US2156786A - Variable band-pass filter circuits - Google Patents

Description

4 sheds-sheet 1 1 &

INVENTOR JAMES J LAMB w H Qs W May 2, 1939. J. J. LAMB VARIABL BAND-PASS FILTER CIRCUITS Filed July 9. 1957 M Y mxg FWI 1 M v v w 5. V Q $3552 AAAAA vvvvvv ATTORNI-Y y 2, 1939- J. J. LAMB VARI ABLE BAND-PASS FILTER CIRCUITS Filed July 9, 1937 4 Sheets-Sheet 2 /I/IIINAN luw\\ (0 +5 '+'/0 +/5 FROM RE80NANCE w 2 I "SEE figs l I I l I J l0 l5 Z0 25 30 .35 JAMES J. LAMB TOTAL BANDW/DTH K C. BY 6 ATTORNEY--.

J. .1. LAMB 2,156,786

VARIABLE BAND-PASS FILTER CIRCUITS 4 Sheets-Sheet 3 May 2, 1939.

Filed July 9, 1957 70 #2 11 EAMI? I-Z DETEC'TOR \NVENTOR JAMES J. LAMB ATTORNEY variable selectivity Patented May 2, 1939 UNITED STATES PATENT OFFICE 2,156,786 VARIABLE BAND-PASS FILTER CIRCUITS James J. Lamb, West to Radio tion of Delaware Hartford, Conn, assignor Corporation of America,

a corpora- Application July 9, 1937, Serial No. 152,728

4 Claims.

This invention relates to variable band-pass filter circuits and has more particularly to do with a device having full-range superheterodyne selectivity in an intermediate frequency amplifier.

over 20 kilocycles. Furthermore, in order to cope with the wide variety of interference conditions Accordingly, provide a band-pass filter system for use in an continuously 100 cycles to It is another object of my invention to provide In carrying out my invention I have found that the full-range encompassed may be covered in substantially three steps, each capable of giving continuously variable band-width between its minimum and maximum limits. These are Such an assumption is quite reasonable as has my paper entitled Receiver selectivity characteristics published in Q. S. T., May 1935. i

For the highest selectivity range the familiar quartz crystal filter is used; for the medium'range I prefer to use what has been termed a Transfilter unit in place of the quartz crystal; and for the broadest'range vari- The Transfilter unit has been article entitled A described in my new I. F. coupling system for published in the April 1937 This Transfilter unit comof a transformer and an Like the quartz filter it is vbines the properties electric wave filter. electro-mechanical this bar has mechanically coupled to small plates of Rochelle salt.

Qualitatively the Transfilter displays the same electrical characteristics found in quartz resonance; that is, essentially capacitive reactance at most frequencies with a sudden deviation from this reactance at the frequency of resonance of Hence it may be used in bridge circuits in the same manner as a quartzfilten However, the Rochelle salt driver steel bar has somewhat higher mechanical damping with the result that the filter 15 not as sharp as quartz; hence the using such a filter in the intermediate band-Width of selectivity; The Transfilter, on the other hand, provides much higher selectivity than is available with the use of audio tuned circuits.

lished in I. R. E., April 1922 and covered by United States Patent No. 1,450,246.

The features of novelty of my invention are believed to be defined by the appended claims. The invention itself, however, as to its structural details and mode of operation, will be best understood upon reference to the accompanying drawings when viewed in the light of the following description. In the drawings:

Figure 1 shows a circuit diagram of a superheterodyne receiver embodying an intermediate frequency amplifier and a combination of filter systems which may be selectively used for varying the band-width;

Fig. 2 shows a selectivity curve appropriate to the operation of the circuit arrangements of Fig. 1;

Fig. 3 shows curves of total band-width obtainable by the operation of the circuit of Fig. 1;

Fig. 4 shows details of a modified arrangement of an impedance matching Transfilter;

Fig. 5 shows for purposes of comparison a variation from the circuit arrangements of Fig. 4;

Fig. 6 shows a selectivity curve appropriate to the circuit arrangement of Fig. 4;

Fig. 7 shows total band-width curves for resistance variation;

Fig. 8 shows a modification of the invention in which a band-pass circuit is used comprising two Transfilters in parallel, and

Fig. 9 shows maximum and minimum selectivity curves in respect to the circuit arrangement of Fig. 8.

Referring first to Fig. 1, I show therein an illustrative embodiment of the invention which includes a first detector stage tube I, a heterodyne oscillator 2, a first intermediate frequency amplifier 3, a second intermediate frequency amplifier 4, a second detector 5, and interconnecting circuit arrangements presently to be described in more detail. The transformer 6 coupling the antenna circuit or radio frequency amplifier to the first detector I may be of any conventional type. Transformers 1, 8 and 9 are preferably of the airtuned, air-core type. The second detector 5 is preferably a double diode. The audio frequency amplifier is connected to the output leads l0.

, The two intermediate frequency transformers 8 and 9 are adjusted for a relatively broad frequency characteristic to provide a fair amount of tolerance near resonace thereby to accommodate minor deviations in frequency of the several piezo-electric or Transfilter devices used.

As shown in the drawing, the first intermediate frequency stage is preceded by an arrangement of alternative filter circuits. For the band-width of sharpest selectivity the piezo-electric crystal H is switched into circuit between one terminal of the secondary of the transformer I and the adjustable coupling condenser l2. The switching arrangement includes a double-pole, triple-throw switch indicated generally at l3. The crystal II is connected in circuit with the switch arms if in the position shown. In the intermediate position of the switch arms a Transfilter unit I4 is substituted for the quartz crystal. The steel bar l5 of the Transfilter unit is grounded. In the third position of the switch l3 both the quartz crystal I I and the Transfilter unit are cut out and tuning is obtained for the broadest selectivity band by merely short-circuiting the switch points Ill.

The variable capacitor l1 in shunt with the secondary of the transformer I is used to control the band-width or selectivity. Another variable capacitor 18 in opposition to the filter units controls the rejection of interfering signals. In varying the band-width by adjustment of the parallel tuned impedance as indicated in the diagram, maximum band-width (minimum selectivity) occurs with this circuit tuned to crystal resonance and decreasing band-width (increasing selectivity) occurs as the parallel tuned circuit becomes reactive on either side of resonance. With the impedance matching which this circuit provides, the over-all carrier wave gain of the receiver is practically the same with the input circuit adjusted for optimum (medium high) selectivity as it is with the crystal shorted out and the input circuit adjusted for maximum straight superhet gain. Either side of this point the over-all gain decreases slightly, both toward maximum band-width and toward extreme minimum band-width.

The Transfilter unit I4 is of fairly low impedance and accordingly cuts the gain of the input amplifier or first detector when fed directly from its anode Hi. When the Transfilter is used by setting the switch |3 to its intermediate position, selectivity is varied by the same method as when the crystal filter is used; that is, by variation of the parallel tuned impedance which constitutes the input to the divided circuit. Although the selectivity control condenser l1 settings are not exactly the same as for a quartz crystal filter of corresponding frequency, minimum selectivity occurs with the input circuit resonant to the Transfilter frequency and increasing selectivity occurs as the input circuit is tuned either side of resonance. The resonance setting (maximum band-width) comes at lower tuning capacitance with the Transfilter than with the crystal because the Transfilter capacitance to ground is apparently greater by as much as 10 micro-micro-farads. The adjustment is still well within the range of the condenser, however.

Many of the details of the circuit arrangements shown in Fig. 1 are self-explanatory but in some respects they follow the teachings of my Patent No. 2,054,757, granted September 15, 1936, reference to which may be made for further explanation of these details.

The range of selectivity obtainable with the crystal filter circuit is shown by curve A in Fig. 2 as a maximum and by curve B when the crystal is adjusted for minimum selectivity. Between these two curves A and B a characteristic shading has been introduced for representing the band-width variation under different adjustments. A different shading has been used to show the band-width variation under the different adjustments between a maximum selectivity limit 0 and a minimum selectivity limit D when the Transfilter is used. Curve E is for the transformer-coupled selectivity characteristic of the intermediate frequency amplifier without either filter, that is, when the switch I3 is moved to the extreme right for short-circuiting the contact points Hi.

It is interesting to note that the selectivity range with the Transfilter practically continues on from where the crystal range reaches its broadest. This is illustrated even more clearly by the total band-width curves of Fig. 3 which are plotted from the same experimental data as used in plotting the curves of Fig. 2. The principal difference between the selectivity of the crystal filter at its broadest and the Transfilter at its sharpest is that the Transfilter selectivity characteristic is somewhat broader near reso- Transfilter selectivity nance. width. Under practical working conditions the circuit arrangements of my invention as herein dehave been shown to be well giving a slightly greater effective bandsponse characteristic for elimination of a particular interfering carrier evenwithin the normal band-width range.

range carries on from this point to a band-width sufiibiently great for speech reception with entirely adequate fidelity. In

fact, the Transfilter selectivity at its broadest is generally useful for broadcast program reception,

providing fidelity fully as good as those customary with the average broadcast receiver.

This rangeis especially adapted to short-wave broadcast reception where it is desirable to conference which is aggravated by the fading so characteristic of these frequencies. True highfidelity reception is practically never feasible on the high-frequency bands, and considerable highquency band-width control with the Transfilter in much more satisfactory fashion than it can be obtained by an audio-frequency tone control accomplishes the same eiTect of reducing the noise but does so without introducing the amplitude distortion which may occur with audio-frequency tone control. Furthermore, it does the job prior to the second detector and removes noise and adjacent-channel sideband components before they have a chance to intermodulate with the desired signal in the second detector to produce low-frequency audio components which cannot be removed byaudio-frequency filtering subsequent to detection.

A matter of some importance in judging the relative merits of selective intermediate frequency circuits, in addition to their contribution of selectivity, is their effect on the over-all gain and effective sensitivity. In connection with crystal filters, for instance, there is considerable divergence of opinion as to whether this or that particular arrangement is the better in point of how little it reduces the gain of the receiver. In my experience, the impedance-matching crystal tion. This refers particularly to the carrier wave v by adjustment provided,

On the other hand, the

output is less, as is also the effective sensitivity of the receiver. I

When using the Transfilter, the gain is also negligibly afiected .as compared to the straight superhet gain. In practice, differences of a few decibels in over-all gain are readily compensated of the receivers gain control of course, the receiver has a proper margin of surplus amplification to start with. This should be true with any good receiver having a two-stage intermediate amplifier;

Of more importance than gain is the effective sensitivity of the receiver. This effective sensitivity is by no means. a simple matter of how much amplification the receiver has. It is, rather, a matter of signal-noise ratio. It is best expressed in terms of the receivers noise equivalent. The noise equivalent is the signal input required to give signal power output equal to the noise power output. The noise concerned is the receiver "hiss noise, which would be the lowest possible noise background under ideal receiving conditions. The noise equivalent will be determined primarily by the signal-noise ratio at the input of the receiver but will be afiected by the subsequent selectivity because the noise power output is generally reduced in proportion to the reduction in effective bandwidth of the receiver.

In tests which have been made with the filter system of my invention it has been observed that a considerable improvement was obtained upon increasing the selectivity. In the case of carrier wave reception with the crystal filter at maximum selectivity, for instance, the sensitivity is about 700 percent of the straight superheterodyne sensitivity, while the speech and music sensitivity with the Transfilter'set for sharp band selectivity is raised to over 300 percent. The broad band selectivity adjustment of the crystal filter is likewise raised to over 300 percent.

In the range of adjustment of the selectivity or band-width control with these circuits, the resonance frequency of the crystal filter varies but a few cycles. This variation is so small that if the .Ignal is first. tuned in with the crystal set selectivity.

Referring now to Fig. 4, I show an arrangement whereby the selectivity may be adjusted by varying the resistance in the ground connection for the Transi'ilter. This arrangement including a potentiometer it provides an impedance matching circuit. When the resistance is entirely out out the circuit 1. In this position maximum selectivity is obtained if the condenser C1 is adjusted for slightly higher capacitance than the resonance setting.

The impedance adjustment between the Transfilter and ground as shown in 5 is found to be suitable when the transformer secondary has a low impedance instead of the divided capacitance step-down arrangement used in Fig. 4. The circuit arrangement of Fig. 5 is, however, less flexible thanthat of Fig. 4 and, in fact, when used with a crystal filter in place of the Transfilter the circuit becomes one of the fixed selectivity type.

Fig. 6 shows a selectivity curve, the datafor which was plotted from tests made with the circuit arrangement shown in Fig. 4. .The curves for zero resistance and for 2500 ohm resistance are not shown since they practically coincide with the 1000 ohm curve. The most interesting feature of these selectivity curves is the notch which appears with 20,000 ohm resistance. This double-hump effect indicates the equivalent of over-coupling with a transformer. As compared to the selectivity curves of Fig. 2, it is apparent that increased resistance tends to broaden the nose of the selectivity characteristic less effectively, while the skirts of the curves spread out more rapidly. They also show that the selectivity characteristic is generally less symmetrical with resistance variation than with variable impedance control. The curves of Fig. '7 show the total bandwidths for the various values of the resistance.

The gain of the circuit falls off somewhat more rapidly with increasing bandwidth as compared to the gain variation with impedance control of selectivity, although the loss is not especially noticeable in practice. On the whole, adjustable impedance control of selectivity appears to be preferable to resistance control with the Transfilter, just as it has been found to be preferable with the quartz crystal filter.

An interesting band-pass type of selectivity characteristic was obtained with two similar Transfilter units connected in parallel in the circuit of Fig. 8. Except for the additional unit, the circuit is identical with that of Fig. 1. The two units had the same rated frequency of 465 kilocycles and actually differed only 200 cycles in resonance frequency. The band-pass curve of A of Fig. 9 was'obtained with the band-width control condenser C1 critically adjusted so that the same output was obtained on both humps with constant signal input. The mid-frequency of this selectivity curve is approximately 1.2 kilocycles lower than the maximum-selectivity curve obtained with the input condenser C1 ad-.

justed for slightly greater capacitance than the broad-band adjustment. The greater broadening of the selectivity curve near resonance is especially desirable in broadcast program reception, although the over-all carrier wave gan with this circuit is practically the same as with a single unit.

Among other variations which I have tried, a particularly interesting one is use of a variableselectivity Transfilter and a quartz crystal filter of the same type in cascade; that is, the crystal filter circuit as the coupling element between the first detector and the first intermediate frequency amplifier, and the Transfilter as the coupling element between the first and second intermediate frequency amplifiers, provision being made to switch either one in and out. A notable improvement with the Transfilter adjusted for medium selectivity is that the crystal filter selectivity characteristics are steepened in the skirts. While such cascade filters require fairly close tolerances in the resonance frequencies of the Transfilter and crystal, there appears to be no great difficulty in meeting the requirements with production types. The fact that the Transfilter frequency can be shifted over a range of a few hundred cycles, by tuning the input circuit above or below resonance, aids in accomplishing close alignment. Tests on three sample production-type Transfilter units have shown a maximum resonancefrequency difference of 380 cycles, the variation being a plus or minus 200 cycles or less from the average.

Further interesting and useful selectivity characteristics are obtained with two variable-selectivity crystal filter circuits similarly in cascade. With one filter adjusted for minimum selectivity and the other for optimum selectivity, for instance, independent rejection control in carrier wave reception makes it possible to eliminate two interfering heterodynes of different frequencies, whether both are on the same side of resonance or on opposite sides of resonance. The crystals may differ 100 cycles or so in frequency without appreciably impairing operation, it has been found. In fact, such a difference actually may prove advantageous, since it gives a band-pass characteristic in the region near resonance.

Fig. 9 shows maximum and minimum selectivity curves obtained with the band-pass filter circuit of Fig. 8. The mid-frequency of curve A is approximately 1.2 kilocycles lower than the resonance frequency of curve B. Curve C is the straight superheterodyne selectivity curve without the filter. The selectivity curves are otherwise self-explanatory.

In carrying out my invention it will be apparent that various modifications may be made by those skilled in the art without departing from the spirit and scope of the invention itself. It is to be particularly understood that the curve diagrams are in themselves merely exemplary of certain circuit arrangements and that where the values of different components in these circuits are varied, different curve diagrams would be produced. The practical application of the invention will be particularly appreciated where systems of the type herein shown are adopted for short-wave reception and other amateur use. The scope of the invention is, however, by no means limited in that manner.

I claim:

1. An electric wave filter system comprising a plurality of selectively operable band-pass circuits one of said circuits including a quartz piezo-electric device, another of said circuits including a mechanically resonant element in association with input and output piezo-electric elements of Rochelle salt, means including a variable impedance effectively in series with a selected one of said circuits for varying the frequency band width of said filter system, and means in shunt with said selected one of said circuits for applying thereto a phase displaced voltage whereby undesired signal frequencies are suppressed.

2. An intermediate frequency amplifier stage in combination with a filter system as defined in claim 1 and including switching means for selectively inserting a particular one of said bandpass circuits in series with said amplifier stage.

3. In an electric wave filter for a superheterodyne receiver, the method of varying the bandpass width and, conversely, the selectivity, which comprises feeding the energy to be filtered through one or another of several paths having different selectivity characteristics, the characteristic of one of said paths being due primarily to piezo-electric action, and the characteristic of another of said paths being due to a combination of mechanical resonance and piezo-electric action, subjecting said energy to a series impedance, varying said impedance between a predominantly inductive and a predominantly capacitive value, thereby to tune the same on either side of the resonant frequency of said energy and Width being filtered.

4. An electric wave filter comprising an input circuit, an output circuit, circuit means comprising a selective electro-mechanical transducer having piezo-electric input and output coupling elements for transferring energy from said input to said output circuits, 2, variable impedance undesired signal frequencies are suppressed.

JAMES J. LAMB.

Images