SPECIFICATION

Measurement of chamber volume This invention relates to apparatus for and methods of measuring the volume of a chamber with 5 particular reference to the precision measurement of the effective volume of the chamber of a meter prover used for testing the accuracy of fluid and in particular gas meters.

A method of and apparatus for meter prover calibration utilising the present invention are described and claimed in our co-pending British patent application number 8108387 (2072854) of even date herewith.

Reference may also be had to our co-pending British patent application 8108389 (2072855) describing and claiming fluid flow meter calibrating apparatus.

The object of the invention is to provide improvements in techniques for measurement of chamber volume.

According to the invention there is provided:

Claims

(a) apparatus for measuring the volume of the chamber of a meter prover as defined by Claim 1 of the appended claims, or (b) a method of measuring the volume of a chamber as defined by Cl&im 10 of the appended claims. Brief Description of the Drawings A detailed description of one preferred embodiment of this invention is hereafter described with specific reference being made to the drawings in which: Figure 1 is an elevational view of a meter prover system in accordance with the teachings of this invention; Figures 2A and B are detailed, partially sectioned views of the meter prover of this invention as more generally shown in Fig. 1, Fig. 2C shows the structure for housing the meter prover as shown in Figs. 1 and 2A and a control console whereby the operator may control and observe the data output from the meter prover, and Fig. 2D shows the display panel of the system control and status module as shown in Fig. 2C; Figure 3 is a functional block diagram of the architecture of the computer system used to sense the several variables of the meter prover system and to control the movement of the piston through the cylinder of the prover as shown in Figs. 1 and 2A, as well as to provide an accurate indication of the meter registration of the tested meter. Figures 4A to 4Eshow the various signal conditioning and interface circuits as are needed to 35 provide signals into and from the computer system as shown in Fig. 3, Fig. 4F shows a perspective view of the meter and the manner in which a proximity detector is disposed with regard to its encoder mechanism, and Figs. 4G through M show detailed schematic diagrams of the signal conditioning and interface circuits generally shown in Figs. 4A and E; Figure 5 shows a high level diagram of the program as executed by the computer system of 40 Fig. 3; Figures 6A and 6B show in a more detailed diagram, the initialization process effected by the computer system of Fig. 3; Figure 7 shows in a more detailed fashion the steps necessary to calibrate the inputs as applied to the signal conditioning and interface circuits for applying the measurements of 45 temperature and pressure to the computer system as shown in Fig. 3; Figures 8A to 8H, 8J to 8N and 8P disclose in detail the flow diagram for reading out data stored within the computer system and to enter the conditions under which the meter prover system will test a given meter (the letters I and 0 are not used for clarity sake); Figures 9A to 9H, 9J to 9N, 9P and 9Q show the steps effected by the computer system of 50 Fig. 3 to carry out the various tasks and to provide manifestations thereof (the letters I and 0 are, not used for clarity sake); Figure 10 is a schematic diagram of a circuit for applying a high frequency signal to the microwave antenna within the cylinder of the meter prover as shown in Figs. 1 and 2A, for varying the signal's frequency as applied to the antenna whereby the volume of the cylinder 55 may be determined with great accuracy to thereby accurately encode the output of the meter prover's linear encoder; Figures 1 1A through E show variously the input signals applied to and the output signal as developed by the signal conditioner and logic circuits 1 70a and 1 70b as shown generally in Fig. 4C and more specifically in Fig. 4K; Figures 12A and B show respectively a perspective view of a cover to be placed over the piston as shown in Figs. 1 and 2A, and a cross-sectional view of the spring-like seal disposed about the periphery of the piston cover; Figure 13 is a graph illustrating the response of a chamber of a right circular cylinder configuration to being excited with high frequency electromagnetic fields in terms of the varying 65 13 GB 2 142 730A 13 encoder 40 is initialized. In particular, the logic circuit 1 70a is initialized, the interval timers 174 and 17 6 are cleared, the PI /0 circuit 184 is disposed to a selected mode, the logic circuits 177 and 178 are reset and the programmable counters 1 74a and 1 76a are programmed with the factors to be entered therein to be counted down. In step 310, a command is sent via the PI/O circuit 196 to the logic circuit 1 94b to effect opening of the second exhaust valve 36. In 5 particular, the proximity detector associated with the valve 36 is accessed and if the valve 36 is not open, the program will wait until the valve 36 is open. Further in step 310, the control circuitry associated with the servomotor 20 is initialized in that the speed of the servomotor 20 is set to zero, and a signal is applied to the servomotor 20 to maintain it in a stationary condition, while a logic circuit for sensing the torque of the servomotor 20 is reset; the noted 10 logic circuit is coupled to sense the control current flowing to the servomotor 20. In step 312, the logic circuit 1 92b is initialized to its static 1/0 mode with "hand- shaking" output capabilities, and the logic circuit 1 90b clears and prepares the associated printer 111 a to begin printing. In step 314, the SDC 1 92a is programmed to insure that the data may be transmitted between the CRT terminal 112 and the CPU 120, and the logic circuit 190a of Fig. 4D similarly 15 instructs CRT terminal 112 to clear and prepare itself for operation to receive data as well as to clear the CRT display. Next in step 316, any data stored in the A/D converter 166, as shown in Fig. 413, is cleared and the multiplexer 163 is set to its first channel whereby the output of the transducer 51 is applied by the multiplexer - 163 to the A/D converter 166. Step 318 clears the 20 conditioning and interface circuit 130, and in particular clears any data stored in the A/D circuit 20 158, as shown in Fig. 4A, as well as to set the multiplexers 149 and 154 to their first channels whereby the output voltage of the prover temperature transducer module is applied via the multiplexers 149 and 154, and the operational amplifier 152 to the A/D circuit 158. In step 320, command signals are sent via the I/D bus 140, and the PI/O circuit 196, as seen in Fig. 25 4E, to cause the logic circuit 1 94b to actuate the solenoid 66 to open and then close the first or 25 meter valve 34 and to actuate the solenoid 64 to close the second or exhaust valve 36. When it is desired to open or close one of the valves 34 or 36, the proximity detector associated therewith is interrogated and if it is determined to be in the desired position, no further action is taken; if, however, the valve is not in the desired position, the logic circuit 1 94b provides an output to actuate the associated pneumatic solenoid to cause the valve to open or close, as desired. In step 322, the program continues to a subroutine as will be explained with respect to Fig. 6B whereby the piston 14 is returned to its park position, i.e., the lowermost position corresponding to the position of the proximity detector 50 as seen in Figs. 1, 2A and 2B. In Fig. 6B, the subroutine 322 for returning the piston 14 to its lowermost position is shown. In step 324, a control signal is sent via the I/D bus 140, the PI/O circuit 196 to cause the logic circuit 1 94b to close the first or meter valve 32. Next, step 326 determines the position of the piston 14 which may be anywhere from its lowermost to its topmost position as seen in Fig. 1; in particular, step 326 causes the PI/O circuit 196 to interrogate the logic circuit 1 94c to determine if the output of the proximity detector 50 as seen in Figs. 1 and 2 is high or one, and if so, the subroutine moves to the final step 346 wherein a command is sent to the PI/O circuit 40 196 of Fig. 4E to condition the logic circuits 1 94a and 1 98b so that the motor control for the servomotor 20 is set for zero speed and further to set the interval timer 176 as seen in Fig. 4C to indicate that the output of the encoder is in its starting position, i. e., to set the counter 1 76a to zero preparing it to start generating one second sampling pulses. If the piston 14 is not at its park position, the subroutine proceeds to step 328, wherein a determination is made as to whether the piston 14 is disposed in an intermediate position, i.e., the abutment 92 is disposed between the proximity detectors 50 and 52 as shown in Figs. 1 and 2A and B; if piston 14 is so disposed, the subroutine moves to step 330, whereby control signals are sent via the PI/O circuit 196 to set the logic circuit 1 94a to drive the servomotor 20 in a counterclockwise direction, and further in step 332, the speed is controlled to accelerate to a selected speed, i.e., 50 the ninth of sixteen speeds. However, if step 328 determines that the piston 14 is in its uppermost position, i.e., its abutment 92 is above the proximity detector 52, step 334 sets the logic circuit 1 94a of Fig. 4E to set the lowest speed, i.e., speed 1 to 16 and to rotate the servomotor 20 in a counterclockwise direction, before proceeding to step 336 wherein the logic circuit 1 94a effects acceleration of the servomotor 20 to its next highest speed up to a maximum of its twelfth speed. At that point in step 338, the logic circuit 1 94c is accessed to determine whether the piston 14 is disposed at the proximity detector 52 and if not, step 336 accelerates the piston to its next higher speed until step 338 determines that the piston 14 is at the proximity detector 52. At that point, the logic circuit 1 94a maintains the current speed of servomotor 20, until in step 342, the park proximity detector 50 detects the presence of the 60 piston 14 at which time, step 344 controls the logic circuit 1 94a to decelerate the piston 14 to a stop position before entering the step 346 wherein the motor control is set for zero speed. If the operator has set the calibrate/run switch to its calibrate position, the program moves to the subroutine 400 as more fully shown in Fig. 7. Initially in step 402, a command is sent to the logic circuit 1 90a of Fig. 4D to clear the CRT terminal 112. Step 404 displays a suitable 65 14 GB 2 142 730A 14 message upon the CRT screen indicating the meter prover system 10 has entered into a calibrate mode indicating, as shown in Fig. 7, the various parameters that may be so calibrated. Next, step 406 returns the cursor as displayed upon the CRT to its left-hand margin waiting for the operator to make a suitable entry. Next- after operator entry via the keyboard of the CRT terminal 112, step 408 interrogates the keyboard to determine which of the possible keys has been depressed. For example, if it is determined that one of the sets of keys T and 0, T and 1, T and 2, T and 3, T and 4, T and 5, T and 6, or T and 7 has been depressed, the subroutine moves to step 410, wherein a corresponding channel is selected by the multiplexer 149. Next in step 412, a select command signal is transmitted via the PI/O circuit 160 to the multiplexer 149, whereby the selected channel is connected via the multiplexers 149 and 154, the amplifier 152, to the A/D converter 158. Next, step 414 applies a convert signal to the A/D circuit 158 whereby in step 416 the temperature as measured is displayed. At this point, the operator can calibrate the selected temperature transducer module to provide a correct reading by adjusting the zero and span of the operational amplifier of the module, this procedure being repeated until an accurate reading is displayed upon the CRT. Though detailed explanation will 15 not be given, it is realized that the similar keys P and 1, P and 2 and B may be also operator activited, whereby the gains of the operational amplifiers 16 1 a, 161 b, 16 1 c of Fig. 4B may be similarly adjusted in order to give accurate readings. In similar fashion, the keys T and 0, T and 1, T and 2, T and 3, T and 4, T and 5, T and 6, and T and 7 may be depressed and their respective amplifiers and circuits may be adjusted to provide accurate readings. Upon pressing 20 of the keys V and P, as determined by step 408, the routine moves via step 450 to step 452, whereby a command signal is sent via the PI/O circuit 168 of Fig. 4B to cause the multiplexer 163 to apply the voltage calibrate input signal V to the A/D converter circuit 166, which converts these inputs in step 454 to digital values. The voltage calibrate input signal is adjusted to set the zero and full scales values of the A/D coverter 166, these values being appropriately 25 displayed in step 456 upon the CRT terminal 112. In similar fashion if the V and T keys of the keyboard is depressed, the routine proceeds via step 442 to step 444, wherein the voltage calibrate input signal VT is applied to the A/D converter circuit 158 of Fig. 4A and its zero and full scale values may be similarly adjusted. If any other key is depressed upon the keyboard, step 408 branches via step 456 to step 458, wherein that character is rejected to return to the 30 starting point of step 408 again. The data input and retrieval subroutine as broadly shown in Fig. 5 as step or routine 500, is more fully explained with respect to Figs. 8A to 8P, an overview of the routine 500 being shown in Fig. 8A. Initially, a command is given via the I/D bus 140 to the logic and driver circuit 1 98a of Fig. 4E to energize the standby light 111 e. Next in step 504, which is more fully shown in Fig. 8C, a determination is made if a self-test is desired, and if so, that subroutine is executed. Next in step 506, it is determined whether it is desired to display the room temperatures, and if so, a display upon the CRT of the terminal 112 is made, as will be more fully explained with regard to Fig. 8H. In step 508, an examination of the keyboard is made to see if the operator wishes to recall data pertaining to previous meter accuracy tests and if any 40 data is required, that data is displayed upon the CRT display; the subroutine 508 will be more fully explained with regard to Fig. 8J. In step 510, appropriate flags are set automatically to permit, at a later time during the meter test, the pressure and temperature parameters to be monitored and displayed; the subroutine for executing such display is more fully explained with respect to Fig. 8K. In step 512, the subroutine for setting the test volume desired and the 45 manner in which it is entered into the programmable counter 1 74a of the interval timer 174 is set out in more detail with the subroutine as shown in Fig. 8L. Next in step 514, it is necessary for the operator to enter via the CRT terminal keyboard a select test flow rate or rates, as will be more fully explained with regard to the subroutine of Fig. 8M. In step 516, the operator sets the number of times to repeat a certain test; for example, the meter may be tested three times for a 50 flow rate or for a selected volume and flow rate. The entry of the repeat commands is explained with respect to the subroutine of Fig. 8N. Thereafter, step 518 sends clear commands via the logic circuit 1 90a whereby any data stored in the buffers associated with the CRT terminal 112 are cleared. At this point, step 520 makes a decision as to whether the volume of the meter 38 is to be determined with the output of the proximity detector 27 or wi th the output of the rotary 55 encoder 40. If the switch is set to a logic "zero", a flag is set in step 522 to conduct an encoder type test, whereas if the switch is set to a logic "one" position, a flag is set in step 524 to conduct the proximity detector type test. Continuing the routine 500 as shown in Fig. 813, the system obtains and displays upon the CRT the meter temperature and pressure, and the prover temperature and pressure as measured 60 during the meter test, as follows. Next, step 528 displays, if instructed, upon the CRT data indicative of the pressure and temperature within the meter 38 and the prover 10. Once the test volume, either one cubic foot, one-half cubic foot, or one-fourth cubic foot, has been set by entry of the operator upon the CRT terminal 112, the entered value is decoded in step 530, and the chosen value is displayed in step 532 by the CRT as: TEST VOL. = X CU. FT. Thereafter, 65 GB 2 142 730A 15 the desired flow rate or rates that are entered upon the keyboard are decoded in step 534 and the chosen rates are displayed upon the CRT in step 536 as: FLOW RATES are QX,QX-QX. In similar fashion in step 538, the number of repeats for any particular test(s) with respect to flow rates or volumes are decoded and in step 539, the selected number of repeats per flow rate or volume is displayed upon the CRT as; NO. OF TESTS PER FLOW RATE = X. At this point, a carriage return line feed instruction is carried out in step 540 whereby a cursor as placed by the CRT terminal upon the CRT is removed. The subroutine 504 is more fully shown in Fig. 8C, wherein a desired selftest(s) is selected. Initially in step 542, the data storage buffers associated with the CRT of the CRT terminal 112 are cleared by commands generated by the logic circuit 1 90a. In step 544, the CRT displays, as10 shown in Fig. 8C, an indication of the various self-tests that may be conducted, e.g. volume V, leakage L, or NO; the operator selects one of these self-tests by depressing the appropriate key of the keyboard. In step 546, the cursors, as shown in Fig. 8C, are flashed to prompt the operator to respond.to select the self-tests, whether volume V, leak L or NO. In step 548, the subroutine moves to the desired self-test dependent upon which key has been selected in step 544. If for example, the volume self-test V has been selected, a volume self-test sequence is output to the CRT in step 552 and the volume self-test is executed in step 554 as will be more fully explained with respect to Fig. 8F. If a leak test has been chosen, the subroutine moves through step 558 to provide a display upon the CRT indicating the selected test and to initiate the leak self-test, which is executed in step 560 as will more fully be described with respect to 20 the subroutine of Fig. 8D. Next the subroutine determines in step 556 whether there is another input to be dealt with and the subroutine returns to step 546. If the operator pressed the escape key as determined By/ step 548, the subroutine exits via Step 562 to the entry point 212 of the main diagram as shown in Fig. 5 whereat the program begins. If the operator strikes the N and 0 keys, as detected by the step 548, the subroutine exits via step 566 and returns to the routine as shown in Fig. 8A to continue with the next step 506. If any other key is struck, -the subroutine exits via step 568 to reject that character in step 570 and to return to step 548 to recognize another key. The leak self-testing subroutine 560 is more completely shown with respect to Fig. 8D. The leak self-test is a test tO provide an indication of the integrity of the prover system's seal. Initially in step 572, the logic circuit 1 98a in Fig. 4E energizes the self-test and test in progress lights 111 b and 111 c. Next, the subroutine 322 For returning the piston to its park position as explained above with respect to Fig. 613, is executed. Next in step 574, a command signal is issued in step 574 via the PI/O circuit 168 to set the Multiplexer 163 to receive the signal via the second channel from the pressure transducer 51 (pressure PP1). Next in step 576, the logic 35 circuit 1 94b is commanded to apply actuating signals to the solenoids for effecting the closing of the first inlet and second exhaust valves 34 and 36. In step 578, a command is issued via the PI/O circuit 196 to actuate the logic circuit 1 94a of Fig. 4F to direct the servomotor 20 to drive the lead screw 18 in a clockwise direction at a relatively slow speed corresponding to " 1 ". The servomotor 20 cannot be energized as directed by step 578, until the prohibition built into 40 the program is defeated as by depressing a momentary switch (not shown) coupled to the logic circuit 1 94c, thereby permitting the opEraTion of thE servomotor 20 with both valves 34 and 36 closed. Normally, the servomotor 20 may not be operated if the valves 34 and 36 are closed, thus preventing possible damage to the piston seal. Next in step 580, an indication of pressure is input iNto the A/D converter 166 in step 580 and a check is made in step 582 to 45 determine whether the pressure has increased by a certain amount as indicated by the data stored in the A/D circuit 166, i.e., has the vacuum increased exceeded 1. 38 inches of water? If the vacuum has not increased to beyond 1.38 inches of water, the subroutine cycles back and repeats step 580. If the vacuum inside the chamber 28 has increased to beyond 1.38 inches of water, the subroutine moves to step 584 whereby a command is given through the PI/O circuit 50 196 to cause the servomotor 20 to stop. In step 586, a 20-second wait is loaded into a register within the RAM 126 and counted down to provide a time for leakage to occur in the chamber 28. Next, in step 588, a convert signal is applied to the A/D converter 166 of Fig. 413, to determine the pressure now being sensed by the pressure transducer 51. In step 590, the subroutine waits for the one-second clock signal which is developed from the logic circuit 178 55 of Fig. 4C and is applied via the logic circuit 1 94d of Fig. 4E to decrement in step 592 the 20 second count placed in the register within the RAM 126. Step 592 determines whether this register in the RAM 126 has been decremented to zero, and if not, the subroutine returns to step 588, whereby for each second of a 20-second interval, the piston pressure as derived from the transducer 51 is obtained and is displayed upon the CRT thus providing a continuous monitor of the prover pressure so that the operator may determine whether there are any leaks within the chamber 28 during the 20-second test. The count decrementing continues until the register is zero as determined by step 592, at which time a command is issued in step 594 to cause the logic circuit 1 94b to actuate the solenoid to open the second exhaust valve 36. At this point, the position of the piston 14 is determined in step 596 by checking the status of the65 16 GB 2 142 730A 16 proximity detectors 50 and 52 by accessing the logic circuit 1 94c. At this time the test is complete and a command is issued to the logic circuit 1 90a to energize the test complete and self-test indicators. Finally, in step 322, the piston 14 is returned to its park position by the sub routine as shown in Fig. 613, before returning to step 556 as shown in Fig. 8C. In Fig. BE, the volume self-test subroutine 554 is shown in more detail. First, step 600 commands the logic circuit 1 98a to energize the self-test and test in progress indicators 111 b and 111 c. The volume self-test is entered to obtain an indication of the accuracy of the linear encoder 40. Step 332, as shown in Fig. 613, returns the piston 14 to its park position before command signals are issued in step 602 via the logic circuit 1 94b to actuate the solenoids to open the first, inlet valve 34 and to actuate the solenoid to close the second exhaust valve 36, 10 Next, a command is made in step 604 to load the speed buffer, i.e., an addressed portion within the RAM 126, with its maximum speed of 16. Next in step 606, the servomotor 20 is accelerated to the speed as stored in this buffer in accordance with the subroutine as will be discussed with respect to Fig. 8F. Next in step 608, the counter 182 and the programmable counter 1 74a as shown in Fig. 4C are inhibited and are reset to zero. Step 610 checks the 15 status of the proximity detector 52 to determine whether the piston 14 has been drawn to its start-test position, and if not, the check is repeated until the piston 14 reaches this position. When the piston 14 has reached its start-test position, the program moves to step 612 wherein the programmable counter 1 74a is disabled and the linear counter 182 is enabled to count the linear encoder pulses as the piston 14 is drawn from the proximity detector 52 to proximity 20 detector 54. In step 614, the status of the disabled proximity detector 54 is checked periodically until the piston 14 is disposed thereat, at which time the subroutine moves to step 616 to disable the linear counter 182. In step 618, the cumulative number or count from linear encoder counter 182 is transmitted via the PI /0 circuit 184 to a designated location within the RAM 126. In step 620, a suitable series of command signals are transmitted to the logic circuit 25 194a whereby the servomotor 20 is decelerated to a stop condition as will be described with respect to the subroutine shown in Fig. 8G. Next in step 622, flags are set indicating that the piston 14 is not between the proximity detectors 50 and 52. In step 624, the logic and driver circuit 1 98a energizes the self-test and test complete lights 111 b and 111 d. At this point in step 626, the count of the counter 182 indicative of the travel of the piston 14 between the 30 proximity detectors 52 and 54 is transferred from its location within RAM 126 to be displayed by the CRT of the terminal 112 and at the same time clears counter 182. Next, the piston 14 is returned to its park position by step 332 as shown in Fig. 613, and step 628 causes the logic and driver circuit 1 98a to energize the standby light 111 e, before returning to step 556 as shown in Fig. 8C. The accelerating motor subroutine generally indicated in Fig. 8E by the numeral 606 is shown in more detail in Fig. 8F, wherein in step 630 a command is sent via the I/D bus 140 to disable the counters 182 and 1 74a of Fig. 4C. Next, an enable command is sent to the logic circuit 1 94a by step 632 to command the servomotor 20 to rotate in a clockwise direction, thus raising the piston 14. At this point, speed 1, i.e., the lowest speed, is loaded into a speed 40 compare buffer within the RAM 126 by step 634. In step 636, a command is given to the logic circuit 1 94a to increase incrementally the speed of the servomotor 20 and the actual speed as stored in the speed compare buffer of the RAM 126 is compared with the final designated speed stored in the speed buffer and if they are equal, i.e., the servomotor 20 has been accelerated to the desired speed, the subroutine moves to step 646. If not, step 640 implements a wait period of approximately 0.25 seconds before sampling via the logic circuit 1 94c the status of the proximity detector 52 and if not at this position, step 644 increments the desired speed by 1 before returning to step 636 in which the new speed is disposed in the speed compare buffer of the RAM 126. If the piston 14 has been driven into its intermediate position as detected by step 640, the subroutine moves to step 642, to set a flag indicating that 50 the piston 14 has been prematurely brought to its intermediate position and the subroutine is returned to step 604, as shown in Fig. 8E. As seen in Fig. 8F, the normal operation is for the piston 14 to be accelerated to the desired speed and then to periodically test the output of the proximity detector 52. Upon determining in step 646 that the piston 14 has reached the start- test position as indicated by the output of the proximity detector 52, the subroutine branches to 55 step 970 as will be explained with respect to Fig. 9E to assess and determine the occurrence of the rising edge of the next pulse A from the rotary meter encoder 40, whereby as will be explained further the meter test is begun by applying the pulse outputs derived from the rotary meter encoder 40 and the linear encoder 26 into their respective counters 1 74a and 182. The subroutine to decelerate the servomotor 20 as indicated in step 620 of Fig. BE is more 60 fully explained with respect to the subroutine shown in Fig. 8G. Initially in step 648, the counter 1 74a and the counter 182 as seen in Fig. 4C for respectively receiving the rotary and linear encoder outputs are first disabled and then reset. In step 650, the load speed compare buffer within the RAM 126 is loaded with the speed 16, i.e., the highest speed available. Next, in step 652, the actual speed as stored in the speed buffer of RAM 126, is compared with the 65 17 GB 2 142 730A 17 high speed 16 as loaded into the speed compared buffer and if not equal, the speed as loaded into the speed compare buffer is decremented until the actual speed equals the speed loaded into the speed compare buffer, at which time the subroutine moves to step 656 and the speed as stored in the speed compare buffer is decremented. In step 658, a commandis issued to the 5 logic circuit 1 94a to decelerate the servomotor 20 to that new speed as disposed within the speedbuffer of the RAM 126. In step 660, the actual speed is compared with zero, i.e., the servomotor 20 is stopped, and if not stopped, a 0.2 second wait or delay is implemented in step 662, before again decrementing the motor speed by steps 656 and 658. This process continues until the servomotor 20 has been brought to a stop, i.e., the speed equals zero at which time step 660 return the subroutine to step 622 of Fig. BE. In Fig. 8H, there is shown the subroutine 506 for operator entry via the keyboard of the CRT terminal 112 of the selected temperatures within the prover room 106 to be displayed and the steps for so displaying them upon the terminal's CRT. It has been shown that knowledge of the temperature within the prover room 106 is helpful to investigate the level of control that may be exercised upon the meter prover 10 within the room 106. The subroutine as shown in Fig. 8H 15 permits the operator to display any of the four room temperature outputs RO, R1, R2, and R3, from the temperature measuring devices shown within Fig. 2C as being disposed about the prover room 106. These temperatures can continue to be accessed and displayed upon the CRT of the terminal 112 until the operator terminates this action by pressing the N and 0 keys of the terminal's keyboard. Initially in step 664, a command is sent to the logic circuit 190a, whereby 20 via the control input, the buffers associated with the CRT terminal 112 are cleared. Thereafter, step 666 causes the logic circuit 1 90a to display an indication, as shown in Fig. 8H, of the possible room temperature devices that may be accessed, i.e., the room temperature devices RO, R1, R2, or R3. If the operator does not desire to access and display any of these temperatures, the N and 0 keys are struck. In step 668, a carriage return line feed is output so 25 that the cursor is flashed, indicating that an operator response is required. In step 670, an - interrogation is made of which key is actuated via the logic circuit 1 90a and if the keys R and 0 corresponding to the room temperature device RO, are actuated, the subroutine moves via step 672a to step 674a whereby a control signal is issued via the I/D bus 140, PI/O circuit 160 to cause the multiplexer 149 to select its fifth input whereby a voltage representative of the room 30 temperature transducer RO is applied via the multiplexers 149 and 154 to the A/D converter 158. In step 676a, a convert signal is applied via the PI/O circuit 160 to the A/D converter circuit 158 to convert the analog input signal to binary digital data. Next, step 678a converts the binary digital data to decimal data and, step 680a converts that data to a floating point decimal format by processes well known in the art. In step 682a, the digital floating point 35 decimal data is converted to digital F in accordance with the formula: [(RAW DIGITAL NUMBER) (7.326007326 X 10 - 3) (9)] + 32 = F' In step 684a, suitable data is generated out from the logic circuit 1 90a to display the room temperature in degrees Fahrenheit upon the CRT display. In similar fashion, step 670 is able to detect the operator actuation of each of the pairs of keys corresponding to RO, R1, R2, and R3 room detectors to access these detectors as connected to the multiplexer 149 of Fig. 4A and to 45 process and display the selected data upon the CRT display. In step 694, a determination is made as to whether another input is desired, and the process returns to the beginning of step 670. If the operator presses the escape key, the routine moves through step 686 to return via entry point 212 to the beginning of the program as seen in Fig. 5. If the N and 0 keys are depressed, the program proceeds to step 688 and returns to the next subroutine or step 508 as 50 seen in Fig. 8A. If an extraneous key is depressed, the subroutine exits through step 690 and rejects the character in step 692 returning to the beginning of step 670. The subroutine 508 as generally indicated in Fig. 8A for sampling data from previously conducted tests is carried out by the subroutine 508 which is more clearly shown in Fig. 8J. The subroutine may be entered after a meter test has been executed. First, in step 696, the CRT, in particular its buffers and display, are cleared of previously stored data. Next, step 698 displays by actuating the logic circuit 1 90a to have the CRT display the possible test parameters that may be accessed and displayed, namely K, the number of counts from the linear encoder 26; TP, the temperature of the prover; TM, the temperature of the meter; PM, the pressure of 60 the meter; PP, the pressure of the prover; and the percentage of error from the previous test. If 60 none of these values are desired to be displayed, the operator may depress keys N and 0 corresponding to NO at which time the program advances to the next step. In step 700, the cursor is flashed indicating to the operator to select one of these parameters or NO by pressing the corresponding keys. In step 702, if the linear encoder counts are selected by depressing the key K, the subroutine branches through step 704a to step 706a, whereby the count of the 65 18 GB 2 142 730A 18 linear encoder register within the RAM 126 is accessed and read out and applied via the logic circuit 1 90a to be displayed upon the CRT. In similar fashion, if the keys corresponding to the temperature of the meter prover are actuated upon the keyboard by the operator, the process proceeds from step 702 via step 704b to step 706b wherein the contents of the prover temperature buffer, a designated location within the RAM 126, is transferred from the RAM 5 126 to be displayed upon the CRT. In similar fashion, each of the meter temperature TM, the prover temperature TP, the prover pressure PP, or the meter pressure PM may be similarly accessed from a corresponding location within the RAM 126 and displayed upon the CRT. Also the percentage of error of the difference between the standard volume as drawn by the prover system 10 and that actually measured by the meter 38 is calculated with respect to the standard volume and may be likewise displayed upon the CRT. If a second or further key has been pressed, step 716 returns the subroutine to the beginning of step 700. If the escape key has been depressed, the subroutine moves through step 708 to exit via entry point 212 to the beginning of the program as shown in Fig. 5. If any extraneous key is depressed, the routine proceeds through step 712 to reject the mistaken character in step 714 before returning to the 15 beginning of step 702. After the operator has observed any and all parameters that he desires to observe upon the CRT, he presses the N and 0 keys whereby the program returns to the next step or subroutine. Routine 510 as generally shown in Fig. 8A is capable of accessing and displaying the parameters as are automatically measured and monitored during the course of a test, and is 20 shown in greater detail in Fig. 8K. First in step 718, the buffers associated with the CRT are cleared, before steps 720, 722, 724 and 726 set flags in appropriate locations within the RAM 126 to access the prover and meter temperatures and the meter and prover pressures. Thus, the flags are set to permit display upon the CRT of these parameters automatically, and thereafter the routine 510 returns to step 512 as shown in Fig. 8A. Step 512, as generally shown in Fig. 8A, for setting the desired volume to be drawn through the meter 38 is more fully explained with respect to the subroutine shown in Fig. 8L. Initially, the CRT terminal 112 is cleared in step 742, before in step 746 displaying upon the CRT the possible volumes, e.g., one cubic foot,.5 and.25 cubic feet, that may be selected by the operator to be drawn through the meter 38 to be tested. In step 748, the cursor is flashed to 30 prompt the operator to make his volume selection. Upon actuation by the operator of the key corresponding to each of the selectable volumes as sensed in step 750, the subroutine 512 moves to enter the corresponding volume. For example, if the appropriate key is depressed indicating a desire to enter a value corresponding to one cubic foot, the subroutine moves via step 752a to step 754a wherein the scaling factor corresponding to one cubic foot is set in a known position in the RAM 126. The scaling factor will be subsequently transferred from this RAM.location to the rotary encoder counter 1 74a. After selection of any of the desired test volumes, the subroutine returns to the next step 514 as shown in Fig. 8A. If the escape key is selected, the subroutine moves through step 756 to return to the main program via entry point 212. If a wrong key is accidentally depressed, the subroutine moves through step 758 to reject 40 that entered character in step 760 before returning to the beginning of step 750. Next in the program is step or subroutine 514 wherein the desired flow rate at which the fluid is to be drawn through the meter 38 is set; as explained above, the selected flow rate in turn determines the speed at which the servomotor 20 is controlled to rotate and raise the piston 14 as shown in Fig. 1. Referring now to Fig. 8M, step 762 initially clears the buffers associated with the CRT terminal 112 and in step 764, the possible flow rates from QO to QF are displayed upon the CRT, it being understood that QO corresponds to a minimum flow rate, e.g., cubic feet per minute, and (IF corresponds to a maximum flow rate, e.g., 400 cubic feet per minute. Step 766 causes the cursor to flash on and off to prompt the operator to make his selection of flow rate. In step 768, an interrogation is made of the keyboard to determine which 50 key the operator has depressed and if for example, the flow rate QO is depressed, the subroutine exists via step 770a to step 772a wherein the flag for the particular rate QO is set in a designated location within the RAM 126. Next, step 774a causes the selected flow rate QO to be displayed upon the CRT; then, the system returns with the response of "ANOTHERT' and flashes the cursor to indicate that a user response is required. If a further rate is to be also tested, step 784 returns the routine to the beginning of step 766 whereby a second and perhaps further flow rates may be tested within a single set of tests. If the operator depresses the escape key, the routine moves through step 776 and entry point 212 to the main program as shown in Fig. 5. If a wrong key of the keyboard of the CRT terminal 112 is depressed, the routine moves through step 780 to reject the incorrect character in step 782 before returning to 60 the beginning of step 768. If no further flow rates are to be selected to test the meter 38, the operator presses the N and 0 keys, whereby the routine returns via step 778 to the next step 516 as shown in Fig. 8A. Step or routine 516 for setting the number of repeat tests that is to be set for each of the selected flow rates may be entered via the operator keyboard in a manner more specifically 65 19 GB 2 142 7 30A 19 shown in Fig. 8N. Initially, the buffers associated with the CRT terminal 112 and the display screen of the CRT are cleared in step 786, before the possible number of tests or retests for flow rate, i.e., 1, 2, or 3, is displayed upon the CRT. Step 790 flashes the cursor on the CRT to prompt the operator to make his selection of the flow rate. Step 792 detects which of the terminal keys is depressed and if for example, a key is depressed indicating only one test per 5 flow rate is to be conducted, the subroutine proceeds through step 794a to step 793a, wherein the output " 1 " is displayed upon the CRT before returning to the next step 518 of the main program as seen in Fig. 8A. If two tests are to be made, the subroutine exits via step 794b to step 793b wherein a command is made to copy the flow rate flags for one repeat and in step 796b to set a flag for a second repeat in a designated area of the RAM 126. In step 796b, an 10 indication of 2 is displayed upon the CRT. If a selected test is to be repeated three times, the subroutine carries out a similar series of steps 794c, 793c, and 796c. If an incorrect character is depressed upon the keyboard, the routine moves through step 795 to reject the incorrect character in step 797 before returning to the beginning of step 792. If the operator presses the escape key, the routine exits via step 798 to entry point 212 whereby a return is made to the 15 main program as shown in Fig. 5. Subroutine 528, as generally shown in Fig. 8A, displays temperature and pressure data as the meter test is being conducted, as more fully shown by the subroutine illustrated in Fig. 8P. Initially, the CRT cursor is moved to a home position and the CPU 120 addresses that location within the RAM 126 where the TM flag has been set automatically during the previous subroutine 510. Since the TM flag has been set, the content of the meter temperature buffer, i.e., an addressable location within the RAM 126, is transferred via the logic circuit 1 90a to be displayed upon the CRT as indicated in step 802. After the display, the subroutine moves to step 806 which examines that location of the RAM 126 where the TP out flag indicating that there was an indication to monitor during a test the prover temperature TP has been set, and 25 the output of the prover temperature buffer is transferred to be displayed upon the CRT. After such display, the subroutine moves to step 810, which accesses the previously set meter pressure PM flag and outputs the content of the pressure buffer, i.e., a known location in the RAM 126 to be displayed upon the CRT. After display, step 814 accesses the previously set PP out buffer flag, and the contents of the prover pressure buffer is transferred to be displayed 30 upon the CRT. After the display of such information, the subroutine returns to the next step 530 as indicated in Fig. 8B. The values output at this time correspond to the values currently stored in the RAM buffers. If a test has not yet been run, zeroes are initially stored in the RAM 126 and will be output and displayed as TM-0.0 Deg. F. This step formats the display so that when a test begins, the system will replace the zeroes with appropriate numbers. The execution of the tests of the meter 38 by the meter prover system 10 as well as the output of the results to the printer and CRT is performed by the routine generally identified in Fig. 5 by the numeral 900, as will now be explained in greater detail with respect to Figs. 9A to 9P. In Fig. 9A, there is shown a high level diagram of the various steps or subroutines that are necessary to effect the test of the meter 38 starting with step 322 as shown above with respect 40 to Fig. 6B to return the piston 14 to its park position, i.e., its bottom- most normal position with respect to the cylinder 12 as shown in Figs. 1 and 2A. Next, routine 902, as will be more fully explained with respect to Fig. 913, determines the type of test, i.e., will the volume measured by the meter 38 be determined by the proximity detector 27, or by the rotary optical encoder 40 as shown in Fig. 4C, and the particular volume that is to be drawn through the meter 38; in 45 particular step 902 loads the appropriate volume factor into the programmable counter 1 74a as well as to store the appropriate divisor in the buffer FORP of the RAM 126. Next, step 904 determines for the condition that one test per flow rate is to be made, the flow rate to be executed for that test; in particular the system addresses that location where the flow rate flags are stored within the RAM 126 to determine the appropriate flow rate to be used in the particular test. If the flow rate is so located, the routine moves to step or routine 906a to determine which flow rate is to be executed and to obtain that flow rate and execute that subroutine as more specific ally shown in Figs. 9C and D. If it is decided in step 908 that two tests for each flow rate are to be conducted, the routine addresses those areas in the RAM 126 for the corresponding flow rate flags and executes the tests at the programmed flow rate in step 55 906b. If three tests per flow rate are called for, it is first determined in step 912 whether the flow rate for the three corresponding rates have been entered, and if yes, step 906c calls the flow rate flags from the designated locations within the RAM 126 and executes the tests of the meter 38 at these flow rate(s). Thereafter, the command is sent to the logic and driver circuit 1 98a to energize the standby indicator before returning to step 216 as shown in Fig. 5. Determining of the test type and volume to be measured is generally shown as subroutine 902 in Fig. 9A, and will be more fully explained with respect to Fig. 9B wherein initially the programmable counter 1 74a and the counter 182, as shown in Fig. 4C, are disabled and reset in step 918. At this point, in step 919, the status of the encoder/ proximity detector switch is determined to decide whether the fluid volume passing through the meter 38 is to be GB 2 142 730A 20 determined by the output of the proximity detector 27 or by the output of the rotary encoder 40. If the output of the rotary encoder is chosen, the subroutine moves to step 920 to test for the presence or status of one cubic foot flag within the designated location of the RAM 126, and if present, that factor, which is equivalent to one cubic foot of output pulses from the output of the optical rotary encoder 40, is stored by step 922 in the RAM EORP, i.e., a designated location of the RAM 126, and the factor number is loaded into the programmable counter 1 74a. At this point, the subroutine moves to return to the next step 904 of the program shown in Fig. 9A. If the one cubic foot flag is not present, step 924 checks the status of the one-half cubic bit or flag and if present as determined by step 924, the programmable counter 1 74a as shown in Fig. 4C is loaded by step 926 with an equivalent count and a corresponding 10 factor is stored in the EORP buffer location of the RAM 126, before returning to step 904 of Fig. 9A. If the one or one-half cubic foot flags are not present, step 928 looks for the presence of the one-fourth cubic foot flag as previously entered by the operator into its system RAM 126, and step 930 loads the programmable counter 1 74a with a count corresponding to one-fourth cubic foot, and a corresponding factor is stored in the EORP buffer location of the RAM 126. If 15 no flag is present corresponding to one cubic foot, one-half cubic foot, or onefourth cubic foot, step 932 automatically determines that a one cubic foot test should be run to test the meter 38. If on the other hand, step 919 determines that the output of proximity detector 27 is to be used to measure the fluid flow through the meter 38, the subroutine moves to step 934 wherein the status of the one cubic foot flag within the RAM 126 is checked and if set, the programmable counter 1 74a of Fig. 4C is set by step 936 with a binary number equivalent to one cubic foot of volume and a factor, normally to 8, is set in the EORP buffer location of RAM 126 before returning to the main program. If the one cubic foot flag has not been set, the subroutine moves to step 938 wherein a check of the status of the one- half cubic foot flag is made and if present, a factor corresponding to one-half of one cubic foot, normally 4, is stored 25 in step 940 in the programmable counter 1 74a and also within the EORP buffer location of the RAM 126, before returning to the main program. If the one-half cubic foot flag has not been set, the subroutine moves to step 942, wherein the status of the one- fourth cubic foot flag is checked and if present, a factor equivalent to one-fourth of a cubic foot, normally 2, is entered by step 944 into the programmable counter 174a and in the EORP buffer location of RAM 126 30 before returning to the main program. If none of the one, one-half, or one-fourth cubic foot flags have been set, the subroutine automatically in step 946 sets the meter prover system 10 to conduct a test drawing one cubic foot of fluid through the meter 38. The next routine 904 and the following routines, as shown in Fig. 9A are a series of steps to determine which of the flow rates are to be conducted and how many times each flow rate is to 35 be tested. Referring now to Fig. 9C, step 948 actuates the logic and driver circuit 1 98a to energize its test in progress light 111 c before moving to step 950 to determine whether the flow rate(s) as stored in the RAM 126 to be conducted is one of GO to G7 and if so, the routine moves to step 952a wherein it is determined whether the test is to be conducted at the GO flow rate and if yes, a speed 0 is loaded in step 954a into the speed buffer location of RAM 126 and 40 thereafter in step 956a, the meter prover system 10 executes the test of the meter 38 at that selected flow rate and outputs the results to the printer, as will be more fully explained with respect to Figs. 9F and F. However, if the test is to be conducted at the Q1 rate, a similar set of steps 954b are conducted whereby the speed 1 is loaded and is executed in these respective steps. Similar sets of steps 952c to 952h, 954c to 954h, and 956c to 956h are conducted 45 whereby corresponding flow rates are entered into the speed buffer and are executed. If the decision in step 950 was no; the routine proceeds immediately to step 958 as shown in Fig. 91), wherein a decision is made by examining the flags disposed in the designated areas of the RAM 126 to determine whether a flow rate of G8 to QF has been selected and if yes, the subroutine continues to step 960a; if not, the subroutine returns to the next step 908 as seen in 50 Fig. 9A. Step 960a determines whether a meter is to be tested at the flow rate corresponding to the flow rate GB and if yes, step 962a loads a speed 8 into the speed buffer of the RAM 126 and at that point executes by step 964a the testing of the meter 38 by the meter prover system at that specific servomotor speed and flow rate. In similar fashion, steps 960b to 960h, 962b to 962h, and 964b to 964h are conducted whereby tests at the various flow rates and 55 corresponding motor speeds are conducted before returning to the next step which may be step 908 or 9 12 to determine the number of tests per flow rate that are to be conducted. In Figs. 9C and D after the flow rate at which a particular meter is to be tested was chosen, the actual test was conducted by steps 956 and 964. The actual meter test conducted by the meter prover system-10 is carried out in the subroutine shown in Figs. 9E and F. First, step 966 60 commands the logic and driver circuit 198a of Fig. 4E to energize the test in progress light 11 '1 c. Next, step 332, as shown in Fig. 6B, returns the piston 14 to its park or lowermost position, as shown in Figs. 1 and 2A. Step 968 commands the logic circuit 194b to effect the opening of the inlet valve 34 and closure of the exhaust v alve 36. As indicated above, a particular speed has been stored in the speed buffer of the RAM 126 and the servomotor 20 is 65 21 GB 2 142 730A 21 driven under the control of the logic circuit 1 94a to that speed in accordance with the subroutine 606 as shown in Figs. BF. Of significance, routine 606 senses that the piston 14 has been accelerated from its park position to the start-test position as indicated by the output of the proximity detector 52, and that the piston velocity has stabilized. Next-in step 970, the outputs as derived from the rotary encoder 40 and the linear optical encoder 26, are inhibited by the signal and logic conditioning circuits 1 70a and 1 70b, respectively, from being entered into their respective counters 1 74a and 182, and the time divisors or factors as originally selected based on the test volume are entered into the programmable counter 1 74a. In step 972, after the piston has reached the start-test position as determined by step 600, the output of the rotary encoder 40 is accessed periodically, e.g., every 40 microseconds, to detect the 10 occurrence of the rising edge of the next pulse therefrom as provided at the input to the interval timer 176. If this has not occurred, the subroutine checks in step 974 the status of the proximity switch 54 and if the piston 14 is at that position, i.e., towards its uppermost position within the cylinder 12, there is an indication of a failure and the subroutine exits via exit point 975 into an interrupt subroutine as will be explained with respect to Fig. 9G. Otherwise, upon 15 occurrence of the leading edge of the output of the rotary encoder transducer 40, the routine moves to step 976, wherein the A output of the rotary transducer 40 is applied by the signal conditioner and logic circuit 1 70a to the interval timer 174 and in particular to its programma ble counter 1 74a. Further, step 976 enables the counter 182 to start counting the pulses derived from the linear encoder 26, and the programmable counter 1 74a to count the pulses 20 from the rotary encoder 40. In addition, step 976 enables at this point in time the counter 1 76a to count the system clock indicated in Figs. 4C as clock A (e.g., 200 KHz) by dividing the system clock by an appropriate factor to output via the logic circuit 178 a one-second sampling signal, whereby the various values of temperature and pressure are sampled. In addition, an output is derived via the interval timer 174 to enable the logic circuit 177, which is otherwise 25 disabled, thus permitting an end of test interrupt to be developed and in particular to apply a one logic or high signal upon the counting down of the inputted count by the programmable counter 1 74a to apply a signal to the Interrupt 2 input of the CPU 120 to initiate an interrupt subroutine which will be explained in detail later with respect to Fig. 9J. Next in step 528, the data. in terms of the pressures and temperatures currently being measured in the meter 38 and 30 in the chamber 28 are outputted as discussed above with respect to Fig. 8P. Next, in step 980, the status of the counter 1 76a within the interval timer 176 is tested, and if no pulses have been accumulated, the subroutine returns to the beginning of step 980 to wait for the occurrence of the first pulse output from the counter 1 76a. If a pulse or pulses have been detected at the output of the counter 1 76a, the subroutine moves to step 982 wherein the 35 counter 1 76a is reset to zero. Thus, the counter 1 76a will be reset to zero upon the occurrence of the one per second pulse output from the logic circuit 178. Next in step 984, the meter pressure MP2 as derived from the outputs of the pressure transducer 46, the prover room pressure derived from the pressure transducer PB, the prover pressure PP1 as derived from pressure transducer 51, the meter temperatures TM3 and TM4 as derived from the temperature 40 transducers 42 and 44, respectively, and the prover temperatures TP1 and TP2 as derived from the temperature transducers 57 and 48, are placed within designated locations within the RAM 126 to be available for the calculations to be performed upon them, as will be explained. Next in step 986, which will be more fully explained with respect to Fig. 9L, the pressure and temperature data is converted to a floating point decimal form and in step 988, as more fully 45 explained with respect to Fig. 9M, the data is accumulated and averaged. As will be explained in detail later, the values of pressure and temperature within the meter 40 and within the chamber 28 are accumulated periodically, e.g., every second, from a point in time when the volume meter test is begun until it is terminated, whereby each of the samples so taken may be summed and divided by the number of samples to time average the temperature and pressure 50 parameters. Thus, the temperature and pressure parameters are considered to be taken or monitored continuously during the test of a fluid flow meter. A final sample of these parameters is taken even after the programmable counter 1 74a has counted down. Next in step 990, a determination is made of whether the test is over by interrogating the INT2 register and if the test is over, the routine moves on to the further steps as shown in Fig. 9F; if not, the subroutine 55 returns to thebeginning of step 528 to output the current values in the pressure and temperature buffers of the RAM 126 and to continue the meter test. The generation and storing of an end of test flag permits the meter test to end anywhere in the interval between the sampling pulses for taking the measurements of pressure and temperature. Thus, as will be explained with respect to Fig. 9J, the end of test flag is set upon the occurrence counting down 60 the meter encoder register 1 74a within a designated portion of the RAM 126 indicating that the test has ended. However, the one-second sampling pulses derived from the logic circuit 178 of Fig. 4C is still to occur in order to obtain the last pieces of temperature and pressure data. Thus, the subroutine as seen in Fig. 9E and in particular step 990 permits the test to end and the final items of pressure and temperature data to be gathered before the meter prover system 10 is 65 22 GB 2 142 730A 22 shut down. The use of the logic circuit 177 to sense the timing out of the meter encoder counter 1 74a permits a rapid and efficient end of test signal or flag to be generated. If software were employed, it would be necessary to repeatedly access the status of the programmable counter 1 74a to determine whether it has counted down, thus considerably complicating the program to be stored within the PROM 124 and reducing the accuracy of the timing of test termination. From the above, it is seen that a meter test is initiated and terminated in response to the output of the meter rotary encoder 40. As explained above, the mechanism coupling the encoder 40 to the diaphragm of the meter 38 generates, in a sense, a frequency modulated signal thus making it desirable to initiate and to terminate the counting of its output at approximately the same point in the cycle of rotation of the rotary encoder 40. To this end, the piston 14 is accelerated from its park position to a substantially constant velocity before passing the start-test position as determined in step 646, at which point the test assesses periodically, e.g., every 40 microseconds, the occurrence of the rising edge of the next A pulse from the rotary meter encoder 40. That next pulse A from the rotary meter encoder 40 initiates the 15 counting down of the programmable counter 1 74a into which has been loaded a count indicative of the test volume to be drawn through the meter 38 under test and at the same time to initiate the counting of the counter 182, which counts the output of the linear encoder 26 to provide an accurate indication of the test volume drawn into the prover 10. Upon the counting down of the counter 1 74a, a signal is applied to the logic circuit 177 to effect a CPU Interrupt 20 2 which as shown in Fig. 9J disables the counters 1 74a and 182, thus terminating the counting of each. Thus, the count as stored within the counter 182 indicative of the pulses derived from the linear encoder 26, is actuated to initiate and to terminate its counting in response to the output of the rotary meter encoder 40, thereby insuring that the count stored in the counter 182 accurately corresponds to a count as stored within the programmable counter 25 174a, which is indicative of the fluid measurement made by the test meter 38. Referring now to Fig. 9F, in step 992 the final values of the aforementioned pressures are obtained after the end of test flag has been set and are converted to a floating point decimal before being accumulated and averaged. The final values of pressure and temperature are outputted via the logic circuit 1 90a of Fig. 4D to be displayed upon the CRT terminal 112 in a 30 manner in accordance with the subroutine as described above with regard to Fig. 8P. Next in step 994, the count accumulated in the counter 182 is converted to a floating point decimal form and in step 620, the servomotor 20 is decelerated to a stop in accordance with the subroutine 620 discussed above with respect to Fig. 8G. Thereafter, step 996 provides command signals via the PI/O circuit 196 and logic circuit 194b to actuate the solenoid associated with the second exhaust valve 36 to open same and to actuate the solenoid associated with the first, inlet valve 34 to close the inlet valve. Next in step 998, the various parameters as have now been finally collected are used to calculate the volumes as indicated by the output of the linear optical encoder 26 and the meter rotary optical encoder 40 in a manner that will be explained in greater detail with respect to Fig. 9N. In step 1000, a command is sent 40 to the logic and driver circuit 1 98a to energize the test complete indicator 111 d and thereafter, in step 1002, the output results as calculated in step 998 are transferred to the hard copy printer as will be more fully explained with respect to Fig. 9P, and the piston 14 is returned to its lowermost normal, or park position by step 322, as more fully shown in Fig. 6B. At this point the program returns to the beginning of step 952a of Fig. 9C to determine whether 45 repeated or different flow rate test are to be made. In Fig. 9G, the interrupt subroutine 975 is shown and is entered primarily when the stop pushbutton 111 g is depressed by the operator, or when in step 974 of Fig. 9B there has been a determination of failure due to the combination of an absence of rotary encoder counts, and the presence of the piston at its upper position corresponding to the placement of the proximity 50 switch 54. Upon pressing the stop button 111 g, the servomotor 20 is quickly decelerated to a halt and the system is "locked up" in its stop mode until the primary power is removed and then reapplied. Upon reapplication of power, the meter prover system 10 will return to its standby mode. Under either condition, the program enters via point 975 as shown in Fig. 9G and in step 1004, causes the logic and driver circuit 1 98a to actuate a light behind the stop 55 pushbutton 111 g. In step 1006, the current motor speed as derived from the logic circuit 1 98b is stored in the speed compare buffer of the RAM 126. Next, the servomotor 20 is decelerated in accordance with the deceleration subroutine 620 as shown in Fig. 8G. Thereafter, step 1008 sends commands to the logic circuit 194b to cause the associated solenoid to open the second exhaust valve 36 and to actuate the associated solenoid to close the first, inlet valve 34. At this 60 point the subroutine goes into a wait step 10 10 until power is removed and then reapplied. As shown in Fig. 9H, an Interrupt 1 routine is entered when the operator presses the reset button 111 g of the system control and status module 111 whereby the program being conducted is interrupted and exits to step 1012, wherein the logic and driver circuit 1 98a is controlled to energize the light behind the reset button 111 g. Next in step 10 14, the current 65 23 GB 2 142 730A 23 speed of the servomotor 20 is read via the logic circuit 1 98b and is stored in the speed compare buffer, i.e., an addressable location of the RAM 126. At that point, the servomotor 20 under the control of the logic circuit 1 94a is decelerated to a stop condition by the subroutine 620 as shown in Fig. EIG, and the return is made to the program via point 212 and in particular 5 to the step 214 as shown in Fig. 5. A third interrupt subroutine, as referred to above with respect to steps 976 and 990 of Fig. 9E, is shown in Fig. 9J as automatically implemented at the end of the meter test when the interval timer 174 outputs a pulse via its logic circuit 177 to the Interrupt 2 input of the CPU 120, this pulse occurring after the down counting of the input of one of the factors corresponding to one, one-half, or one-fourth cubic foot of the test volume to be drawn by the 10 meter prover system 10 through the meter 38. Initially, the status of the system is stored in step 1018 in appropriate locations within the RAM 126 and in step 1020, the counters 1 74a and 182 of Fig. 4C are disabled and the stored count of the linear counter 182 is stored in appropriate locations within the RAM 126, before each of the counters 1 74a and 182 is reset. In step 1024, a one or flag is disposed within the test over register of RAM 126, upon the occurrence of the counting down of one of the aforementioned factors or parameters by the counter 1 74a. Thus, the end of a test may occur midway between the one-second sampling pulses that are used to obtain measurements of pressure and temperature. Thus, it is necessary to wait to obtain that final set of measurements of pressure and temperature and this is done by setting the end of test flag immediately upon the occurrence of the down counting of the aforementioned factors. The meter prover system 10 will continue to operate to accumulate data as indicated in Fig. 9E until step 990 occurs, at which time the test over register of the RAM 126 is accessed to see whether a test over flag has been set and if so, then the test is brought to a halt. If still running, in step 1028, the status of the system is recalled and the meter prover system operation continues at the point of interruption within the meter test routine as shown in 25 Fig. 9E. The routine 984 generally indicated in Fig. 9E for inputting the values of pressure and temperature is more fully shown in the subroutine shown in Fig. 9K, wherein the initial step 1032, an indication of the differential pressure as obtained from the differential pressure measuring device 46 is converted into binary data by the A/D converter 166 and is stored in a 30 designated location within the RAM 126. Similarly, a differential pressure as measured by the piston pressure transducer 51 is converted by the A/D converter 166 into binary data and is stored by step 1034 in a designated location of the RAM 126. Similarly, the barometric pressure as measured by the transducer 109 disposed in the prover room 106 in which the meter prover system 10 is housed, is converted by the A/D converter 166 into binary data and 35 is stored by step 1036 in a designated location in the RAM 126. Similarly in steps 1038, 1040, 1042, the temperature inputs TP1, TP2, TM3, and TM4 from the temperature measuring devices 57, 48, 42, and 44, are converted to binary data by the A/D converter 158 and are stored within designated locations at the RAM 126, before returning to step 986 of Fig. 9E. Step 986 as generally shown in Fig. 9E is more fully shown in the subroutine of Fig. 9L where in step 1046, each of the temperatures corresponding to the prover temperatures TP1 and TP2, and the meter temperatures TM3 and TM4 is converted from a binary voltage to degrees Fahrenheit in a floating decimal format and then is stored in a designated location within the RAM 126. Next in step 1048, in similar fashion, the binary voltages indicative of the 45 pressures MP1, MP2, and the barometric pressure PB are taken from their designated locations within the RAM 126 and are converted to pounds per square inch (psi) in a floating point decimal format and are restored in designated locations in the RAM 126, before returning to step 988 as shown in Fig. 9E. The subroutine or step 988 as shown in Fig. 9E for accumulating and averaging the test data 50 is more fully shown with respect to the subroutine shown in Fig. 9M. In step 1050, the digital values of the first and second prover temperatures TP1 and TP2 are added together to provide an Average Prover Temperature (ATP). This value is in turn added to the accumulated previous values of ATP, thus accumulating a series of N samples of the Average Prover Temperature during the conducting of a meter test. In step 1052, the meter temperatures TM3 and TM4 are 55 added together to obtain an Average Meter Temperature (ATM) and this new value is added to the previously accumulated values of the Average Meter Temperature ATM to obtain a quantity indicative of N consecutive samples of the ATM taken during the course of a meter test. In step 1054, a value of the ambient or barometric pressure PB is added to the sum of the barometric pressure PB plus the differential pressure MP2 taken between the ambient pressure and the pressure established within the meter 38 to obtain an Average Meter Pressure (APM). The Average Meter Pressure is added to the previously accumulated values of the Average Meter Pressure to obtain a quantity indicative of N consecutive samples of the APM taken during the meter test. In step 1056, a value of the ambient or barometric pressure PB is added to that differential pressure as obtained from the differential pressure PP1 transducer 51 to obtain an 65 24 GB 2 142 730A 24 Average Prover Pressure (APP). The Average Prover Pressure (APP) is added to the previously accumulated values of the Average Prover Pressure to obtain a quantity indicative of N consecutive samplings of the APP taken during the meter test. In step 1058, the number of samples N of each of the values taken above in steps 1050, 1052, 1054 and 1056 is determined, whereby in succeeding steps an average value of each of these parameters may be obtained by dividing the accumulated quantity by the factor N. Step 998, as shown in Fig. 9F, will now be explained in more detail with respect to Fig. 9N. As indicated above, the accumulated values of average pressure temperature (ATP), average meter temperature (ATM), average meter pressure (AMP), and average prover pressure (APP) are indicative of N samples taken during the course of the meter test and in step 1060 these average values are obtained for each of the temperatures and pressures. Next, in step 1062, the percent error of the output of the meter encoder 40 of the fluid passing through the meter 38 with respect to the standard volume passing therethrough as drawn by the meter prover system is calculated in accordance with the indicated equation. The values of meter pressure PM, prover pressure PP, prover temperature TP, and meter temperature TM are those averaged values as determined from step 1060, whereas the prover volume count K was obtained as an indication of the accumulated counts that was produced by the linear encoder 26 and accumulated within the counter 182, these counts being indicative of a selected standard volume, e.g., one cubic foot, one-half cubic foot, or one-fourth cubic foot. As shown in Fig. 9N, the ratios of PM to PP, and TP to TIVI provide, respectively, pressure and temperature correction 20 factors. As explained above, there exist differences of fluid temperature and fluid pressure between the fluid within the meter prover 10 and the fluid meter 38 under test. The noted ratios as determined by step 1062 provide appropriate correction factors whereby the differ ences in temperature are compensated for to give an accurate indication of percent error of the meter reading of fluid volume with respect to that calibrated indication or reading provided by 25 the meter prover 10. The factor K is representative of those counts as stored in the programmable counter 1 74a indicative of the fluid flow as measured by the meter 38. As shown in step 1062, the factor K is multiplied by a constant C to convert the count, e.g., 40,000 for one cubic foot, to a value of the volume to be drawn through the meter 40. In the illustrative example, where the factor K as entered into the programmable counter 1 74a is 40,000, the constant C is chosen as 1 40,000 Further, the number of counts as derived from the linear encoder 26 and counted by the counter 182 is multiplied by a calibration factor which, as will be explained later in detail, is derived by accurately measuring the volume of the chamber 28 of the meter prover 12 and correlating the precisely measured volume to the series of pulses outputted by the linear encoder 40 26. In an illustrative example of this invention, for a selected volume of one cubic foot of the chamber 28, it was determined that 38,790 output pulses were derived from the linear encoder 26. Thus, the calibration factor would be selected as inverse of this count or 1 38,790 In similar fashion, if a different volume was to be drawn through the meter 38 and the programmable counter 1 74a was to be programmed with a count corresponding to a different 50 flow, it would be necessary to determine a different calibration factor corresponding to that volume as defined within a selected region of the chamber 28. In practice, the count of 38,790 is obtained empirically by moving the piston 14 along a length of the chamber 28 correspond ing to precisely one cubic foot and calculating or otherwise determining the number of pulses as outputted by the linear encoder 26. Similar measurements and encoding will be taken empirically to determine the counts for one-half and one- fourth cubic feet flow corresponding to the output of the linear encoder 26 as the piston 14 is moved distances from the start-test position along the cylinder 12 corresponding illustratively to one-half cubic foot and to onefourth cubic foot volumes within the cylinder. Thus, the actual counts and therefore the corresponding calibration factors for one-half and one-fourth cubic feet will correspond precisely 60 to these volumes as defined by the cylindrical wall of chamber 28. In this manner, a calculation of the percent error between the reading of the meter 38 and the actual volume as drawn into the chamber 28 can be made with great accuracy, by the use of calibration factors that have been determined. in accordance with the precise volume of the corresponding regions of the cylindrical wall of the chamber 28. After the step 1062 has been executed, the program returns65 GB 2 142 730A 25 to the next step 1000 as shown in Fig. 9 F. The calculation of the average temperatures and pressures as generally disclosed in step or subroutine 1060 of Fig. 9N is more completely shown by the subroutine as set out in Fig. 9Q, as previously indicated in step 1050, each measurement of temperature is added to the previous measurement to obtain a total indicative of accumulated discrete measurements of prover temperature ATP during the course of a single test. As indicated in step 1064 as shown in Fig. 9Q, this accumulated value of ATP is divided by the number of samples AVCTR = N + 2 and added to 459.67 to provide the final averaged temperature FTP of the prover in degrees Kelvin. In similar fashion in step 1066, the value of the temperatures TM3 and TM4 are added successively to each other during the course of the meter test to provide an accumulated value 10 of the average temperature ATM of the meter during the course of the test, and the ATM is divided by the number of samples or counts of the one- second sampling clock which is in turn added to 459.67 to give a final average temperature FTM in terms of degrees Kelvin. In similar fashion, in step 1068, the accumulated values of the average pressure APIVI of the meter as accumulated during the test is divided by the number AVCTR of samples taken during the test 15 to provide a final FPM of the meter in terms of pounds per square inch absolute. In step 1070, the accumulated values of the average prover pressure APP as accumulated during the course of a meter test are divided by the number AVCTR of samples to provide an indication of the final pressure of the prover FPP in terms of pounds per square inch absolute, before returning to the next step 1062 as shown in Fig. 9N. The accuracy of the meter prover 10 to measure the volume drawn through the meter 38 is dependent and is limited by the variations in the measurements of fluid pressure and temperature during a meter volume test. The variations in fluid pressure and temperature as occur during a meter volume test run are greater than those variations that may occur in the measurement and calibration of the volume of the meter prover. Thus in general, the actual volume of fluid delivered through the meter 38 as the piston 14 is moved from a first to second position may be approximated by the following expression: V'(Vl-V2) 11 + (3T /T1)-(31P/ P,)]+ VA(AT/T1)-(AP/PJ] 30 where V, P, and T, are respectively the initial volume, fluid pressure and temperature existing within the chamber 28 of the prover 10 at the beginning of a meter volume test run, AP and AT are the corresponding changes in pressure and temperature that occur during the test, and ST and P are the mean deviations from T, and P, at the fluid meter 38. As explained above, the logic circuit 178 as shown in Fig. 4C provides a one-second clock signal that samples the 35 output of the temperature transducers 42, 44, 48, and 57, as well as the pressure transducers 46 and 51 to provide samples of these variables throughout the volume meter test run. As discussed above, the periodic samples of the meter temperature TM3 and TM4, and the prover temperature TP1 and TP2 as obtained in step 984 of Fig. 9E are summed in steps 1050, 1052, 1054, and 1056 of Fig. 9M together to obtain a spatially averaged meter temperature 40 and prover temperature, which are time averaged by being sampled and summed over the test run as explained with respect to steps 1064, 1066, 1068, and 1070 of Fig. 9Q. In this manner, the continuous changes in both pressure and temperature are measured at the fluid meter and the prover 10. By so averaging in a spatial and time sense, the variations in temperature and pressure that would otherwise cause an error in the calculation of the change 45 of volume may be avoided. Temperature changes in the fluid as drawn into the meter prover as caused by adiabatic expansion across a pressure drop of 1.5"H,O can be as large as 0.5F, causing an error in the calculation by as much as 0. 1 %. Thus, it is necessary to monitor accurately the fluid pressure and temperature throughout the system and apply the necessary corrections in order to obtain the "true" volume drawn into the meter prover 10. For an accuracy of 0.05% in volume, the parameters ST and SP between the chamber 28 and the meter inlet need to be monitored to an accuracy of better than 0.1 F and 0.05111-1201 respectively. The temperature transducers and pressure transducers described above with respect to Fig. 1 have such a capability to achieve the desired accuracy for temperature and pressure measurement and provide the desired accuracy in the measurement of the volume. Step 1002, as shown in Fig. 9F as outputting the results of the conducted tests upon the meter 38 to the hard copy printer, is more fully shown in Fig. 9P. In the first step 1072, the printer is strobed by the logic 1 90b to initialise its operation and thereafter the percent error as calculated previously in step 1062 is printed out upon the printer. Next, in step 1074, a line feed command is applied for the printer to advance its paper by one line, before in step 1076, 60 the flow rate at which the test was conducted is displayed in step 1076. In step 1078, ten line feed command signals are applied to the printer to feed the paper ten lines before returning to step 322 in Fig. 9P. An important aspect of the proving process is the method of accurately measuring the volume of the chamber 28 of the cylinder 12 and to use that measurement to accurately calibrate the 65 26 GB 2 142 730A 26 output of the linear encoder 26 to provide a manifestation of volume per pulse output of the linear encoder 26. As indicated in Fig. 2A, the linear encoder 26 includes a scale 24 with a high number of accurately, closely spaced markings 102, whereby the optical encoder 26 provides a train of output pulses corresponding to the passage of each such marking past the optical encoder. By accurately correlating the number of such output signals to the precisely determined volume of the chamber 28, the linear encoder 26 may be encoded with a corresponding high degree of accuracy to thereby improve the calibrated indication of meter registration or error of the meter 38 under test. In such a calibration procedure, it is first necessary to measure the volume of the chamber 28 with great accuracy. A method of volume measurement is employed including the establishing 10 of a high frequency, electromagnetic field within the cavity; the principles of such measurement method will now be explained. The inside dimensions of the regularly shaped chamber 28 completely surrounded by conducting walls can be accurately determined by measuring the frequencies at which resonant conditions occur for the electromagnetic field established within the chamber 28. For a given geometry, the electromagnetic fields within the chamber 28 can 15 assume a variety of spatial configurations. At discrete frequencies, the electromagnetic energy confined within the chamber 28 is stored (resonates) over time intervals long compared with its wave period and these resonant solutions are designated as the normal modes of the chamber 28. The ratio of the energy stored to that dissipated per cycle of the resonant frequency is defined as the quality factor or "G" of the resonance. The quality factor Q is a measure of the 20 dissipative losses due to the ohmic resistance of the walls to the electrical currents induced by the electromagnetic fields. For a given normal mode, the resonant frequency is uniquely determined by the dimensions of the chamber 28 and the propagation velocity of light in the medium filling the volume. Consequently, by simultaneously measuring the resonant frequencies of a number of normal modes equalling the linear dimensions specifying the volume (one for a 25 sphere, two for a right circular cylinder, and so forth), the volume of the chamber 28 can be determined to an accuracy comparable to the precision of the frequency measurements. For a given chamber geometry, there is an infinite set of normal modes whose resonant frequencies will have a lower bound value corresponding to a free-space wavelength of the order of the linear dimensions of the chamber 28, but no upper bound. A volume change caused by a 30 mechanical displacement of one of its linear dimensions will, in turn, change the resonant frequency of the low ordered modes by approximately the same fraction. Briefly, as shown in Fig. 10, means in the form of the antenna 70 is provided to establish or generate an electromagnetic field within the chamber 28 and to extract therefrom a relatively small portion of the stored electromagnetic energy to be measured by the circuit of Fig. 10. As 35 will be explained in detail later, measurements are made to determine the frequencies f at which resonance occurs within chamber 28. The selection of the configuration of the chamber 28 is significant for if an arbitrarily shaped chamber is used, the mathematical relationship between the resonant frequencies f and the dimensions of such a chamber may have no closed or analytical solution. Thus, the chamber 28 is selected to be of a regular geometry to assure that 40 the dissipated losses and the effects of coupling are negligible and to provide known Maxwell equations defining the relationship between the chamber dimensions and the resonant frequen cies subject to the boundary conditions that the electrical field E be radial and the magnetic field H be tangential to the totally enclosed surfaces of the regularly shaped chamber 28. Under these conditions, the field solutions for the chamber 28 will take on a relatively simple form and 45 recognizable perturbations can be predicted. Thus, in a preferred embodiment of this invention, the chamber 28 is selected to be a right circular cylinder as is formed by the inner surface of the cylinder 12, the head 60, and the exposed surface of the piston 14. It is recognized that a chamber 28 of such a configuration permits the piston to be driven therethrough to displace a known fluid volume, as explained 50 above. The normal modes of the electromagnetic field established in the chamber 28 being configured as a right circular cylinder are divided into two general classes: the transverse electric modes (TE) for which the electric field is zero along the cylindrical (z) axis of the chamber 28, and the transverse magnetic modes (TM) for which the magnetic field is zero along its cylindrical axis. They are further specified by three integers 1, m, and n, which are defined for the TE 55 modes in terms of the cylindrical coordinates, r, 0, and z by: I = number of full-period variation of Er with respect to 0; m = number of half-period variation of E,9 with respect to r; n = number of half-period variation of Er with respect to z. A similar set of indices exist for the TM modes for which the integers 1, m, n are correspondingly defined in terms of the components of the magnetic field; Hr and H.. The solutions for the resonant frequencies of the normal modes are expressed in terms of the geometrical dimensions and roots of Bessel functions by the general expression: 27 GB 2 142 730A 27 XIM n D f = (c/ D) [( -)2 + (_2_)2( L)2]1/2 (1) 17 where D is the diameter, L is the length, c is the speed of light in the medium filling the cavity 5 volume and X,, are given respectively by: XIM = Mth root of X,(x) = 0 for the TE-modes, x, = mth root of J,(x) = 0 for the TIVI-modes. Numerical values for these Bessel roots corresponding to the various lower ordered TE and TM modes are taken from the following table: Transverse Electric Modes (TE) Transverse Magnetic Modes (TIVI) TE X,. TM X]. 11 1.84118 01 2.40483 21 3.05424 11 3.83171 01 3.83171 21 5.13562 31 4.21009 02 5.52008 20 41 5.31755 31 6.38016 12 5.33144 12 7.01559 A plot of the quantity (fD)2 VS (D / L)2 of equation 1 is a straight line with intercept (CX,./,T)2 and slope (cn/2)2. Such a mode chart is shown in Fig. 13 for the lower modes of the right cylindrical chamber 28 for values of n up to 2. Fig. 13 shows graphically the relative resonant frequency values as a function of the geometric parameter (D/L)2 and the variation in resonant frequencies as the linear dimension L is changed. It is also useful in predicting the number of resonances expected to be encountered in a given frequency range for any fixed D/L ratio as well as the values of L at which two different modes are degenerate in frequency, and hence would interfere with each other. The mode chart of Fig. 13 is used to determine the modes of electromagnetic field excitation, the expected resonant frequencies of the selected modes, and the dimension in terms of the diameter D and length L of the chamber 28. In an illustrative embodiment of the chamber as shown in Fig. 10, the diameter D is set equal to 12" and the variation in the length L is from 1011 to 30". Consequently, the corresponding values of (D/L)2 have the range of values from 1.44 to 0. 16. If the frequency range of observation is between 500 MHz to 1000 MHz, then (fD)2 Will vary between 2.32 X 1020 and 9.29 X 1020 (cycles/seC)2CM2. The number of resonant modes expected to be encountered over the variation in length L for this configuration between 40 the range of 500 to 1000 MHz is then just given by the number of lines included in the rectangle shown on the mode chart of Fig. 13. As can be seen, at L = 10", there are only three modes which occur between 500 to 1000 MHz. These are, namely, the TM,,, mode occurring at approximately 755 MHz, the TE,, mode occurring at 830 MHz and the TM01' mode occurring at 960 MHz. On the other hand, for L = 30", there are eight modes, namely TE,, TE,12, TMO,O, TM0,, TMO12, TE21, TE2,2, and TE,1, which for purposes of clarity, are not shown in Fig. 13. At any intermediate position, the number of modes and their resonant frequencies f can be obtained from the intersection of the vertical line corresponding to the desired (D/L)2 value with the lines designating the various modes. Since the measurements of the resonant frequency of any two distinct modes uniquely specify both D and L, the other modes may be used either as a redundancy check on the measurements or as a means for evaluating the effects of perturbation caused by deviation of the geometry from the idealized case as well as any higher ordered correction terms which may be present. In a qualitative sense, it is understood that the electric and magnetic field components of the electromagnetic field are perpendicular to each other and have a defined relation at a resonant 55 frequency to the dimensions D and L of the chamber 28 of a regular geometry and in particular the diameter and length of the chamber 28, as shown in Fig. 10. As will be explained, the relationship between the resonant frequency and the dimensions D and L for a particular mode can be related by a mathematical expression. The chamber 28 having a right circular cylinder configuration has two unknown dimensions that define its volume, i.e., its diameter D and its length L. Thus, it is necessary to provide two mathematical expressions that may be solved simultaneously for the unknown values D and L, and therefore it is necessary, as indicated above, to establish electromagnetic fields within the chamber 28 of two distinct modes and obtain the resonant frequencies of these two modes whereby the values of D and L and thus the volume of the chamber 28 may be calculated. As will be explained, the modes of excitation are 65 28 GB 2 142 730A 28 selected to determine resonant frequencies with correspondingly high quality factors whereby the effects of perturbations to introduce errors into the measurement of D and L may be minimized. The TMO,O mode line has zero slope, as shown in Fig. 13, and is therefore independent of the dimension L of the chamber and is only a function of the diameter. This unique property can therefore be used to identify this mode in experimental measurements. The TMO,O mode of excitation can be used to obtain an independent measurement of the dimension D. Furthermore, it can be seen that the rate of change of the resonant frequency for any given mode as a function of the dimension L is solely determined by the last index n. Consequently, the frequencies of the modes such as TIE,, TMO,,, and TE2,1 will shift by the same amount as L is 10 varied whereas the TE,12, TMO121 and TE2,2 modes will all shift together at twice the rate. Consequently, the tuning properties of these modes can be used to determine the relative change in the length L to very high accuracy once the absolute diameter of the cavity has been determined. The quality factor or Q of the chamber 28 at resonance as excited in a given normal mode is important in two respects. First, it determines the sharpness of the resonant frequency response and therefore limits the ultimate accuracy with which the resonant frequency can be measured experimentally. More important, the quality factor Q is a measure of the order of magnitude of the expected deviation of the resonant frequency from the idealized results as given by equation 1. In general, the higher the Q for a given mode, the more accurate the theoretical expressions 20 are in terms of determining the geometrical dimensions of the cavity. Thus, the modes for exciting the chamber 28 are selected to provide the highest quality factors Q, and thus provide the more accurate determination of the dimension and thus the volume of chamber 28. For a volume comprised of perfectly conducting walls, the unloaded or intrinsic 0, of the chamber is infinite and the frequency solutions to the resonant normal modes are exact. For walls of finite resistivity, the electric and magnetic fields within the chamber 28 penetrate into the walls to a distance defined as the skin depth 8, which is given by the expression 8 = [ApI 1 20g2g]l /2CM (2) where it is the permeability of the wall material, X is the free-space wavelength in cm, and p is the resistivity (d.c.) of the walls in ohm-cm. Since a finite skin depth makes the apparent dimensions of the chamber as seen by the electromagnetic fields somewhat larger than the actual geometrical dimensions, the dissipative effect caused by the ohmic losses on the walls perturbs the resonant frequency of the normal modes and shifts them to a lower value. A 35 measure of this perturbation is the value of the ratio of the skin depth to the free-space wavelength or (S/X), which can be calculated via equation 2 by using known values of the d.c. resistivity and permeability of the wall material. For a cavity made of copper (g = 1, p= 1.72 X 10-6 ohm-cm), the skin depth is equal to 3.8 X 10-5A1/2cm. At 1000 MHz (,X= 30 cm), the ratio (S/X) is approximately equal to 7 X 10-6. This correction to the resonant frequency caused by the finite conductivity is clearly negligible for the present application. For series 300 stainless steel, (g = 1, p = 72 X 10-6 ohm-cm), the raio of (S/X) at 1000 MHz is 4.5 X 105 which is becoming non-negligible compared to the desired overall accuracy of 0.01 % in absolute volume accuracy. In a preferred embodiment, the walls of the chamber 28 are chrome-plated, (g = 1, p = 13 X 10 - 6 ohm-cm), the corresponding value of (S/X) is equal to 45 2 x 10-1 at 1000 MHz. Consequently, on the base of theoretical calculations, the correction to the resonant frequency of the normal modes in the chamber 28 of a right circular cylinder configuration should be negligible and the expression given by equation 1 should be adequate in giving an overall volume determination to an accuracy better than 0.0 1 %. In practice, however, the theoretical skin depth value for a given material is never achieved 50 due to a variety of reasons, including wall imperfections, material impurities, and residual surface contaminations. Consequently, it is necessary to experimentally determine the effective skin depth by measuring the G of the chamber 28 for selected of the preferred modes of excitation, and if necessary, apply the appropriate correction to the volume determination due to this effect. The expression Q. of the chamber 28 is obtained by evaluating the ratio of the energy stored to the dissipative losses occurring on the walls per cycle of the electromagnetic oscillation, which can be written as, 21 VH2dy _ 60 a % 6J 5 H 2 ds 9 (3) 29 GB 2 142 730A 29 where H is the normal mode magnetic field vector and 8 is the skin depth. The G factors in terms of the normal modes and geometrical shape for the chamber 28 of a right cylindrical cavity are given by the following expressions: For TE modes: a 11-(, lxt m) 2 1 [X 2 + p2R2 1 312 Lm Q01. a 2W [X2 Am + P2R3 + (1-R)_( AiLM - 73 ' (4) 10 For TM modes: EX2 + P2R23112 a LM Qi a 5(1+R) for 00 (5) a XIM, for n - 0 20 (z + it) where R = (D/L), p = nir/2 and all other symbols are as previously defined. By evaluating the right hand sides of the above expressions for a given mode and cavity 25 geometry and dividing it by the experimentally determined Q, an effective 8/X can be obtained. A direct comparison of the result obtained with the calculated via equation 2 then gives a quantitative evaluation of the magnitude of the perturbation expected in the volume determina tion of the chamber 28. The chamber 28 is coupled to an external measurement circuit as shown in Fig. 10 whereby 30 microwave power is introduced into the chamber 28 to establish an electro- magnetic field therein and to extract reflected power therefrom. The chamber 28 is a reflection type cavity requiring only one coupling device in the form of the antenna 70. As will be explained in detail later, the resonant characteristic, and in particular the resonant frequency of a given normal mode, is determined by measuring the power reflected from the antenna 70 as a function of the 35 incident microwave frequency. The use of a reflected type chamber 28 permits minimal perturbation from an idealised cavity response and permits an implementation by the use of a directional coupler 1106 which allow sampling of the reflected power without interference from the incident power. As shown in Fig. 10, the power output of a sweep generator 1100 is applied to excite the antenna 70 while the reflected power may be transferred by the coupler 40 1106 to a crystal detector 1110. The presence of the coupling device in the form of the antenna 70 introduces an additional loss term in the dissipation of the resonant electromagnetic energy stored within the chamber 29, namely, the amount of power extracted for measurement. This is usually defined as an equivalent coupling Q or Q,, as distinguished from the -unloaded- Q of the cavity or Q.. In addition, the interaction of the cavity with the rest of the circuits of Fig. 10 via the antenna 70 need to be taken into account in order to derive an accurate description of the -cavity- coupling system for evaluating the appropriate expressions in the present application of the invention. It can be shown that for a reflection-type chamber 28, i.e., a chamber for which power is introduced and extracted by a single antenna 70, the ratio of the reflected power to the incident 50 power near a resonance is given by the expression, 114( 1 - 1) 2 + (v - v)21v 2 QO QC 0 0 - (6) 55 (P /P) - 1 1 2 2 2 r o 1/4(t + + (V0 - v) 1V0 0 c where P is the frequency, x,, is the resonant frequency, Q. is the unloaded Q of the cavity 60 resonance, and Q,, is the coupling Q which is proportional to the power loss through the coupling device. At the resonant frequency z,,,, the ratio of the reflected power to the incident power is defined as the reflection coefficient for the given normal mode, i.e. GB 2 142 730A 30 (P,/ po) = p a = p = (Q._QJ2/(Q C + QJ2 (7) For P = 0, the cavity is considered to be 100% coupled and Q. = Qc. This condition corresponds to the low-frequency equivalent circuit case where the load impedance is matched to the 5 generator impedance. Combining equations 6 and 7, and eliminating Qc, the unloaded Q. of the normal mode is given in terms of the measurable parameters by, Q. = p 0 [(P,/ PO)-P/ 1 _(PrI P0)]l /2/(p-p 0) (1 + fil/2) (8) Therefore, by measuring 8 and the frequency width at some arbitrary power level (PjP.) on the cavity resonant response curve, the quantity Q. can be determined. If a -half-power- point is defined such that then the frequency difference corresponding to the two half-power points is known as the 15 -half-widthof the resonant response and is given by P1120 +,8)/2, (9) Ap = 2(P,12-Vo) where V1/2 is the frequency corresponding to the half-power points. For this case, equation 8 20 further reduces to, Q. = + P1/2). (10) The value of Q., therefore, varies between (P./Ap) and (2vJAP), depending on the degree of 25 coupling. A pictorial representation of the chamber response curve, together with the defined para meters, is shown in Fig. 14. The center frequency of the response, v., corresponding to a minimum of the reflected power, is the resonant frequency of the normal mode. It is understood that the frequency measuring circuit as shown in Fig. 10 effects a change of 30 the resonant frequency, i.e., frequency pulling, caused by the interaction between the chamber 28 with the elements of the circuit of Fig. 10. In the circuit as shown in Fig. 10, wherein the frequency of the sweep generator 1100 is varied by a factor of approximately 2, the directional coupler 1106 receives the output power of the sweep generator 1100 and applies a relatively small portion thereof to the antenna 70. In particular, the power applied to the auxiliary or output port of the directional coupler 1106 is approximately 80% of the output of the sweep generator 1100. The chamber 28 is connected to the main input port, and the crystal detector 1110 is connected to the third port of the directional coupler 1106 to measure the microwave power reflected from the chamber 28. The use of the directional coupler 1106 results in a factor of 100 attentuation or "padding" between the sweep generator 1100 and the load as imposed 40 by the antenna 70. As a result, the sweep generator 1100 is respectively isolated from the antenna 70, i.e., the coupling system, to insure that no interaction occurs to perturb the chamber response. As a result of the isolation imposed by the coupler 1106 between the antenna 70 and the generator 1100, the amount of frequency pulling is determined by the quality factor Q of the chamber 28, the coupling coefficient 8, and the VSWR between the crystal detector 1110 and the cavity coupling system. Assuming that the sweep generator 1100 is totally decoupled from the cavity 28, a cavity 28 having a quality factor a of 5,000, a coupling coefficient,8 of.5 and a residual VSWR of 2, will provide a frequency pulling in the order of 1.3 X 10-5 or.0013%. This deviation is an order of magnitude between than required to achieve desired overall measurement of the chamber volume to an accuracy of better than 50 01%. Further, an analysis of the cavity 28 and its associated resonant frequency measuring system, as shown in Fig. 10, have shown that perturbations due to distortions in the geometry of the chamber 28 whether due to the out of roundness of the cylindrical shape of the cavity 28 or to small localized surface irregularities and deformations, indicates that such perturbations can be 55 compensated for by taking the following precautions. First, if the machine tolerance of the chamber 28 is such that the diameter D of its cylindrical configuration is maintained to within the limits of 12 mills out of roundness, then the diameter D and thus the volume of the chamber 28 can be determined with an accuracy in the order of 10-5 or less and, thus, such deformation may be neglected for the present method of measuring volume. Similarly, the shift 60 in the chamber resonant frequencies for any normal mode caused by a small inward or outward dent on the interior wall of the chamber 28 is deemed negligible if the hole dimension is well below the cut-off wavelength of the electromagnetic field established within the chamber 28 and the hole does not couple electromagnetic field to another structure; under these conditions, the frequency pulling will be approximately proportional to the ratio of the cube of the hole diameter 65 31 GB 2 142 730A 31 to the volume of the chamber 28. Thus, as a practical matter, the frequency pulling for inward or outward dents in the walls of the chamber 28 is negligible; for example, a 1 inch diameter hole as placed within a chamber 28 having a diameter of approximately 12 inches and length of inches will provide a change of the resonant frequency by only 7 parts in 105. Considering the effect of the inlet 62 as well as the openings to receive the transducers 51 5 and 57 within the piston 14, these openings or holes may be filled with metallic plugs during the volume measurement and calibration to virtually eliminate these sources of error from effecting the determination of the resonant frequency. In order to satisfy the accuracy requirement of .01 % in the measurement of the displacement volume of the chamber 28, the resonant frequency measuring circuit as shown in 10 Fig. 10 should be capable of measuring the resonant frequencies of the normal modes as established within the chamber 28 to 1 part in 105. The circuit, as shown in Fig. 10, is designed to reduce systematic error which can effect the measurement of the resonant frequencies f, such distortions due to impedance mismatch between the sweep generator 1100 and the antenna 70, variations and fluctuations of the microwave power, extraneous noise, and 15 sensitivity to component drifts. As shown in Fig. 10, the microwave power source takes the form of the sweep generator 1100 which may illustratively take the form of that generator manufactured by Texscan under their model designation VS80A. The sweep generator 1100 may be operated illustratively at a fixed frequency (CW) or automatically to sweep through a range between 50 KHz and 300 MHz at a rate set between.05 Hz and 30 KHz. In addition, the 20 output of the sweep generator 1100 may be controlled by its vernier knob 11 00a to sweep at an operator control change of frequency through the noted range. The output of the sweep generator 1100 is applied by a coaxial cable to a directional coupler 1102 which acts as a transformer whereby a portion of the energy applied across the coupler 1102 is transferred to a frequency counter 1108, which in an illustrative embodiment of this invention may take the form of that counter manufactured by Fluke Corporation under their model designation No.

1 920A. As will be explained later, the counter 1108 displays the frequency at which a standing wave is established within the chamber 28. In turn, the output of the coupler 1102 is applied by a similar coaxial cable to a second coupler 1104 and in turn via the directional coupler 1106 to be applied to the microwave antenna 70. As shown in Fig. 10 and in more detail in Fig. 2A, 30 the microwave antenna 70 is a simple metallic loop 70a and is insulated by an insulator 70b from the head 60 of the chamber 28.

As is well known in the art, the energy reflected through the antenna 70 from the chamber 28 upon occurrence of a standing wave decreases significantly in comparison to that enery reflected at other frequencies. This is known as the resonance condition, and the associated frequency as the resonant frequency. Thus, as the frequency of the output of the sweep generator 1100 is varied, a resonant frequency is selected at which a standing wave occurs within the chamber 28 dependent upon the configuration and dimensions of the chamber 28.

The frequency at which the standing wave is established determines, as will be explained, the chamber dimensions in terms of the diameter D and the length L, and therefore the volume of 40 the chamber 28. To detect the power drop at the resonant frequency, the coupler 1106 is connected to the crystal detector 1110, which converts the microwave power reflected through the antenna 70 to a d.c. signal. In turn, the crystal detector 1110, which may illustratively take the form of a Hewlett Packard crystal detector Model No. 423A (NEG), applies its d.c. output to the Y input of an oscilloscope 1112. The oscilloscope 1112 may illustratively take the form of a 45 Textronix oscilloscope manufactured under their Model No. T922R. The X input to the oscilloscope 1112 is provided by the sweep generator 1100, so that when the generator 1100 is set in the sweep mode, the reflected power response of the chamber 28 is a function of the input signal frequency and is displayed upon the oscilloscope 1112. As seen in the expanded 50 display 111 2a, the power reflected through the antenna 70 dips to a minimum 1113, at the resonance frequency in a manner shown in Fig. 14. The frequency at which the minimum 1113 occurs is displayed upon the counter 1108. The second coupler 1104 applies microwave power to a crystal detector 1114 which provides a corresponding d.c. signal to be amplified by an operational amplifier 1115 and applied to the sweep generator 1100 to provide a level control upon the output of the sweep generator 1100, whereby substantially an even power drain is placed upon the sweep generator 1100 as it sweeps through that frequency at which a standing wave occurs.

The circuit of Fig. 10 is operated in the following fashion to obtain a measurement of the resonant frequency. First, the sweep generator 1100 is set for wide sweep to permit essentially all of the normal mode resonances to be displayed simultaneously on the screen of the oscilloscope 1112 whereas the resonant response for any particular mode can be displayed individually by an appropriate choice of sweep width and sweep center frequency. Unambiguous identification of the modes can be made by measuring their resonant frequencies for a particular setting of the piston position within the chamber 28, and using equation 1, or alternatively, by moving the piston position and comparing the rate of change of their resonant frequencies a 32 GB 2 142 730A 32 function of piston position with those shown in Fig. 13.

The resonant frequency f of any given mode is measured by first displaying the response curve on the oscilloscope screen and then switching the sweep generator 1100 to its CW mode and manually tuning the fine frequency knob 11 00a until the voltage displayed on the oscilloscope 1112 is a minimum. The display on the frequency counter 1108 when this minimum is reached is then the resonant frequency of the cavity normal mode. As a convenience, the second beam of the oscilloscope 1112 can be used to better define the position of this minimum by setting the system in the sweep mode and manually changing the vertical position of the second beam so that it just touches the bottom of the resonant response curve. When the generator 1100 is switched to operate in its CW mode, the resonant frequency10 f of the chamber 28 is then the frequency setting of the generator 1100 corresponding to condition when the two beams coincide. In a preferred embodiment, by simultaneously reading the frequency counter output while the two beams are exactly coincident, the effect due to frequency drifts of the generator 1100 is eliminated so that the measurement can be made to a much higher degree of accuracy than the inherent stability of the sweep generator 1100. In addition, since the measurement depends only on establishing the minimum in the chamber response curve, it is independent of non-linearity in the response of the detector 1100 as well as fluctuations in the incident microwave power with time. Repeated measurements on a given mode indicate that resonant frequency of the chamber 28 can be determined to an accuracy better than 3 KHz or approximately 5 parts in 106.

Now, a first embodiment of the method for precisely measuring the volume of a section of the chamber 28 will be explained in greater detail. Generally, the method of this invention involves measuring the resonant frequencies of the chamber 28 for two different modes of excitation. As an example, consider the right cylindrical chamber 28, and measure the resonant frequencies in the T!V1,0 and TE... modes. Both modes are non-degenerate and their resonant frequencies can be determined to an accuracy of 1 part in 107 or better with standard techniques. The TIVI0,0 mode (parallel plate mode) is dependent only on the average diameter (D) of the chamber 28 and is independent of the cavity height (L), whereas the TE... mode is dependent on both (D) and (L). Therefore, from the measurement of the two frequencies, the volume of the chamber 28 can be determined. For the case of TIVI0,0 mode:

WD A 1 A 01 resonant wavelength = X or D 7r (11) 1 01.

where D = diameter, X0, = 1 st Bessel root or J.(X) = 0 For the case of TE,1, mode:

a AP = resonant wavelen [2Xl 1 2 + 1 ?- (12) 9th- -1r-27-1 - 1 TTD) L. 45 where D = diameter, L = length or height of cavity, X, = 1 st root of X, (x) = 0. 50 Combining the 2 results give D= LO1 Xl, L= 112 X ' 111-1/2 1) 2 1T -[X2 f - X1 L' ko (13) 55 In terms of the resonant frequencies, the results can be expressed as:

D=)C015 9 L= C [1 - R; 1 X'1 114 - 1 / 2 60 Trf 2f? 7?) jj J fa RO, where X',, = 1.8412, X, = 2.4048, and c = speed of light in the medium filling the cavity, (air65 33.

GB 2 142 7 30A 33 for the present application).

In terms of the total volume of the chamber, 1]p -112 Volume= UD21= X1p-;k2X01p- X,, -1 R_ 4 11 - X, 201 -3 2 c X 01 -112 10 slyfl 2;? 2 fl X', 2] 15 As can be seen, the volume is, to first order, proportional to X3 or 1 /f3. Therefore dv 3 dX r 3 df 20 v (--zl c) Hence, the frequencies can be measured -accurately to 1 part in 107 as by the counter 1108, and thus the theoretical accuracy for v is of the order of 3 parts in 107.

The method can be used to continuously measure the change in volume of a right circular cylinder 12 caused by a positive displacement of its piston 14. Both the volume of the chamber 28 before and after the piston motion, as well as the rate of change of volume can be measured in a simple manner. For the arrangement shown in Fig. 10, as the piston 14 moves from position X to position Y, the TM010 mode resonant frequency will remain constant (or change 30 slightly due to non-uniformity in the diameter of the cylinder) and the TIE,, mode resonant frequency will shift by an amount proportional to the displacement AL. At position X, the resonant frequencies f, and f2 are measured and are inserted within equation 2 to provide an indication of a first volume V, Thereafter, the piston 14 is moved to a second position Y and a second set of resonant frequencies f,', f,' for the modes TM,,, and TIE,,, respectively, are taken 35 and a second volume V2 is calculated in accordance with equation 15. Finally, a displacement volume AV is calculated by subtracting the first determined volume V, as determined at position X from the value V, of the second volume as determined at position Y. In addition, by continuouslymonitoring the TMO,O resonant frequency, the variation of the diameter of the chamber 28 (due to machining imperfections) between L, and L2 can be measured as a function 40 of L. In a similar manner, the rate of change of volume can be measured by continuously monitoring the resonant frequency of the TE... mode.

Perturbations to the above relationships include dielectric properties of air, the presence of coupling conduits 30 and 32, and the other gas inlet 62, surface irregularities, finite electrical conductivity of the wall material of the chamber 28 and degeneracy due to mode crossing. To 45 1 st order, so long as the irregularities are small compared to A (which will be on the order of 30 cm or larger), the perturbations will be proportional to the volume change. Hence, the method will average over deformities and give a measurement which will be proportional to the true volume of the chamber 28.

The coupling conduits 30 and 32 and gas inlet 62 are made with sizes well below the cut off 50 wavelength of the microwaves and should perturb the resonant frequency at most by 1 part in and can be corrected for in the 1 st order.

Similarly, the perturbation due to the finite electrical conductivity of the wall material of the chamber 28 should be of this same order of magnitude if the walls are fabricated or plated with a high-conducting metal such as copper, silver, gold, or aluminum, and reasonable care is taken 55 in polishing. As an example, the theoretical skin-depth for copper at 300 Mc/s is 3.8 X 10-4 cm. The perturbation on the volume is of the order of the ratio of the skin-depth to the linear dimension of the resonant cavity which, for a right circular cylinder with a radius of 50 cm, is approximately 7.6 X 10-6. The actual skin depth can be estimated from the dissipative losses in the cavity which are directly related to the quality factor or Q of the chamber, which can 60 generally be experimentally measured to about 1 % accuracy. Consequently, a first order correction can be applied which will reduce the uncertainty to better than a few parts in 107 The resonant frequency change between vacuum and air in the cavity is given by, (fvauu/fai,) = (E)1/2 (17) 34 GB 2 142 730A 34 where c is the dielectric constant for air at microwave frequencies, which for dry air at STP, has the value -s,,-1 = 536.5 X 10-rl. Hence the frequency change from vacuum to air is of the order of 2.7 X 10-4. Since c for dry air is accurately known at microwave frequencies as a function of pressure and temperature, this shift can be corrected for to an accuracy of at least 1 part in 106.

The expression W- 1)t,p/("- 1)20c,]aim] = (P/760)/[1 + 0.00341 (t-20)] can be used to correct for the pressure and temperature dependence of e to better than 0. 1 % accuracy. Since the perturbation in frequency is only 2.7 X 10 - " initially, we can expect an overall accuracy in the resonant frequency determination of the order of 10-7 if the barometric pressure is monitored to better than 0. 1 % (or about 1 mm of mercury).

The water vapor (relative humidity) contribution to the dielectric constant of air can be 15 expressed as:

( ACE- 1) -6 - 5.00 273.16 2 (18) water vapor X 10 1 T] P 20 where T is the temperature as measured by a precision temperature device in degrees Kelvin and P is the partial pressure of water vapor in millibars. For T = 20C (29 3 K), the saturation vapor pressure (100% relative humidity) is 23 millibars. Hence, for this extreme case, (V--- 1).1.,.por X 10 - 6,-.-t! 100 which is approximately 1 /3 as that for dry air. Again, this effect can be corrected to the 1 st order by measuring the relative humidity, and an accuracy of the order of 10-7 can be achieved in determining the vacuum resonant frequency of the cavity.

Both the TE,11 and the TIVIO10 are not degenerate in frequency with any other resonant TEM modes. Accidental degeneracy due to spurious mode-crossing can be avoided by choosing the 30 dimensions of the volume properly. The conditions for mOde-crossing, between D/L>O to D/L= 3 are:

D/L = 0.45; D/L = 1; and D/L = 2.14 (at D/L = 0.45, the TMO,O mode is degenerate with the TE, mode; at D/L = 1, the TM,O mode, is degenerate with the TE.. mode; at D/L = 2.14, the TIE,, mode is degenerate with the TM,O mode). Therefore, by choosing D/L ratios other than those values, interactions with spurious modes are avoided and the resonant behavior of the cavity will be well-defined and the formulas for calculating the resonant frequencies from the dimensions of the chamber 28 are 40 rigorously valid.

As an illustrative example, suppose we choose to work in the region of 1 <D/L<2.14 and require that the net traverse of the piston 14 displace a volume equal to 8 cu. ft. (2.26 52 X 105 cc). Then the following configuration can be used:

D = 104.88 cm L, = 52.44 cm = final piston position Y L2 = 78.66 cm = initial piston position X.

Hence, the net displaced volume is, 7r(104.88)2 4 (26.22) = 2.2652 X 105 cc.

Also, as can be seen, the D/L ratio varies from 1.33 for the initial position to 2 for the final 55 position, which are well within the desirable operating range. For this case then: f, = resonant frequency of TM,,, mode = 219.0 Mc/s f2(i) = initial value of TE... mode = 253.9 Mc/s (D/L = 1.33) f2(f) = final value of TE1, mode = 331.5 Mc/s (D/L = 2) Similarly, for the case of 2 cubic feet total volume, the frequencies are:

fi = 347.6 Mc/s f,(i) = 403.0 Mc/s f2(f) = 526.2 Mc/s The dependence of frequency for an incremental change in L can be expressed as:

GB 2 142 730A 35 Af - - 2 AL 1. 3 7."3--9 ( 2 2) +1- 1 L 1 which for D/L = 1.33 gives, (Af/f) = 1. 1 (AL/L) (20) and for D/L = 2 gives, (Af/f) = 1.5 (AL/L).

Li 1 ji (15) As can be seen, the uncertainty in measuring L is nearly equal to that for the frequency 15 measurement. Consequently, very high precision can be obtained for determining the volume displaced in this dimension configuration.

It is also informative to estimate the quality factor Q of the resonant modes, since the precision in measuring the resonant frequencies will depend, to a great extent, on the sharpness 20 of the resonances. For the TMOO mode:

8 Q- 0.22 for D/L = 1.33, A 8 Q-0.19 for D/L= 2. X Where 8 is the skin depth given by 8 = [(Xp)l 1 2072g]1/2, p is the resistivity of the wall material 30 of the chamber 28, A is the wavelength, and g is the permeability of the wall material.

If the chamber 28 is made of copper, p = 1. 7 X 10, It = 1 and 8 = 4.43 X 10-4 cm at 219 Mc/s. Therefore Q= 6.8 X 104 for D/L = 1.33 and G= 5.9 X 104 for D/L = 2.

Depending on the coupling coefficient, the width of the resonance curve at the half power points varies between (2f./Q) and (fJO) where f,, is the resonant frequency. Consequently, the width of the resonance curve for the values of G computed are:

at % = 219 Mc/S: 3.2:5Af:56.4 kc/s for G = 6.8 X 104 and 3.7:5Af:57.4 kc/s for G = 5.9 X 104 Since f. can be usually determined to an accuracy of 102 of f or better, we can expect an accuracy of the order of 10-7 in determining f.. This in turn implies accuracy of this order in the 45 measurement of the diameter D of the chamber 28.

Similarly, for the TE... mode:

8 Q- = 0.28 for D/L = 1.33, 50 X a Q- = 0.27 for D/L-2, A which gives, G= 7.5 X 10-4 at 253.9 Mc/s and Q= 5.5 X 10-4 at 331.5 Mc/s. As before, the widths of the resonance curves are, 60 3.4---f:56.8 kc/s at 253.9 Mc/s and 6.0.<f:512.0 kc/s at 331.5 Mc/s. Again adapting the criteria that f. can be determined accurate to 10-2 of Af, then for the worst case of (12.0 kc/s) Af./f,=4 X 10 - 1. From the expression previously derived for D/L-2, 36 GB 2 142 730A 36 Af 0 f 0 = 1. 5 (',L) Therefore, (AL/L) can be determined to (4 X 10-7)/1.5=-2.7 X 10-7. It is also possible to calculate the perturbation on the resonant frequency of a cavity mode due to the presence of a gas-inlet and outlet opening 62 on the head 60 of the cylinder 12.

From the Adiabatic Invariance theorem and a knowledge of the field configuration inside the 10 cavity, the frequency pulling caused by the hole can be estimated in a straightforward manner. If the hole dimension is well below the cut-off wavelength (which will be rigorously true for the case under consideration), then the frequency pulling will be proportional to the ratio of the cube of the hole diameter to the volume of the chamber 28.

Illustratively, the expression for the change in the resonant frequency of the TIVI010 mode 15 caused by the opening 62 located at the center of the plate 60 is given by, (Af/fJ = (d 3) /81) 2 L(X01)J 12(X01) (21) where d is the diameter of the hole J1(X0,) is the value of the Bessel function J, at X,, and Af is 20 the frequency shift.

Upon numerical evaluation using D = 104.88 cm, L-52.44 cm, X, = 2.40483 and J12 (X01) = 0.2695, then (Af/fJ = 3.35 X 10-7 d 3 As can be seen, for d of the order of 2 cm or less, the frequency shift is only of the order of 2 X 10-6. Consequently, the shift is very small and with an appropriate initial calibration procedure, such as covering the opening 62 with a matching metallic plug, this effect can be virtually eliminated as a systematic error in the precision of the method.

Coupling of the microwave energy to the chamber 28 for the two modes TM,O and TE,1 can best be accomplished by placing a coaxial feed line terminated in the antenna 70 at a position approximately 1/2 way out from the center of the end plate 60 of the cylinder 12 with the loop oriented along a radius. The magnetic field at this location is about 90% of the maximum field intensity inside the cavity for both modes. Consequently, both modes are energized to the same 35 degree of coupling with high efficiency. In addition, by placing the coupling on the head 60, the coupling will not be affected by the movement of the displacement piston 14.

In the following, a description will be given of a second preferred method of measuring a displacement volume within the chamber 28 and using this accurately determined volume to calibrate the train of pulses as provided by the optical, linear encoder 26. In a similar manner to 40 that described above, the piston 14 is moved from a first position as indicated in Fig. 10 by designation L, to a second position indicated by the designation L, having moved through a displacement of AL. The cylinder 12 is inherently rigid whereby the calibration process, as will be described, may be only carried out occasionally to insure that no long term systematic changes, such as dimensional deformation of the cylinder 12, misalignment and malfunction of 45 the optical linear encoder 26, or distortion of the piston 14 has occurred. In order to maximize the absolute measurement accuracy of the microwave volume calibration, it is necessary that the mechanical configuration of the cylinder 12 be as close as possible to that of a perfect, totally enclosed right circular cylinder and thereby eliminate or reduce all possible sources of systematic perturbations which could potentially affect the microwave measurements.

Referring now to Fig. 12A, certain mechanical modifications are made. First, the physical gap that exists between the piston 14 and the walls of the chamber 28 must be effectively blocked to prevent the escape of microwave energy through that gap. As explained in the aboveidentified application, filed concurrently herewith and entitled "Piston Sea[", because of the nature of the seal between the piston 14 and the wall of the chamber 28, the gap is considerable. A cover 11 is made of a suitable metallic material, such as stainless steel, and further, has a series of springlike fingers 15, as shown in detail in Fig. 1213, disposed between the piston 14 and the inner periphery of the chamber 28, Which, when the cover 11 is in place over the piston 14, project into the gap between the piston 14 and the wall of the chamber 28, the fingers being in close contact with the piston 14 and the wall. The fingers 15 act as a short 60 circuit reflecting the electromagnetic field that would otherwise be directed through the noted gap. In an illustrative embodiment, the spring-like fingers 15 are made of a beryllium copper.

Further, the pressure and temperature sensors 51, 57, 48, and 68 are removed and are replaced by appropriate blank metallic plugs, configured to provide a substantially flush surface with the inside walls of the chamber 28. Further, the fluid inlet opening 62 in the head 60 is 65 37 GB2142730A covered by a metallic plate to provide a substantially flush surface across the top of the head 60. In addition, the inside peripheral walls of the chamber 28 are cleaned with a suitable solvent to remove any residual traces of the oil as may have seeped from the pistol seal. Noting that the required calibration is determined by a displacement volume AV and not by the absolute volume of the chamber 28, the abovedescribed mechanical modifications do not affect the accuracy of the calibration process. Once the measurements, as will be described, are made for the modified chamber 28, the same set of measurements may be carried out immediately afterwards with the chamber 28 restored to its normal operating configuration and a set of appropriate calibration factors can be generated to relate the two sets of measurements. The results of the second measurements can then be used as a data base from which subsequent 10 checks of the absolute calibration can be compared without going through the full procedure of modification and reassembly of the chamber 28.

Briefly, the volume measuring and calibration process includes the step of moving the piston 14 to a first position indicated by the designation L, in Fig. 10, by manually rotating the rotary member 19 of the servomotor 20. At the first position, the antenna 70 is energized with electromagnetic energy of a first mode TE,11 and a second mode TE,21 selected to minimize the above-discussed perturbations. The frequencies f, and f2 at which resonance is established for each mode are detected by observing the counter 1108. Then, the piston 14 is moved through a distance AL to a second position as indicated by the designation L21 whereat the antenna -70 is energized again with electromagnetic energy of the first and second modes and corresponding 20 frequencies at which resonance is established for each of the modes are noted. The output of the optical, linear encoder 26 is applied to a counter, which counts the linear encoder pulses as the piston 14 is moved through the distance AL. The diameters D, and D2 of the chamber 28 at each of the first and second positions corresponding to the designations L, and L2 are calculated. At this point, a calculation of the AL is made using the previously calculated values 25 of D, and D2. The calculated value of AL is divided by the number of pulses derived from the linear encoder 26 as counted during the movement of the piston 14 through the distance AL to provide a length calibration factor using the measurements of D, and D2. The volume AV corresponding to that volume as defined by planes passing through the points L, and L2 and the inner periphery of the chamber 28 is expressed by a mathematical expression in terms of the 30 diameters D, and D2 and AL. If the output of the optical, linear encoder 26 is to be calibrated for a given volume, e.g., one cubic foot, that value is disposed in this equation and it is solved for the calculated values D, and D, to provide that value of AL corresponding to the movement of the piston 14 to draw one cubic foot of fluid through the meter 38. The calculated value of AL is multiplied by the previously calculated length calibration factor to provide that number of 35 pulses that will be output by the optical, linear encoder 26 as the piston 14 is moved a length AL to draw the one cubic foot into the chamber 28. As explained above, the count as derived from. the linear encoder 26 is used to calculate the calibration factor as incorporated within the calculation carried out in step 1062, as shown in Fig. 9N. In particular the calibration factor is the reciprocal of the counts so derived for one cubic foot of fluid drawn into the chamber 28 40 and provides a correction to the calculation of percent error in the reading of the meter based upon a precise measurement of the volume of the chamber 28, as explained above.

First, it is necessary to measure the frequencies at which the standing wave conditions are established at the positions L, and L2. The calculation of diameters D, and D2, as will be explained, requires a value of the speed of light, which changes for varying ambient conditions 45 of temperature, pressure, and relative humidity. Corrections for changes in the speed of light are expected to be small, and the calculation of the speed of light is made typically once or twice during the course of a calibration process of the optical, linear encoder 26.

The speed of light in vacuum, Co, is 2.997925 X 1010 cm/sec. The corresponding value c for air is obtained by dividing Co by the refractive index of air at the wavelength of observation. For 50 the microwave region (f<30GHz), the refractive index, n, is related to the atmospheric parameters by the equation:

6 77.6 4810e (n-1) x 10 = --f - (P + T (22) where P is the total pressures in millibars (1 bar = 106 dynes /CM2 = 0. 986923 standard atmosphere 75.0062 cm Hg at O'C), T is the temperature in degrees Kelvin, and e is the partial vapor pressure of water in millibars. The speed of light is then given by c = Co/n = Co,,"[1 + 77.6 (p + 4810e) x 10-6] (.13) 2 T T 38 GB 2 142 730A 38 The temperature and barometric pressure can be directly obtained from the readings of a thermometer and a barom-eter placed near the meter prover 10. The partial vapor pressure of water can be deduced from the relative humidity data obtained with a sling psychrometer through the use of the psychrometer formula, or more conveniently, via the use of a standard table such as the Smithsonian Physical Table #640.

In order to calculate a value of AL, there is needed to determine the average value of the diameter of the chamber 28 and more specifically, to determine the values of the diameters D, and D2 at the locations L, and L2, respectively. The calculation D, and D2 is carried out with great care since the resulting uncertainty in the volume is approximately twice the uncertainty of this measurement. As explained above, the piston 14 is moved to the first position correspond- 10 ing to the designation L, at which the frequencies f, and f, for which the resonant standing wave condition is established for the two different modes. The preferred method is to measure simultaneously the resonant frequencies f, and f2 of two different modes of the same electrical characteristics as a function of the piston position L and solve for the average diameter 6 by using the appropriate theoretical expression. 1 In a preferred embodiment wherein the chamber 28 has the configuration of a right circular cylinder, the pair of modes preferred for this purpose are the TE... and TE,12 modes. As will be discussed, it has been demonstrated that the quality factor 0 obtained by excitation in these modes is high thereby reducing the effects of perturbations upon the measurements of resonant frequency. The average diameter of the cylinder at any fixed position of L is given by the 20 expression:

- 1.015098c D (L) - L 4 f 2 2 (L) _f 1 2 (L) 1 1/2 (24) 25 where f, is the resonant frequency of the TE.. mode and f, is the resonant frequency of the TE112 mode, and c is the speed of light in air as calculated by equation 23. By using two different modes of electromagnetic wave energy excitation, the various perturbations such as skin-depth variation, reactive frequency pulling caused by the antenna 70, the degree of divergency of the inner periphery of the chamber 28 from being a perfect right circular cylinder are compensated for and the absolute value of I) is obtained with great accuracy. By exercising 35 care in the taking of measurements of the frequencies upon the counter 1108, as shown in Fig. 35 10, absolute accuracies of the values 15) as a function of L may be obtained in the order of one part in 105 or 0. 1 mill out of a 12 inch diameter. This degree of accuracy is of the same order of the changes in the volume of the chamber 28 due to thermal expansion and contraction as disposed in a temperature stabilized environment where the temperature is maintained within range of i: 1 F.

* In order to confirm these measurements as well as to provide a quantitative means for evaluation of the order of magnitude of the expected perturbations in the system of measurement, the diameter may be independently determined by measuring the resonant wave frequencies by generating electromagnetic waves of the TM0,0 mode within the chamber 28.

With such a mode of excitation, the resonant frequency is independent of the length L and therefore for a perfectly uniform cylinder, should not change as the position of piston 14 is varies. However, excitation in the TM010 mode is subject to other various perturbations which need be considered to achieve the same degree of accuracy as for the two modes discussed above. For the TM,10 mode, the average diameter is given by the expression:

- 50 D(L) = 0.7654799c/f (25) where f is a resonant frequency of the TM,, mode.

Once the average diameter D of the chamber 28 as a function of L has been determined to the desired degree of accuracy, the value of AL is obtained and related to the observed number 55 of pulses from the optical, linear encoder 26 in order to obtain the length calibration factor in terms of length per pulse interval or number of pulses per inch. The piston position is set at L, and the resonant frequencies at f, and f, for the selected modes TE... and TE,2 are measured.

The piston 14 is then moved by cranking the rotary member 19 to a new position L2 and the resonant frequencies of the same modes are remeasured, while counting the number of optical 60 encoder pulses during the movement of the piston 14 from its first to its second position. The number of pulses is divided by AL = L,-L2 to provide the desired length calibration factor. The distance AL = (L11-2) should be large enough such that the calibration accuracy is not limited by the accuracy in the pulse count ( 1 in this case) and the calibration should be performed over a number of AL intervals to insure that no non-linearity effects exist in these measurements.

39 GB 2 142 730A 39 For the TE1, mode, the change in distance AL is given by the expression:

-112 1191 - 11 (, f 1 0.5860671C P- rO.5660G71 &L:g. (Lg- 2 - 1 1- D (L1)21.

L) = Q [[f2 - 1, where f2 and D(L2) are the TE,11 mode resonant frequency and the previously determined average diameter at piston position L2, and f, and D(L1) are the respective values at position L, 10 For the TE,12 mode, the change in distance AL is given by the expression:

11i lc (27) A&(L2-LI)=c f2'-[ ú(Lo)j 1- 0061 2 -f - P.580 1 1 2 f2 -9 D(L1))a] where the various quantities are defined in a similar manner as above.

Similar expressions can be written for any mode of excitation, and more than one mode can be used to check the internal consistency and absolute accuracy of these measurements.

The absolute calibration of the displacement volume AV between the piston positions IL, and L2 is provided by the following expression:

AV - V 2 -v 1 = (1r/CLD 2 2 (A L) + (D 2 2 -D 1 2) L (28) where D2 and D, are the averaged diameters of the cylindrical cross- sections as taken at piston positions L, and L2 and AL = (L2-L,). It is evident from an observation of the equation 28 that knowing values of D2 and D, if we assume for calibration purposes a given value of the absolute displacement volume AV, e.g., one cubic foot, that the corresponding value of AL, i.e., that distance through which the piston 14 must be moved in order to draw one cubic foot of 35 fluid into the chamber 28 of the meter prover 12, may be calculated. The object of the calibration is to obtain the number of pulses as derived from the optical linear encoder 26 that are output for any desired displacement volume AV and is obtained by multiplying the obtained value of AL for a given volume by the length calibration factor to provide the equivalent number of pulses that are output by the optical linear encoder 26.

The selection of the TIE,, and TE,12 modes to excite the cavity 28 was based upon repeated determinations using a number of normal mode resonances to determine the quality factor Q for each of the modes. These determinations of the Q of a normal mode require the measurements of the ratio of the reflected power (Pr) to the incident power (PO) at the resonant frequency, and the frequency width of the response curve corresponding to the half-power level defined by P1/2 (Po + Pr)/2. It is desired that the d.c. voltage response of the crystal detector 1110 be linear with input microwave power. This condition can be satisfied by operating the crystal detector 1110 in the so-called "square law detection" region corresponding to a d.c. level typically less than 20 millivolt. If necessary, the linearity of the response can be verified by the use of the step attenuator located on the sweep generator 1100. Once established, the coupling 50 coefficient (Pr/Po) can be measured directly on the screen of the oscilloscope 1112 in terms of the corresponding voltage ratio. The half-power level then can be calculated as an equivalent voltage. The half width of the response curve, as shown in Fig. 14, is just the difference between the two frequency settings as set on the sweep generator 1100 corresponding to the half-power levels on either side of the resonant frequency as observed on the oscilloscope 1112. The Q of the resonance is calculated by using the expression given by equation 10. From these determinations of Q, it was demonstrated that TE1, mode has a quality factor Q of approximately 6,000 to 7,000 over the running range of the piston 14, while the TE1,2 mode has a quality factor Q of 8,000 to 10,000. As indicated above, the quality factor is a measure of the order of magnitude of the expected deviation of the resonant frequency from the idealised 60 results as provided by equation 1. Thus by using these modes, the resonant frequency may be measured with a greater accuracy as those perturbations as would arise due to surface imperfections as well as for the effects of skin depth and frequency pulling may be minimised. Thus, the use of the modes TE,1, and TE,12 are believed to provide determinations of greater accuracy of the resonant frequency, and thus of the average diameter D and of the volume GB 2 142 730A 40 displacement between the two piston positions.

Thus, there has been described a meter prover that is capable of measuring fluid and in particular, gas flow through a meter with a high degree of precision. Preferably the volume of the cylinder into which the fluid is drawn is measured with extreme accuracy and is compared to the output of the encoder which detects movement of the cylinder's piston, whereby indication 5 of the volume drawn into the cylinder is provided with a corresponding high degree of accuracy.

This standard or calibrated volume is compared with the output of the meter under test to provide an indication of meter registration, as well as the percentage of error of the meter fluid reading from the actual or calibrated volume indicated by the optical encoder of the meter prover system. Further, the described example of the meter prover system is controlled by a computer system whereby a number of tests are made in which parameters of meter and prover temperature and pressure are taken into consideration to adjust the indication of the measured volume of fluid flow, as well as to take repeated tests under varying conditions. In particular, varying volumes of fluid may be drawn by the meter prover through the meter by entering corresponding count factors into a counter of the computer and counting the selected count to 15 zero, to terminate the meter test. A method is employed in this example for determining with high precision the volume of the cylinder into which the fluid is drawnfor a given displacement of the piston. This accurate measurement is determined by the frequencies at which standing waves are established for first and second piston positions to provide a precise indication of the fluid volume and the output of the optical encoder coupled to detect the movement of the 20 piston.

CLAIMS 1. Apparatus for measuring the volume of the chamber of a meter prover, said prover including a piston adapted for rectilinear movement within the chamber to direct a fluid between 25 a meter under test and the chamber, said volume measuring apparatus comprising:

(a) means for generating and varying the frequency of outputted microwave electromagnetic energy; (b) Antenna means coupled to said generating means for emanating waves of electromagnetic energy into said chamber; (c) means for detecting the electromagnetic energy leaving said chamber; (d) resonant condition detecting means coupled to said energy detecting means for providing an indication of the occurrences of the minimum levels of energy leaving said chamber corresponding to the establishment of a standing wave resonant condition within the chamber; and (e) means for determining the frequencies at which standing wave resonant conditions are established within the chamber whereby the volume of the chamber is accurately determined.

2. Apparatus as in Claim 1 wherein said generating means comprises means for varying the frequency of the microwave electromagnetic energy until a standing wave condition is established within the chamber.

3. Apparatus as claimed in Claim 2 wherein said generating means automatically varies the frequency of the output microwave electromagnetic energy.

4. Apparatus as in Claim 2, wherein said generating means comprises operator manipulable means for varying the frequency of the outputted electromagnetic wave energy.

5. Apparatus as in Claim 2, 3 or 4 wherein said energy detecting means and said generating 45 means are each coupled to said antenna means and to said resonant condition detecting means, whereby a manifestation of the level of energy leaving said chamber is provided as a function of the frequency of the electromagnetic energy generated by said electromagnetic generating means.

6. Apparatus as claimed in Claim 2, 3, 4 or 5 wherein said frequency detecting means is 50 coupled to the output of said electromagnetic energy generating means and comprises means for displaying the frequency of the electromagnetic energy.

7. Apparatus as in Claim 6 wherein said detecting means comprises a counter.

8. Apparatus as in any one of Claims 2 to 7 including means coupled to the output of said electromagnetic energy generating means for detecting the level of energy and for providing a 55 feedback signal indicative of the level of energy to said electromagnetic energy generating means, whereby the level of power required by said electromagnetic energy generating means is stabilised.

9. Apparatus for measuring the volume of the chamber of a meter prover substantially as hereinbefore described with reference to and as shown in the accompanying drawings.

10. A method of measuring the volume of a chamber of a regular geometry, said method comprising the steps of:

(a) generating electromagnetic energy field within said chamber of a normal mode selected so that at a resonant condition within said chamber the electric and magnetic component fields of the energy field are related to the dimensions of the regular geometry chamber; 41 GB 2 142 730A 41 (b) extracting reflected energy from said regular geometry chamber; (c) sensing the frequency of the electromagnetic energy that establishes a resonant condition for the normal mode within said chamber; and (d) determining, based upon the sensed resonant frequency, the dimensions and thus the 5 volume of the chamber.

11. A method as in Claim 10, wherein the chamber is a right circular cylinder and the- step of determining determines the axial length and diameter and therefore the volume of the right circular cylinder chamber.

12. A method as in Claim 10 or 11, wherein the normal mode for generating electromag- netic energy field is selected to minimise the perturbations due to the character, configuration, 10 and surface of the chamber.

13. A method as in Claim 12 wherein the mode selected is the TE, mode.

14. A method as in Claim 12, wherein the mode selected is the TE,12 mode.

15. A method as in Claim 12, wherein the mode selected is the TMO90 mode.

16. A method of measuring the volume of a chamber substantially as hereinbefore described 15 with reference to the accompanying drawings.

Printed in the United Kingdom for Her Majesty's Stationery Office, Dd 8818935, 1985, 4235. Published at The Patent Office, 25 Southampton Buildings, London, WC2A lAY, from which copies may be obtained.