US2462452A - Variable gain amplifier - Google Patents

Description

Feb. 22, 1949. L. R. YATES 2,462,452

VARIABLE GAIN AMPLIFIER Filed March 12. 1945 0 B our/ ar INVENTOR J0 422975? A? K4755 /.0 90 --80 70 60 -50 -40 By W 0. B. //YP07" AGENT Patented Feb. 22, 1949 VARIABLE GAIN AMPLIFIER Lester R. Yates, Baltimore, Md, assignor to Bendix Aviation Corporation, South Bend, Ind, a

corporation of Delaware Application March 12, 1945, Serial No. 582,325

4 Claims. (01. 179-171) This invention relates to electric wave amplifying apparatus and more particularly to such an apparatus capable of improving the apparent signal-to-noise ratio when intelligence of varying level is transmitted in the presence of background noise.

To secure the greatest effective transmission range in apparatus for phonic communication by modulated carrier waves, it is endeavored to, at all times during the transmission, insure substantially complete modulation of said carrier Waves. In the presence of signals of varying intensity, such as arise in a speech transmission system when the operator speaks at different voice levels or at different distances from the microphone, this goal is attained only when the gain of the speech amplifying apparatus is continually adjusted to maintain the desired depth of modulation. To fulfill this purpose, it is customary to provide in the speech amplifier, an automatic gain control system wherein an increase in the output level decreases the transmission gain and vice versa, so that compensation is introduced for the differing input levels encountered in service. Apparatus of this type is generally referred to as a compressor and the function which it performs is termed compression, denoting the compression of the range of levels at the output as compared with the spread in energy levels at the input secured through the gain control action.

When operating in a quiet location this apparatus is quite satisfactory but in many of the most important uses for this type of communication, such as equipment utilized in connection with operation of aircraft and for communication on railroad vehicles and the like, the voice transmission must be carried on in the presence of a high background noise level. In a typical example, the normal background level may be 45 db. above zero level and the total level in the presence of speech may rise to 60 db., a total differential of db. When a signal of this nature is impressed on an amplifier having an effective compressison characteristic the amplifier gain is brought up through action of the gain control system in the intervals between words and syllables so that the sound level output is substantially unaltered. As a result, the person receiving the communication hears the voice against a rising and falling background level in which the background noise fills in all gaps in speech transmission, and the intelligibility of speech transmission is thus seriously impaired.

Accordingly it is a primary object of the invention to provide new and novel means for delivering a substantially constant intelligence level Without diminishing the overall intelligibility of the transmission system.

Another object of the invention is to provide new and novel means for modulating a transmitter to substantially its full capacity in the presence of varying intelligence levels without permitting extraneous noise to fill in the intervals between the bursts of intelligence.

Yet another object of the invention is to provide new and novel means for improving the si nal-noise ratio delivered at the output of an amplifier having a compression characteristic.

A further object of the invention is to provide new and novel amplifying apparatus having an expansion characteristic at low signal levels and a compression characteristic at elevated levels.

Other objects and advantages of the invention will in part be disclosed and in part be obvious when the following specification is read in conjunction with the drawings in which:

Figure 1 is a schematic diagram of an amplifier embodying the principles of the invention.

Figure 2 is a graphic portrayal of the output characteristic of the new amplifier arrangement as contrasted with the conventional amplifier having a linear input-output characteristic.

Generally speaking the apparatus herein described attains the desired objects by the use of an amplifier whose input-output characteristic is expanded in the low signal level region and compressed in the high signal input region through the use of an expansion rectifier operative with low signal inputs and of a compression rectifier operated at high signal inputs with further means whereby the compression rectifier acts upon the input to the expansion rectifier to eliminate expansion action at the high signal level region.

Referring now to Figure 1 there is shown the input transformer ti! whose input terminals may be connected to the output of suitable transducing means, such as a microphone or phonograph pick-up. The secondary of the transformer H] is connected to the potentiometer l2 having one terminal grounded at it of the movable tap it. The signal voltage between the tap It and ground is impressed on the control grid 58 of the vacuum tube 26 of the pentode type having a cathode 22 connected internally to the suppressor grid 26 and to ground through parallel resistor 26 and capacitor 28. Excitation for the space charge or screen grid is derived from the positive terminal of the source 32, having its negative terminal connected to ground, through the dropping resistor 3t grounded for alternating currents at the grid end by capacitor 36. The anode 88 of the tube 28 is connected through the primary of transformer 48 to the positive terminal of the high voltage source 32 and the signal current components appearing within the tube 28 as a result of the action of the signal voltages on grid I8 produces across the primary of transformer 48 signal voltages appearing, by virtue of mutual induction, across the center tapped secondary of transformer 48 having its outer ends connected to the control grids 42 and 44 of the variable gain amplifiers 46 and 48, which are preferably of the variable-mu type to secure a smooth variable gain action. The anode 58 of tube 46 and the anode 52 of the tube 48 are connected to the outer terminals of the center tapped primary winding of output transformer 54, the terminals'of whose secondary may be connected to any desired output load. The positive potential required for the anodes 58 and 52 is supplied by the connection of the positive terminal of the source 32 to the center tap of the primary of transformer 54 and excitation for the space charge grids 56 and 58 is derived from the same point through the screen dropping resistor 68 shunted for filter purposes by the capacitor 62.

The source 82 is connected in parallel with a voltage divider including a fixed resistor 64, a resistor 68 having an adjustable tap thereon, a resistor 68, also having an adjustable tap thereon, and fixed resistor I8. The cathodes I2 and I4 of the tubes 46 and 48 are connected internally to the respective suppressor grids located between the screen grid and the anode of the individual tubes and are also connected together to the tap 68 on potentiometer 68. The direct current grid for the adjustable gain tubes 46 and 48 is obtained by the connection of the center tap of the secondary of transformer 48 to ground through resistor l6 and that portion of the potentiometer I8 lying between the tap I9 and the ungrounded terminal 88 of resistor I6. The operating control grid bias for the variable gain tubes 46 and 48 is thus seen to be equal to the sum of the voltage appearing between tap 68 and ground, the voltage across resistor I6 and that appearing across the tapped off portion of potentiometer I8.

The gain control action producing the desired operating characteristics results from the voltages appearing across resistor I6 and potentiometer 18, whose derivation will now be discussed in detail. In addition to the output Winding, transformer 54 also carries an auxiliary center tapped secondary winding 55 supplying the compression rectifier tube 85 of the double diode type. The diode anodes 82, 84 are connected respectively to the opposing terminals of the compression rectifier secondary 55 and the center tap of this secondary is connected to the ungrounded terminal 88 of resistor I6 shunted by capacitor 11. The cathode 86 of compression rectifier 85 is connected to the tap 6! on the delay control potentiometer 66. The cathode 86 is biased positively to ground by an amount equal to the voltage between tap 61 and ground and hence rectification by the tube 82 cannot occur until the signal level in the output transformer exceeds the level required to develop a rectifier secondary peak voltage equal to the relay bias. When this current flow occurs the terminal 88 becomes negative with respect to ground and a negative bias is thus impressed on the variable- mu tubes 46, 48 reducing their gain and tending to maintain the output level constant. The tubes 46 and 48 are connected in push-pull relationship to avoidthe :alleled by the bypass capacitor 96.

4 distortion in the output which would otherwise occur as a result of operating them over the curved portion of the characteristic encountered in the low Gm region. As will be apparent, this operation is that encountered in most amplifiers having a compressor characteristic, which may be defined as a characteristic in which the output volume range is less than the input volume range. A portion of the signal appearing in the anode circuit of the input amplifier 28 is impressed via capacitor 81 on the control grid 88 of the expansion amplifier 98 having a cathode 92 connected to ground through the bias resistor 94 par- 7 The anode circuit of the tube 98 is completed by the connection of the anode 98 to the positive terminal of the source 32 through the primary of the expansion rectifier transformer I88 having a center tapped secondary with the opposing ends thereof connected to the diodes I82 and I84 of the expansion rectifier tube I86. The cathode I88 of the tube I86 is connected to the center tap of the expansion rectifier secondary winding through the potentiometer 18,. which is shunted by the filter capacitor II8 whose value is selected to insure the desired time constant. The action of expansion rectifier I86 produces across potentiometer 18 a voltage of the indicated polarity, the cathode terminal being positive. The grid circuit of expansion amplifier 98 is completed by the connection of grid leak resistor II2 to the terminal 88 of resistor I6. The negative bias developed during operation of the compression rectifier is thus applied to the control grid 88 of the expansion amplifier 98 to reduce the gain of this tube and with it the voltage developed across potentiometer I8 in the presence of compression action.

The operation of the circuit may now be readily understood. The signal energy is passed through the amplifier 28 in cascade with the push- pull amplifiers 46 and 48 in the conventional manner. As the input signal level increases a positive bias is developed as a result of the signal voltage impressed on the expansion rectifier diode I86, which positive bias is applied to the control grids of the variable- mu amplifiers 46, 48 to provide an output volume range exceeding the input volume range for the low inputs at the moment assumed. The bias on the cathode 86 of compression rectifier prevents rectification in this tube and no voltage appears across resistor 16.

If the signal input is further increased, the output voltage increases also until its peak value exceeds the delay bias applied to cathode 86. Operation of the compression rectifier 85 is now initiated, developing a negative voltage at the ungrounded terminal 88 of resistor I6. This voltage is impressed on the control grids of push- pull amplifiers 46 and 48, tending to reduce the gain thereof and also appears on the control grid of the expansion amplifier 88, thereby reducing the input to the expansion rectifier I86 and the positive bias voltage produced therefrom. 'These two effects combine to rapidly reduce the gain of the amplifiers and maintain the output level substantially constant despite variations in the input energy level. a V

The input-output characteristic resulting from this circuit configuration is shown in Figure 2 as the curve H4, while that of theconventional amplifier is illustrated by the straight line H6. The conventional amplifier has a linear inputoutput relation which is to say that for every decibelincrease in input 'thereis a .decibel in-.

aeez'ase crease in output. As shown by the curve N4 the amplifier described herein has an expanded input-output characteristic until the output level reaches zero decibel so that in the range of input variations not bringing the output to the zero decibel level the output changes an amount more than one decibel with each decibel increase in input. At the zero level in the illustrated eX- ample chosen, the compression rectifier becomes active and there is but negligible further change in output with continued increase in input voltage. Assuming a signal plus noise level, at the transducer output, of '70 db. and a noise level of -80 db., it is seen that in the conventional amplifier the noise level between words is db. while the total energy level during the word is zero db. In the case of the amplifier shown herein, the level during intelligence transmission is 0 db. and that during the intervals between words is -25 db. which effectively improves the signal-to-noise ratio by db. Should the speech intensity increase to 60 db. the compressed portion of the characteristic prevents the output from rising in an amount sufiicient to appreciably overload the modulation stage of the transmitter or other apparatus receiving energy from the output of the amplifier.

In the operation of an amplifier of this type it is desirable to adjust the gain control to place the normal signal-plus-noise level at the intersection between the compression and expansion portions of the amplifier input-output characteristic.

Following is a simple procedure for setting up the amplifier to provide the desired performance. The potentiometer 66 is adjusted to a voltage corresponding with the amplifier output level in which it is desired that the compressor should become operative. With the expansion rectifier tube tilt and compressor rectifier tube 85 removed from their sockets and with a normal input signal to the amplifier, potentiometer 68 is adjusted to increase the grid bias on variable gain amplifiers Q6 and it and thereby reduce the amplifier gain by an amount corresponding to the amount of expansion desired. The expansion rectifier tube W6 is then reinserted in its socket and potentiometer l8 adjusted to bring the output level back to its original value. The positive bias voltage developed by the input signal across the tapped off portion of potentiometer 18 now cancels the bias voltage from tap 69 to ground so that at normal input level, the amplifier sensitivity is normal; while at low input levels, the voltage developed across the tapped ofi portion of potentiometer i3 is not sufiicient to cancel this bias and the amplifier gain is thereby reduced.

In input levels in excess of those required for normal output, the compression circuit biases the grids of amplifiers 46 and 48 negatively and reduces their gain to maintain a substantially constant output level. Due to the application of compression voltage to the expansion amplifier 9d the negative voltage developed by the compressor overrides the positive voltage developed by the expander and the output level is maintained nearly constant for considerable increases in input level. The time constant of R. C. circuits 18, I I0 and it, ll may be made relatively short, of the order of e second, to permit the amplifier gain to fall off between words during speech transmission and so prevent background noise from being heard during these intervals.

Listening tests conducted on an amplifier of this type indicate psychological reduction in background noise equal to the amount of expansion us'e'd:..i These tests show no adverse effect-on speech quality when expansions on the order of 29 25 db, are used. When microphones must be used in 'noisylocations, a considerable improvement in the intelligibility of the transmitted signal is made possible with the amplifier described herein havinga combined expansion-compression characteristic in which one or the other effect predominates at differentpredetermined operating levels.

The tubes shown throughout in the schematic diagram illustrating this embodiment of the invention are of the thermionic cathode type and the circuits for energizing the associated heaters have been omitted for the purpose of simplicity in presentation, it being understood that any of the conventional means for exciting them may be used.

It will be obvious that many changes and modifications may be made in the invention Without departing from the spirit thereof as expressed in the foregoing description and in the appended claims.

I claim:

1. In signalling apparatus, an amplifier including a variable gain electric discharge device, means for impressing signal energy on the input of and deriving signal energy from the output of said device, a first rectifying means, means amplifying said input energy and applying it to said rectifying means, a second rectifying means responsive to said output energy, means applying biasing potential to said second rectifying means, means applying the output of said second rectifying means to said amplifying means in a manner to reduce the output of said first rectifying means, and means causing the outputs of said rectifying means to effect the gain of said electric discharge device in opposite senses.

2. In a signal amplifying channel a variable gain electric discharge device, a first rectifying circuit, means amplifying energy from a portion of said channel preceding said device and applying it to said rectifying circuit, said rectifying circuit being connected to said device in a manner such that flow of current therethrough increases the gain of said device, a second rectifying circuit responsive to the output energy of said device, means applying a biasing potential to said second rectifying circuit, and a circuit coupling said second rectifying circuit to said amplifying means and to said device in a manner such that flow of current through said second rectifying circuit decreases the gain of both said amplifying means and said device.

3. In a signal amplifying channel a variable gain electric discharge device, a first rectifying circuit, means amplifying energy from a portion of said channel preceding said device and applying it to said rectifying circuit, a second rectifying circuit responsive to the output energy of said device, circuit connections applying to the input of said device a positive potential which is a function of current flow through said first rectifying circuit and a negative potential which is a function of current flow through said second rectifying circuit, circuit connections applying said negative potential to the input of said amplifying means, and a source of bias potential connected to bias said second rectifying circuit.

4. In a signal amplifying channel a variable gain electric discharge device, a first rectifying circuit responsive to energy from a portion of said channel preceding said device, a second rectifying circuit responsive to the output energy of 7 8 said device, circuit connections applying to the input of sand. devlce a. positive potential which n is function of current flo through said first The following references are of record in the rectifying circuit and a, negative potential which file O this P is a function of current flow through said second 5 UNITED STATES PATENTS rectifymg circuit, circuit connections applying said negative potential to said first rectifying cir- Number Name Date cuit to diminish the output thereof and a source 1,998,617 Hammond, 23, 1935 of bias potential connected. to bias said second 2,250,559 Weber July 29, 1941 rectifying circuit 10 3 5 Holst M y 1942 LESTER Rh E 2283 241 Van COtt May 19, 1942

Images