JP2014109525A - Thermal noise thermometer and reference noise voltage generator used for the same - Google Patents

Abstract

An object of the present invention is to reduce crosstalk between wirings, change a voltage spectral density in a short time, and further greatly reduce the cost of the entire apparatus.
For example, every time a pulse of logic “1” in a pseudo-random number sequence is input, a pulse multiplier 21 calculates a magnetic flux quantum number or a magnetic flux quantum voltage pulse by a spectral density control signal supplied from the outside. A magnetic flux quantum voltage pulse in which the level in the voltage axis direction due to the superimposition of the signal is multiplied within a predetermined unit time T ₀ is output. The pseudo-random number generator 22 generates an M series using superconducting digital circuit technology. For precision measurement applications, the output circuit composed of the Josephson transmission line 23 and the SQUID array 24 is configured using a Josephson circuit capable of complete quantization. The pulse multiplier 21, the pseudo random number generator 22, and the output circuit are connected to each other by a short-distance wiring having a high shield efficiency and are formed in the reference noise voltage source chip 25.
[Selection] Figure 2

Description

  The present invention relates to a thermal noise thermometer and a reference noise voltage generator used therefor, and more particularly to a thermal noise thermometer for measuring the temperature of a resistor and a reference noise voltage generator used therefor.

An object of electrical resistance value R at temperature T generates a thermal noise (Johnson noise) voltage. This thermal noise voltage is white noise having a uniform frequency spectrum. The root mean square value <ν ² > of the thermal noise voltage is expressed by the following Nyquist equation.
<Ν ² > = 4 kTRB (1)
Here, in the above formula, k is a Boltzmann constant, and B is a measurement bandwidth. The equation (1) indicates that the temperature T of the resistor can be measured by measuring the mean square value <ν ² > of the thermal noise voltage generated by the resistor. Based on this principle, the thermal noise thermometer measures the temperature T of the resistor.

  Electric quantity measurement is a technology that has been developed for a long time, and enables precise measurement with small errors. FIG. 13 shows application temperature ranges of various precision temperature measurement methods. In FIG. 13, [] shows a typical measurement error, and the upper part shows a typical temperature fixed point (created from Non-Patent Document 1). As shown in FIG. 13, according to the thermal noise temperature measurement method, it is known that accurate temperature measurement can be performed with an extremely small error of about 10-50 ppm in a very wide temperature range as compared with other precision temperature measurement methods. ing. For example, in the vicinity of 0 ° C, which is the triple point of water, the acoustic gas temperature measurement method and the constant volume gas temperature measurement method can be used for extremely precise temperature measurement, but according to the thermal noise temperature measurement method, the temperature measurement of the triple point of water is possible. In addition, the temperature of each freezing point of Al, Ag, and Cu, which cannot be measured by the acoustic gas temperature measurement method or the constant gas temperature measurement method, can be measured with an extremely small error of a heat temperature of 50 ppm or less. Therefore, the thermal noise thermometer is used in the field of precision temperature measurement such as temperature standard and thermometer calibration because of its feature of being able to measure temperature with high accuracy over a wide temperature range.

  In the above thermal noise thermometer, in order to accurately measure the noise voltage of the resistance object (resistance sensor), the measurement is caused by the characteristics of the measurement system composed of a cable, an amplifier, an A / D converter, etc. Errors must be eliminated. Therefore, before measuring the temperature of the resistance sensor, the response of the measurement system to a reference noise voltage waveform having characteristics equivalent to the noise voltage of the resistance sensor is evaluated, and the measurement system is calibrated.

  When performing calibration, the reference noise voltage generator that generates the reference noise voltage waveform having the same characteristics as the noise voltage of the resistance sensor has two transmission paths (two transmission systems consisting of an amplifier and an A / D converter). The system is connected to a measurement system, and a correlation between two reference noise voltages that have passed through these two transmission paths is calculated by a digital signal processor, and the characteristics of the measurement system are evaluated. Thereafter, the resistance sensor is connected to the measurement system after the characteristic evaluation, and thermal noise temperature measurement is performed.

  Therefore, in order to accurately measure the noise voltage of the resistance sensor, the reference noise voltage generator used when executing calibration of the measurement system is an important component of the thermal noise thermometer. This reference noise voltage generator must output a pseudo-noise waveform having a frequency spectrum that can be accurately calculated (the waveform is determined with high accuracy). The pseudo-noise waveform is desirably white noise that is the same as the thermal noise of the resistor, but need not have an exact white spectrum if it can be calculated.

  As the reference noise voltage generator, for example, a reference noise voltage generation method shown in the block diagram of FIG. 14 is disclosed by the NIST in the United States and is currently widely used (for example, see Non-Patent Documents 2 and 3). ). The reference noise generator 100 includes a computer 101 that generates a pseudo random number sequence using Σ-Δ modulation, a fast Fourier transformer 102 that verifies the pseudo random number sequence, and a pulse pattern generation that synthesizes a noise voltage waveform from the pseudo random number sequence. And a Josephson junction array 104 that precisely quantizes the noise waveform from the pulse pattern generator 103 and outputs a reference noise voltage. In the reference noise generator 100, the frequency spectrum of the pseudo random number sequence is verified by the fast Fourier transformer 102 in order to guarantee the accuracy of the reference noise waveform.

J. Fischer, M. dePodesta, KDHill, M. Moldover, L. Pitre, R. Rusby, P. Steur, O. Tamura, R. White, and L. Wolber, “Present estimates of the differences between thermodynamic temperatures and the ITS -90 ”, International Journal of Thermophys, vol. 32, pp. 12-35, (2011). SPBenz, A.Pollarolo, J.Qu, H.Rogalla, C.Urano, WLTew, PDDresselhaus, andD.R.White, “An electronic measurement of the Boltzmann constant”, Metrologia, vol.48, pp.142- 153, (2011). S.P.Benz, P.D.Dresselhaus, and C.J.Burroughs, “Multitonewaveform synthesis with a quantum voltage noise source”, IEEE Transactions on Applied Superconductivityvol.21, pp.681-686, (2011).

  However, in the reference noise voltage generator 100 shown in FIG. 14, the input signal is coupled to the reference voltage output waveform due to crosstalk between the input and output signals in the Josephson junction array 104, so that the reference noise waveform accuracy is deteriorated. Measurement errors occur and temperature measurement accuracy is limited.

  In the reference noise temperature generator, the voltage spectral density of the reference noise voltage must be controlled so as to correspond to the measured temperature. In addition, the voltage spectral density must be fine-tuned in real time following changes in the characteristics of the resistance sensor during measurement.

  However, in the reference noise voltage generator 100 shown in FIG. 14, in order to change the voltage spectral density of the reference noise voltage waveform, it is necessary to recalculate the input pseudorandom number sequence to the pulse pattern generator 103. An enormous cost (time, computer usage fee) is required for column generation / verification calculation. For example, in order to generate and verify a pseudo-random number sequence for one measurement temperature point, the computer 101 and the fast Fourier transformer 102 require several days of calculation time using the highest class workstation. In practice, many temperature points need to be measured, and the reference noise voltage spectral density value must be controlled to a corresponding value each time.

  In addition, in precise measurement, in order to control the reference noise voltage spectral density in real time following the change in characteristics of the resistance sensor due to drift, programmable control of the reference noise voltage spectral density in a short time is required. However, in the conventional method, it is difficult to control the reference noise voltage spectral density in real time due to the enormous calculation cost of generating and verifying the pseudo random number sequence.

  Further, the pulse pattern generator 103 needs to use a high-speed and large-capacity memory in a necessary band, but such a memory is extremely expensive, and the price of what is commercially available is tens of millions of yen. . The high price of the pulse pattern generator 103 increases the price of the entire thermal noise thermometer, which is one of the factors that prevent the spread of the thermal noise thermometer.

  The present invention has been made in view of the above points, a thermal noise thermometer that can reduce crosstalk between wirings, can change a voltage spectral density in a short time, and further greatly reduces the cost of the entire apparatus, and a thermal noise thermometer used therefor An object is to provide a reference noise voltage generator.

In order to achieve the above object, the thermal noise thermometer of the present invention passes through a resistance sensor, a reference noise voltage generating means for generating a reference noise voltage having characteristics equivalent to those of the resistance sensor, and two measurement systems. Measuring means for obtaining a measured value by digital signal processing from two noise voltages, and supplying only the reference noise voltage from the reference noise voltage generating means to the measuring means at the time of calibration, and two references through the two measurement systems After calculating the correlation of the noise voltage and evaluating the characteristics of the measurement system, only the measurement voltage from the resistance sensor is supplied to the measurement means at the time of measurement, and two measurement voltages passed through the two measurement systems are obtained. , A thermal noise thermometer comprising a switch means for determining a measured value by correcting using the correlation obtained at the time of calibration, wherein the reference noise voltage generating means is:
A pseudo random number generator configured using superconducting digital circuit technology for generating a pseudo random number sequence that can be calculated, and a unit time when the magnetic flux quantum voltage pulse sequence constituting the pseudo random number sequence is a predetermined random value. A pulse multiplier for outputting a magnetic flux quantum voltage pulse train in which one or both of the number of magnetic flux quantums and the level in the voltage axis direction are increased at a rate variably controlled by an external control signal; An output circuit that outputs the reference noise voltage by completely quantizing and superimposing the magnetic flux quantum voltage pulse train output from the multiplier using the Josephson effect is formed on the same substrate. It is characterized by that.

  Here, in the thermal noise thermometer of the present invention, the output circuit includes a Josephson transmission line for transmitting the magnetic flux quantum voltage pulse train output from the pulse multiplier, a magnetic coupling with the Josephson transmission line, and in series with each other. It comprises a SQUID array composed of a plurality of connected superconducting quantum interference elements, and the reference noise voltage obtained by fully quantizing and superimposing the magnetic flux quantum voltage pulse train is output from the SQUID array.

  In order to achieve the above object, the reference noise voltage generator of the present invention supplies a reference noise voltage to the measurement system during calibration in order to eliminate measurement errors caused by the characteristics of the measurement system of the thermal noise thermometer. A pseudo-random number generator configured using superconducting digital circuit technology, which generates a pseudo-random number sequence that can be calculated, and a magnetic flux quantum voltage pulse sequence that constitutes the pseudo-random number sequence. Outputs a magnetic flux quantum voltage pulse train that increases either or both of the number of magnetic flux quanta per unit time and the level in the voltage axis direction at a rate that is variably controlled by an external control signal at a given random number. And the pulse quantum voltage pulse train output from the pulse multiplier are completely quantized and superimposed using the Josephson effect and output as the reference noise voltage. A circuit, characterized in that but a structure that is formed on the same substrate.

Here, the above-described pulse multiplier of the reference noise voltage generator of the present invention receives a magnetic flux quantum voltage pulse train constituting a pseudo random number sequence and is activated when a predetermined pulse value is generated to generate a pulse with a constant period τ. Oscillating means for counting, and counting means for counting pulses output from the oscillating means and supplying an oscillating operation stop signal to the oscillating means to stop the oscillating operation when the counted value reaches a value specified by the control signal; The magnetic flux quantum voltage pulse train in which the number of magnetic flux quantum per unit time T ₀ (where T ₀ > τ) of the magnetic flux quantum voltage pulse train is variably controlled by the control signal may be output from the oscillation means. .

  Further, the pulse multiplier of the reference noise voltage generator of the present invention includes a voltage multiplier circuit composed of a plurality of amplifiers connected to each other's output terminals, a magnetic flux quantum voltage pulse train that constitutes a pseudo random number train, and a control When the signal is input and the input magnetic flux quantum voltage pulse train has a predetermined pulse value, pulse distribution for supplying the magnetic flux quantum voltage pulse to one or more amplifiers specified by the control signal among a plurality of amplifiers A voltage multiplier circuit that outputs a magnetic flux quantum voltage pulse train whose level in the voltage axis direction per unit time of the magnetic flux quantum voltage pulse train is variably controlled by a control signal.

  According to the present invention, the temperature measurement error due to the crosstalk between the conventional wirings can be greatly reduced, and thereby the measurement accuracy of the thermal noise thermometer can be improved as compared with the conventional one. In addition, according to the present invention, the voltage spectral density can be finely adjusted almost in real time following the characteristic change of the resistance sensor during measurement. Furthermore, according to the present invention, the cost of the entire apparatus can be greatly reduced by eliminating the need for a pulse pattern generator.

It is a block diagram of one embodiment of a thermal noise thermometer according to the present invention. It is a block diagram of one embodiment of a reference noise voltage generator according to the present invention. FIG. 3 is a block diagram of a first embodiment of the pulse multiplier in FIG. 2 and its normalized frequency versus gain characteristic diagram. FIG. 3 is a block diagram of a second embodiment of the pulse multiplier in FIG. 2 and its normalized frequency versus gain characteristic diagram. FIG. 5 is a configuration diagram of an example of a pulse distributor and a voltage multiplication circuit in FIG. 4. FIG. 6 is a circuit diagram of an example of a basic cell of the voltage multiplication circuit in FIGS. 4 and 5 and a signal waveform diagram of each part. FIG. 6 is a block diagram of a third embodiment of the pulse multiplier in FIG. 2 and its normalized frequency versus gain characteristic diagram. FIG. 3 is a configuration diagram, an output voltage versus time characteristic diagram, and an output spectral density versus frequency characteristic diagram as an example of the pseudorandom number generator in FIG. 2. FIG. 3 is a circuit diagram of an example of a Josephson transmission line in FIG. 2 and signal waveform diagrams of each part. FIG. 3 is a circuit diagram of an example of a Josephson transmission line and a SQUID array in FIG. 2. It is a signal waveform diagram of each part of FIG. It is a circuit diagram of the other example of the Josephson transmission line and SQUID array in FIG. It is a figure which shows the applicable temperature range of various precision temperature measurement systems. It is a block diagram of an example of the conventional reference noise voltage generator.

Next, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 shows a block diagram of an embodiment of a thermal noise thermometer according to the present invention. A thermal noise thermometer 1 of the present embodiment includes a resistance sensor 10 that generates a thermal noise voltage to be measured, and a reference noise voltage generator that generates a reference noise voltage having characteristics equivalent to the thermal noise voltage of the resistance sensor 10. 20, a switch 30, amplifiers 40 a and 40 b, A / D converters 50 a and 50 b, and a digital signal processing unit 60. The amplifier 40a and the A / D converter 50a constitute a first measurement system, and the amplifier 40b and the A / D converter 50b constitute a second measurement system. The switch 30 connects either the resistance sensor 10 or the reference voltage generator 20 to the first and second measurement systems.

  As described above, in order to eliminate a measurement error due to the characteristics of the measurement system, before the thermal noise voltage measurement of the resistance sensor 10, only the reference noise voltage generator 20 is switched to the first and second measurement systems by the switch 30. Connect each to perform calibration. Thus, the reference noise voltage output from the reference noise voltage generator 20 is amplified and A / D converted by the switch 30 by the amplifier 40a and the A / D converter 50a of the first measurement system, and then the digital signal processing unit 60. And amplified and A / D converted by the second measurement system amplifier 40b and the A / D converter 50b and supplied to the digital signal processing unit 60. The digital signal processing unit 60 calculates the correlation between the two reference noise voltages supplied through the first and second measurement systems and evaluates the characteristics of the measurement system. During the subsequent temperature measurement, the switch 30 connects only the resistance sensor 10 to the first and second measurement systems, measures the thermal noise voltage generated by the resistance sensor 10, and further compares the measurement result with the correlation described above. By correcting based on this, the temperature of the resistance sensor 10 is accurately measured.

  Although the basic block configuration and operation itself of the thermal noise thermometer 1 described above are the same as the conventional one, the configuration of the reference noise voltage generator 20 in the thermal noise thermometer 1 of the present embodiment is as shown in FIG. There is a feature in that the configuration of the embodiment shown in FIG. 2 is used instead of the conventional configuration.

  FIG. 2 shows a block diagram of an embodiment of a reference noise voltage generator according to the present invention. As shown in the figure, the reference noise voltage generator 20 of the present embodiment includes a pulse multiplier 21, a pseudo random number generator 22, a Josephson transmission line 23, and a plurality of superconducting quantum interference elements ( A SQUID array 24 in which SQUIDs (Superconducting Quantum Interference Devices) are connected in series is mounted on a reference noise voltage source chip 25 that is a single substrate. Here, each component on the reference noise voltage source chip 25 (that is, the pulse multiplier 21, the pseudo random number generator 22, the Josephson transmission line 23, and the SQUID array 24) has high shielding efficiency on the integrated circuit. Since they are connected by short-distance wiring, crosstalk between signals can be effectively suppressed and measurement errors can be reduced. The pulse multiplier 21 and the pseudo random number sequence generator 22 constitute a waveform generation circuit, and the Josephson transmission line 23 and the SQUID array 24 constitute an output circuit.

With the configuration described later, the pulse multiplier 21 has a spectral density control signal supplied from the outside each time a pulse of logic “1” in the pseudo random number sequence supplied from the pseudo random number generator 22 is input. Thus, a magnetic flux quantum voltage pulse in which the magnetic flux quantum number or the level in the voltage axis direction due to the superposition of the magnetic flux quantum voltage pulse is multiplied within a predetermined unit time T ₀ is output. In the present specification, the “flux quantum voltage pulse” is quantized (the time integral is flux quantum Φ = h / (2e) = 2.07 × 10 ⁻¹⁵ Wb) output from a circuit composed of Josephson junctions. Means a voltage pulse that exactly matches an integer multiple of. The spectral density control signal maintains the waveform accuracy by controlling the magnetic flux quantum number included in the output waveform of the pulse multiplier 21 in the time axis direction or by superimposing magnetic flux quantum voltage pulses in the voltage axis direction. It is a signal that changes the voltage spectral density in a programmable manner.

  The pseudo random number generator 22 must be able to calculate the generated pseudo random number sequence (the waveform is determined with high accuracy). Therefore, the pseudo-random number generator 22 is configured to generate an M series, for example, as described later, using superconducting digital circuit technology. By the M series using the shift register, a computable magnetic flux quantum voltage pulse train signal close to white noise having a uniform frequency spectrum can be generated.

  For precision measurement applications, the output circuit must be a Josephson circuit capable of full quantization. In the present embodiment, the output circuit includes a Josephson transmission line 23 and a SQUID array 24. Each of the plurality of SQUIDs constituting the SQUID array 24 is a superconducting closed circuit composed of one or more Josephson junctions and a coil, and is magnetically coupled with a coil constituting the Josephson transmission line 23 with a predetermined mutual inductance. And operate in a cryogenic environment.

  Next, each embodiment of the pulse multiplier 21 will be described. Here, the voltage spectral density of the reference noise voltage generated by the reference noise voltage generator 20 in the thermal noise thermometer must be controlled to correspond to the measured temperature, and the characteristics of the resistance sensor 10 being measured. It is necessary to fine-tune the voltage spectral density in real time following the change. Therefore, the pulse multiplier 21 is required to output a magnetic flux quantum voltage pulse whose voltage spectral density is varied while maintaining waveform accuracy over a wide frequency range.

  FIG. 3A is a block diagram of the first embodiment of the pulse multiplier 21, and FIG. 3B shows the normalized frequency versus gain characteristic of FIG. The pulse multiplier 21A of the first embodiment shown in FIG. 3A is composed of a variable counter 71 and a ring oscillator 72. The ring oscillator 72 outputs, for example, a logical value “1” from the pseudo-random number sequence output from the pseudo-random number sequence generator 22 in which two values of logical values “1” and “0” are randomly synthesized in time series. Oscillation is started when a pseudo-random number is input, and a magnetic flux quantum voltage pulse with a constant amplitude is oscillated and output at a period τ to be supplied to the variable counter 71.

The variable counter 71 counts the magnetic flux quantum voltage pulses supplied from the ring oscillator 72, and generates an oscillation stop signal when the number of counted pulses reaches a number N that is programmable as specified by an external spectral density control signal. The ring oscillator 72 is supplied to stop the oscillation operation of the ring oscillator 72. Here, assuming that the minimum input time interval (bit period) between adjacent logical values “1” of the pseudo random number sequence is T ₀ as shown by the input waveform 73 in FIG. The N magnetic flux quantum voltage pulses are output from the ring oscillator 72 within the time interval T ₀ . Accordingly, the output signal waveform of the ring oscillator 72 is as indicated by 74 in FIG. As an example, τ / T ₀ = 0.0001. The pulse number N is a magnetic flux quantum number obtained by sequentially increasing the magnetic flux quantum voltage pulse in the time axis direction within the minimum time interval T ₀ , and can be set to a desired value programmable by the spectral density control signal from the outside. Can be variably specified.

The normalized frequency versus gain characteristic of the output signal of the pulse multiplier 21A of the present embodiment when the magnetic flux quantum number N is varied is as shown in FIG. , 20, a constant gain characteristic was obtained when the normalized frequency fT ₀ was in the range of 0-100. The normalized frequency fT ₀ is a frequency obtained by normalizing the output signal frequency with the minimum input time interval T ₀ of the input signal. Thus, according to the pulse multiplier 21A of the present embodiment, by controlling the number of magnetic flux quanta included in the output waveform, the magnetic flux quanta in which the voltage spectral density is varied while maintaining the waveform accuracy over a wide frequency range. A voltage pulse can be output.

  Next, a second embodiment of the pulse multiplier 21 will be described. FIG. 4A is a block diagram of the second embodiment of the pulse multiplier 21, and FIG. 4B shows the normalized frequency versus gain characteristic of FIG. The pulse multiplier 21B of the second embodiment shown in FIG. 4A is composed of a pulse distributor 81 and a voltage multiplier circuit 82. The pulse distributor 81 and the voltage multiplying circuit 82 are configured as shown in FIG.

The voltage multiplication circuit 82 is a multi-stage voltage multiplication circuit composed of n amplifiers each having an amplification factor set to 2 ^m (m is 0, 1,..., N−1). The outputs are combined, and the operations of the n amplifiers are controlled by magnetic flux quantum voltage pulses supplied from the pulse distributor 81. With the configuration schematically shown in FIG. 5, the pulse distributor 81 randomly outputs binary magnetic flux quantum voltage pulses of logical values “1” and “0” output from the pseudo random number generator 22 in a time-series manner. For example, when a pseudo-random number having a logical value “1” is input from the pseudo-random number sequence synthesized, the switch is selected so as to obtain a predetermined amplification factor according to the value of the spectrum density control signal from the outside. The input pulses are distributed and supplied to one or more of the n-stage amplifiers constituting the voltage multiplication circuit 82.

  FIG. 6A shows a circuit diagram of an example of a basic cell (one stage) constituting the voltage multiplication circuit 82. The basic cell of the voltage multiplication circuit 82 shown in FIG. 6A is a SQUID 821 that is a closed circuit composed of two coils and two Josephson junctions, coils 822 to 824 connected in series, and connection points thereof. And Josephson junctions 825 and 826 connected between the ground and the coil, and the coil 823 is magnetically coupled to one coil constituting the SQUID 821.

When the magnetic flux quantum voltage pulse ν _in shown in FIG. 6D is input to the coil 822, a current I ₁ shown in FIG. 6C is induced in the coil 823 in the process of passing through the Josephson junctions 825 and 826, The current I ₁ induces a current I _s shown in FIG. 6C in the SQUID 821 through magnetic coupling. Then, the current flowing through the two Josephson junctions in the SQUID 821 increases, and each Josephson junction outputs magnetic flux quantum voltage pulses of voltages ν _s1 and ν _s2 shown in FIG. 6B. The magnetic flux quantum voltage pulse having the voltage ν _s2 is output as the output pulse ν _out . This is equivalent to outputting one magnetic flux quantum voltage pulse from the SQUID 821 when one magnetic flux quantum voltage pulse is inputted. As shown in FIG. 6D, voltages ν ₁ and ν ₂ are generated in the Josephson junctions 825 and 826, and the voltage ν ₃ is output from the coil 824. The amplifier with an amplification factor of 2 ^k constituting the voltage multiplication circuit 82 has a configuration in which the basic cells shown in FIG. 6A are connected in 2 ^k stages in series.

The voltage multiplication circuit 82 configured as described above amplifies the input magnetic flux quantum voltage pulse by one or more amplifiers, synthesizes the outputs thereof, and outputs the resultant as one multiplication magnetic flux quantum voltage pulse. Therefore, if the minimum input time interval of the logical value “1” of the pseudo random number sequence is T ₀ as shown by the input waveform indicated by 83 in FIG. 4A, the voltage increases within the minimum time interval T ₀ . The multiplier circuit 82 outputs a magnetic flux quantum voltage pulse in which the input magnetic flux quantum voltage pulse is superimposed in the voltage axis direction (the amplitude is multiplied) as indicated by 84 in FIG. The synthetic amplification factor (ratio of amplitude multiplication) N with respect to the magnetic flux quantum voltage pulse can be variably specified to a desired value by the above-described spectral density control signal.

The normalized frequency versus gain characteristic of the output signal of the pulse multiplier 21B of this embodiment is as shown in FIG. 4B, and the value N variably controlled by the spectral density control signal is 2, 5, 20 , 75, a constant gain characteristic was obtained when the normalized frequency fT ₀ was in the range of 0-100. The normalized frequency fT ₀ is a frequency obtained by normalizing the output signal frequency with the minimum input time interval T ₀ of the input signal. Thus, according to the pulse multiplier 21B of the present embodiment, it is possible to output a magnetic flux quantum voltage pulse whose amplitude is multiplied while maintaining waveform accuracy over a wide frequency range.

  Next, a third embodiment of the pulse multiplier 21 will be described. FIG. 7A is a block diagram of a third embodiment of the pulse multiplier 21, and FIG. 7B shows the normalized frequency versus gain characteristic of FIG. In FIG. 7A, the same components as those in FIGS. 3A and 4A are denoted by the same reference numerals.

  A pulse multiplier 21C according to the third embodiment shown in FIG. 7A is different from the pulse multiplier 21A shown in FIG. 3A and the pulse multiplier 21B shown in FIG. The magnetic flux quantum voltage pulse output from the ring oscillator 72 is supplied to the pulse distributor 81. The spectral density control signal is supplied to both the variable counter 71 and the pulse distributor 81 and independently of each other.

  In this pulse multiplier 21C, a magnetic flux quantum number multiplication magnetic flux quantum voltage pulse indicated by 86 in FIG. 7A output from the ring oscillator 72 with the magnetic flux quantum number N controlled is supplied to the pulse distributor 81, Here, one or more of n amplifiers constituting the voltage multiplication circuit 82 are obtained so that a predetermined superimposition level (amplification factor) M in the voltage axis direction corresponding to the value of the spectral density control signal from the outside is obtained. The input pulses are distributed and supplied to the amplifiers. Thus, the voltage multiplication circuit 82 amplifies the input magnetic flux quantum number multiplication magnetic flux quantum voltage pulse 86 by one or more amplifiers, synthesizes their outputs, and both the magnetic flux quantum number and the amplitude are multiplied. Is output as one multiplied magnetic flux quantum voltage pulse.

Accordingly, the minimum input time interval between adjacent logical values “1” of the pseudo random number sequence supplied as an oscillation start signal to the ring oscillator 72 is as shown in the input magnetic flux quantum voltage pulse waveform indicated by 85 in FIG. If it is assumed that T _0, the pulse multiplier 21C, both superimposing level M of the flux quantum number N and the voltage axis direction in this minimum time within interval T ₀ is multiplied independently of one another, FIG. 7 ( A magnetic flux quantum voltage pulse indicated by 87 is output to a).

FIG. 7B shows the normalized frequency versus gain characteristic when N and M of the multiplication magnetic flux quantum voltage pulse output from the pulse multiplier 21C of the present embodiment are varied. In FIG. 7B, when the number N of multiplication magnetic flux quantum voltage pulses and the level M in the voltage axis direction are 2, 5, and 20, respectively, the gain is within the range of the normalized frequency fT _{0 from 0} to 100. Certain characteristics were obtained. In the case of N = M = 75, a characteristic that the gain decreases as the normalized frequency increases is obtained, but the degree of the decrease is relatively small. The normalized frequency fT ₀ is a frequency obtained by normalizing the output signal frequency with the minimum input time interval T ₀ of the input signal. Thus, according to the pulse multiplier 21C of the present embodiment, by controlling both the magnetic flux quantum number and the level in the voltage axis direction included in the output waveform, while maintaining the waveform accuracy over a wide frequency range, A magnetic flux quantum voltage pulse in which both the magnetic flux quantum number and the level in the voltage axis direction are variable can be output.

  Next, the pseudorandom number generator 22 will be described. FIG. 8A is a block diagram of an example of the pseudorandom number generator 22, FIG. 8B is an output voltage versus time characteristic diagram of FIG. 8A, and FIG. ) Shows an output spectral density vs. frequency characteristic diagram.

  The pseudo random number generator 22 must be able to calculate the generated pseudo random number sequence (the waveform is determined with high accuracy). Therefore, as shown in FIG. 8A, the pseudo random number generator 22 used in the present embodiment exclusively outputs the seven-stage shift register 91 and the sixth-stage and seventh-stage outputs of the shift register 91 as an example. An exclusive OR circuit 92 which performs a logical sum operation and feeds back to the first stage of the shift register 91, and is generated by a predetermined generator polynomial from the seventh stage of the shift register 91 in synchronization with the supplied clock pulse CK. It is configured as a known M-sequence generator that outputs a sequence.

  Moreover, the pseudorandom number generator 22 configured by the M-sequence generator shown in FIG. 8A of the present embodiment is configured by a superconducting circuit. Since the M series generator composed of a superconducting circuit is known (see, for example, Japanese Patent Laid-Open No. 62-16614), the description of the configuration is omitted. The M-sequence output signal output from the pseudo random number generator 22 shown in FIG. 8A is a magnetic flux quantum voltage pulse having a logical value “1” and a logical value “0” as shown in FIG. 8B. Is a pulse train signal generated randomly and synthesized in time series. Also, the M sequence output from the pseudo random number sequence generator 22 shown in FIG. 8A is a signal whose spectral density is close to constant white noise regardless of the frequency, as shown in FIG. 8C.

  Next, the output circuit of the reference noise voltage generator 20 shown in FIG. 2 will be described. The output circuit is composed of a Josephson circuit capable of complete quantization for precision measurement applications, and here comprises a Josephson transmission line 23 and a SQUID array 24.

The Josephson transmission line 23 transmits the magnetic flux quantum voltage pulse output from the pulse multiplier 21 to the SQUID array 24. FIG. 9A shows a circuit diagram of an example of the Josephson transmission line 23. As shown in FIG. 2A, the Josephson transmission line 23 has, for example, four coils 231 to 234 connected in series, and Josephson junctions 235 to 237 connected between the coil connection points and the ground. It is the structure which was made. In the Josephson transmission line 23, when the magnetic flux quantum voltage pulse ν _in is input to the input terminal shown at the left end of the figure, the input magnetic flux quantum voltage pulse ν _in is converted into Josephson junctions 235 to 237 as shown _in FIG. Are successively passed as magnetic flux quantum voltage pulses ν ₁ , ν ₂ , and ν ₃ , and output as magnetic flux quantum voltage pulses ν _out to the output terminal. Further, as the voltage of the Josephson junctions 235 to 237 is generated by the input magnetic flux quantum voltage pulse ν _in , currents i _{1 to} i ₄ flow through the coils 231 to 234 as shown in FIG.

  The SQUID array 24 generates a magnetic flux quantum voltage that is precisely synthesized by superimposing the magnetic flux quantum voltage pulses input from these transmission lines 23, and outputs the magnetic flux quantum voltage to the outside as a reference noise voltage. FIG. 10 is a circuit diagram showing an example of the SQUID array 24 together with the Josephson transmission line. In FIG. 10, a SQUID array 241 is a SQUID array 24 in which N-stage serially connected SQUIDs composed of two coils and two Josephson junctions are connected. Also, the Josephson transmission line 23A shown in FIG. 10 is an example of the Josephson transmission line 23 in which the Josephson transmission line of the basic cell is connected in N stages in parallel to the input.

In this configuration, when one magnetic flux quantum voltage pulse ν _in is input to the Josephson transmission line 23A, each SQUID generates one magnetic flux quantum voltage pulse. For example, the coil in the Josephson transmission line 23A magnetically coupled to the n-th SQUID of the N stages (n is a natural number of 1 to N) is shown _in FIG. 11B by the magnetic flux quantum voltage pulse ν _in . In _n is induced, and magnetic flux quantum voltage pulses having voltages indicated by ν _n1 and ν _n2 in FIG. 11C are generated in the two Josephson junctions connected to the coil.

In addition, a current I _sn shown in FIG. 11B is induced in the _n-th SQUID through the magnetic coupling by the current In. Thus, since the N-stage SQUIDs constituting the SQUID array 241 are connected in series, the SQUID array 241 outputs an output pulse ν _out shown in FIG. 11A on which N magnetic flux quantum voltage pulses are superimposed.

FIG. 12 is a circuit diagram of another example showing the SQUID array 24 together with the Josephson transmission line. In the figure, the same components as those in FIG. 10 are denoted by the same reference numerals, and the description thereof is omitted. In FIG. 12, the Josephson transmission line 23B magnetically coupled to the SQUID array 241 is an example of the Josephson transmission line 23 in which the Josephson transmission lines of the basic cells are connected in series to the N stages. Even in the configuration of this example, the SQUID array 241 can output an output pulse ν _out on which N magnetic flux quantum voltage pulses are superimposed. However, a time delay occurs in the operation of the SQUID at each stage constituting the SQUID array 241, and the output pulse waveform becomes broad.

  According to the reference noise voltage generator 20 of the present embodiment described above, a waveform generation circuit composed of a pulse multiplier 21 and a pseudo random number generator 22, an output composed of a Josephson transmission line 23 and a SQUID array 24. Since the reference noise voltage source chip 25 is connected to the circuit by a short distance wiring with high shielding efficiency on the same substrate, the temperature measurement error due to the crosstalk between the conventional wiring can be greatly reduced. Thus, the measurement accuracy of the thermal noise thermometer can be improved as compared with the conventional one.

  Further, according to the reference noise voltage generator 20 of the present embodiment, the magnetic flux quantum voltage pulse output from the pulse multiplier 21 in accordance with the pseudo random number sequence output from the pseudo random number sequence generator 22 that generates the M-sequence. The voltage spectral density can be changed in a programmable manner according to the spectral density control signal in a considerably short time compared to the conventional case, so that the voltage spectral density is finely adjusted almost in real time following the characteristic change of the resistance sensor 10 being measured. be able to.

  In addition, according to the reference noise voltage generator 20 of the present embodiment, an M sequence that is a pseudo random number sequence that can be calculated by the pseudo random number sequence generator 22 is generated instead of the pulse pattern. The extremely expensive pulse pattern generator used in the voltage generator is unnecessary, and as a result, the cost of the entire thermal noise thermometer 1 can be reduced.

  Further, the power consumption of the reference noise voltage source chip 25 constituting the reference noise voltage generator 20 of the present embodiment is sufficiently small, can be cooled by an inexpensive small refrigerator, and an expensive pulse pattern generator is provided. Coupled with being unnecessary, the cost of the entire thermal noise thermometer 1 can be greatly reduced. Therefore, according to the reference noise voltage generator 20 of the present embodiment, the spread of highly accurate thermal noise thermometers over a wide temperature range can be expected from the features that the time required for measurement is short and the price is low.

  The present invention is not limited to the above-described embodiment. For example, the pseudo random number generator 22 is not limited to the configuration of FIG. 8 such as the number of stages of the shift register, and other than the M series. It may be configured to generate a pseudo random number sequence (however, the waveform of the generated pseudo random number sequence must be determined with high accuracy and be configured of a superconducting circuit).

  The present invention can be used for the advancement of temperature standards and noise standards. Further, the present invention can be used for an ultra-precise thermometer in which a precision resistor having a small temperature coefficient is arranged in a resistance sensor.

DESCRIPTION OF SYMBOLS 1 Thermal noise thermometer 10 Resistance sensor 20 Reference noise voltage generator 21, 21A, 21B, 21C Pulse multiplier 22 Pseudo random number generator 23, 23A, 23B Josephson transmission line 24, 241 SQUID array 25 Reference noise voltage source Chip 30 Switch 40a, 40b Amplifier 50a, 50b A / D converter 60 Digital signal processing unit 71 Variable counter 72 Ring oscillator 73, 83, 85 Pseudo random number sequence 74, 84, 87 Output multiplication magnetic flux quantum voltage pulse 81 Pulse distributor 82 Voltage multiplier 91 Shift register 92 Exclusive OR circuit

Claims (5)

A resistance sensor;
A reference noise voltage generating means for generating a reference noise voltage having characteristics equivalent to those of the resistance sensor;
A measuring means for obtaining a measured value by digital signal processing from two noise voltages respectively passing through two measuring systems;
At the time of calibration, only the reference noise voltage from the reference noise voltage generation means was supplied to the measurement means, and the correlation between the two reference noise voltages passed through the two measurement systems was calculated to evaluate the characteristics of the measurement system. Thereafter, only the measurement voltage from the resistance sensor is supplied to the measurement means at the time of measurement, and the two measurement voltages that have passed through the two measurement systems are corrected using the correlation obtained at the time of calibration to obtain a measurement value. Switch means for determining;
A thermal noise thermometer comprising:
The reference noise voltage generating means is
A pseudorandom number generator configured using superconducting digital circuit technology to generate a computable pseudorandom number sequence;
When the magnetic flux quantum voltage pulse train constituting the pseudo random number sequence is a predetermined random number value, either or both of the magnetic flux quantum number per unit time and the level in the voltage axis direction are variably controlled by an external control signal. A pulse multiplier that outputs a magnetic flux quantum voltage pulse train increased at a rate of
An output circuit for fully quantizing and superimposing the magnetic flux quantum voltage pulse train output from the pulse multiplier, using the Josephson effect, and outputting the reference noise voltage;
Is a structure formed on the same substrate. The output circuit is
A Josephson transmission line for transmitting the magnetic flux quantum voltage pulse train output from the pulse multiplier;
A SQUID array comprising a plurality of superconducting quantum interference elements magnetically coupled to the Josephson transmission line and connected in series with each other;
2. The thermal noise thermometer according to claim 1, wherein the reference noise voltage obtained by fully quantizing and superimposing the magnetic flux quantum voltage pulse train is output from the SQUID array. A reference noise voltage generator for supplying a reference noise voltage to the measurement system at the time of calibration in order to eliminate measurement errors caused by characteristics of the measurement system of the thermal noise thermometer,
A pseudorandom number generator configured using superconducting digital circuit technology to generate a computable pseudorandom number sequence;
When the magnetic flux quantum voltage pulse train constituting the pseudo random number sequence is a predetermined random number value, either or both of the magnetic flux quantum number per unit time and the level in the voltage axis direction are variably controlled by an external control signal. A pulse multiplier that outputs a magnetic flux quantum voltage pulse train increased at a rate of
An output circuit for fully quantizing and superimposing the magnetic flux quantum voltage pulse train output from the pulse multiplier, using the Josephson effect, and outputting the reference noise voltage;
The reference noise voltage generator is characterized in that is formed on the same substrate. The pulse multiplier is
The oscillating means that receives the magnetic flux quantum voltage pulse sequence constituting the pseudo random number sequence and is activated at a predetermined pulse value to generate a pulse having a constant period τ;
Counting means for counting the pulses output from the oscillating means and supplying an oscillating operation stop signal to the oscillating means to stop the oscillating operation when the count value reaches a value specified by the control signal;
The magnetic flux quantum voltage pulse train in which the quantum number of magnetic flux per unit time T ₀ (where T ₀ > τ) of the magnetic flux quantum voltage pulse train is variably controlled by the control signal is output from the oscillation means. The reference noise voltage generator according to claim 3. The pulse multiplier is
A voltage multiplication circuit comprising a plurality of amplifiers connected to each other's output terminals;
The magnetic flux quantum voltage pulse train constituting the pseudo random number sequence and the control signal are input, and when the input magnetic flux quantum voltage pulse train has a predetermined pulse value, it is designated by the control signal among the plurality of amplifiers. A pulse distributor for distributing the flux quantum voltage pulses to one or more amplifiers;
The magnetic flux quantum voltage pulse train in which the level in the voltage axis direction per unit time of the magnetic flux quantum voltage pulse train is variably controlled by the control signal is output from the voltage multiplication circuit. Reference noise voltage generator.

Images