Automatic Gain Control.

There have been many ways to control the gain of radios over the years and will discuss several of the important ways starting close to the beginning of the AC radio.  While I may not cover all of the ways used for control but will cover the important ways and will discuss some of the reasons for their use.I plan to cover the 1920’s to at least to the present day volume control plus AGC (Automatic Gain Control) gain control – this would be the last of the tube radios. I plan to provide a simple but accurate analysis of the automatic gain control.  Another proper name for AGC is AVC, or Automatic Volume Control, but here we will generally use the term Automatic Gain Control. To be accurate I believe that the term Automatic Volume Control is a more proper name if you are interested in marketing. I tend to use th?e term AGC, but the terms are used interchangeably. I have never seen the AGC analyzed before. The analysis can be very tedious without some simplifying assumptions. I have made these assumptions, but they are realistic.  They also give good results without complex equations.I started repairing radios in 1943 and very few of the sets brought into our shop were TRF’s. I repaired them but did not pay attention to the gain controls nor was I was I concerned that they did not have AGC. Nearly all the sets that I worked on were similar circuit wise to those of today. All the Superhetrodynes that I worked on had AGC. When I started my work many of the sets did not have a loop antenna.
This note will discuss the gain control of the TRF and continue to the gain control of the superhetrodyne. It took many years before we got to the modern superhetrodyne gain circuits. I believe that by 1938 a good percentage of our sets had AGC and many of these sets had loop antennas. The automatic gain control was invented long before 1938 but it took time to get it into the usual configuration where all the circuits were basically the same and reasonably understandable. The sameness was a result of the improved operation plus the cost of the circuits, and components. Of course, better performance could be obtained by circuit modifications, with a usual increase in cost.  However, a cost increase is not necessarily required - improvements in components or cleaver design could at times actually reduce cost.
Harold A. Wheeler invented the Automatic Volume Control, and I believe -correctly used that term - around 1927. The Philco 95 came out with the automatic volume control in 1929 and it was a big hit.  The set used a 27 tube arranged as a diode for the AGC detector. Wheeler started thinking about the automatic volume control (AVC) around 1925. The AVC controls the gain of the RF circuits. The screen grid tube was invented early -about 1919 and the suppressor grid was added around 1926. These were very important additions to the vacuum tube. They made the use of AGC much easier because it did simplify the design.
The triode tube was a problem because it has a large input capacitance. The“Miller Effect” will cause this capacitance to increase almost directly with the gain of the tube. The gain of the tube is varied accor ding to the AGC voltage and this changes the input capacitance of the tube. The grid of the tube receives its signal from a tuned circuit, and the AGC voltage will then vary the frequency of the tuned circuit when the triode is used with AGC. Invention of the screen and suppressor grid reduced this capacitance dramatically and thus made the use of AGC easy to implement.
While I have seen a couple of early TRF sets with AGC and they are rare, but the Wheeler patent shows triodes in a TRF (neutralized) arrangement using AGC. In his arrangement he used a triode as a diode and fed the filtered diode output voltage back to the first RF tube. The volume was controlled by a rheostat (wire wound) in the filament circuit of the first audio amplifier. I believe that the wire wound control was used in radios at least until the early 1930’s. A carbon rheostat was proposed by several people around 1875 with Edison having invented a carbon control about that time, and it was quite large. I would assume that the original carbon controls were for control of motors. It required some years before the invention was made into a form useful for the radio. Thus the carbon control was invented before radio, and not in a form useful for the radio.  The wire wound control was difficult to make in the high resistances required for the radio. The wire wound control was physically large and the change in resistance was not as smooth as the carbon control.  Had the carbon control been available to Wheeler I am sure that he would have used it in his circuit, thus making it somewhat equivalent to our later circuits. I would suggest that had the control been available, the TRF with AGC would have been on the market for a longer period. While Wheeler had worked on the AGC circuits in the 20’s, the patent did not issued until about 1932 according to some sources. It was issued as a Hazeltine patent. There was litigation over this important circuit and some sources give a later date for perhaps another patent issue, but it is also reported that the litigation went to the Supreme Court where the Haziltine patent was reversed.
There were some early sets with AGC before the Philco 95, but they had complex circuits. These early circuits were difficult to service. These may have been used because of patent problems. I have worked on a few of the sets. The circuit used zero volts for the plate voltage of the tube that was used to drive the AGC grid circuits. The tube had a large negative supply for the cathode – generally around two hundred volts (negative). The RF signal was sent to the grid of the tube and the AGC tube was biased in a non-linear manner such that it was basically a detector. When the tube conducted because of the RF signal the plate would then supply a negative voltage to the grids of the tubes that were to provide the gain control. The tube could also be called a non-linear DC amplifier. When all the voltages were correct the circuit worked great. The circuit was, however, sensitive to component drift and changes in the tube bias and tube ageing was critical. These sets were poorly understood and often became unusable because of modifications made by those not qualified to service the sets. I have serviced these sets and they are difficult to repair because of the many resistors used for setting voltages.
My interest in radios was renewed after I retired. I had all the test equipment that was necessary for radio repair. I restored a few radios and began to notice the change in the technology of the radio with respect to the year the radio was manufactured. I became interested in these changes because some of my work involved circuit design with tubes. Around the early 1960’s transistors, and integrated circuits replaced nearly all tubes. Integrated Circuits, and Transistors were another world, but they had great advantages.
I found a TRF set that was built in the late twenties and was struck by the use of what I labeled a very crude gain control. The control had a low resistance wire wound potentiometer directly in the antenna circuit. The antenna went to one terminal, and the other end of the control was grounded. The center post went to the antenna coil.  My first thought was that this was clearly a terrible arrangement by some primitive manufacturer. I started looking at schematics in the years this radio was made, and found much the same arrangement. This brought the question to mind as to why? How could this happen – it causes a large loss in the antenna output to the radio and destroys the noise figure of the radio.  
I began to take a closer look at the available components and on second thought it began to make sense that the arrangement was reasonable- for the time.  The antennas at the time were generally very good. The resistive antenna control would cause a loss in signal in addition to a poor noise figure, but it was not all that bad – the antennas of the day were still better than the later loop antennas. Furthermore, the noise figure is really not so important when receiving signals at such low frequencies. With the available components it was clearly difficult to put a gain control in the audio because there was no AGC and without AGC the TRF radio frequency circuits could easily be overloaded by a strong signal. It was not desirable to put the gain control in the audio since there was no AGC. However, I have seen one early TRF with a volume control in the audio, and without AGC.
It is not my purpose to discuss battery sets, but to consider the AC set. The battery sets had a variety of gain control circuits. The filament control was useful and important. It was not until after retirement that I began to understand the historical importance of the battery set. Some of these battery sets are quite valuable. Many collectors have recognized their importance and have fine collections of these irreplaceable radios. While the battery set is very interesting historically I have little interest in collecting them. They are by necessity a comparatively crude radio. The AC set has far superior circuits with their advanced  components.  The AC set is so easy to plug into the wall.  To me the radio began with the AC set and while that is incorrect I still seem to have that mindset. Of course were it not for the Battery Radio we would not have AC sets.
One of the oldest sets in my collection is of the late 20’s and it has an antenna volume control. The control is popular and unusual in that it not only controls the signal input to the antenna coil but as the control is turned down to reduce the station volume it also increases  the bias on the first RF amplifier.  This circuit would be desirable in a strong signal area because it would reduce the grid circuit signal pickup from the first RF amplifier.
The superhetrodyne became the radio of choice in the early 1930’s and by this time it was in a form very much the same as the modern miniature tube radio. Most had AGC and the AGC voltage was obtained from a diode. The volume control was placed in the audio amplifier where it belonged.
There is not a lot you can do to a basic radio after using four or five tubes. The basic radio required only one IF transformer along with one RF stage. This radio is good for local stations, and more with a good antenna. To improve the performance of the basic radio you may add one or two more RF stages and one or more IF stages. The sound of the radio is often important and it may become desirable to add a push- pull output stage – not always necessary but it allowed the manufacturer to increase the tube count and this was a great selling point for the marketing of the radio. Adding sound improvements would allow tone control circuits; another good selling point.   Some of the tone control circuits were quite complicated. Then the better sets often used short wave bands which could add a tube if so desired by marketing.  Push buttons were a popular marketing device and the push buttons could be mechanical or electrical, A few expensive sets also would add a motor for station search mode. All these additions were desirable for marketing, and with cabinet design it would form a dramatic radio. Little was said about performance or perhaps we should say specifications. This was because the radio is rather technical and the specifications could bring up some questions that were difficult or impossible to explain to the general public. Therefore little or nothing was said about the actual performance by using numbers for things like gain, noise figure, and bandwidth or frequency response. Of course, nothing was said about the performance of the AGC, and this was unfortunate because something like good AGC was desirable, but the AGC performance of many sets was very poor.  It was some time before sharp cutoff tubes were available, and these tubes were desirable for improved AGC because they improved the AGC by adding AGC gain. If you really wanted to improve the AGC performance there were other means available but it was technical and therefore not used by the manufacturer to sell sets. It is much the same today when you buy a radio, TV or computer. Few people will look into or have the understanding of specifications when buying a complicated electrical device. Most shoppers take the advice of the sales person who is generally technically unqualified, but knows some of the sales brochure information. The sales person often gives incorrect answers to fairly simple questions.  Today it is easy to obtain reviews of technical devices, and the reviews are usually very good.
I find that many new transistor radios have excellent to outstanding AGC performance, but it is not used as a sales feature. If you try to sell an amateur radio operator (ham) a radio (transceiver) you will find a long list of specifications for these sets because the ham is usually technically qualified.
Again, it is interesting to me that I had never seen an analysis of the AGC system and the reason is that it becomes very sticky when you have to use the actual tube characteristics in the analysis, that is to say, the actual tube curves. To describe a tube curve mathematically requires a complicated equation. I like to make mathematical system equations simple if it can be done without loss of the actual reality of the system. I made some simplifying assumptions, and they work well. The simplifications provide a realistic mathematical picture.
The actual simplification is to treat the tube curves as a straight line rather than a curve. The equation of a straight line is simple and straight forward. The system can be a bit more complicated if so desired by making the straight line into as many parts (piecewise solution) as desired, but this would be done only if you were actually designing a radio, and wanted the actual numbers rather than a straight forward explanation of how the system worked with a simple expression. The straight line approximation of such a tube curve is shown in Figure 1.The bias equation shown in Figure 1 is actually relatively easy to obtain by using discrete or integrated circuits. My patent (Number 2,951,980) describes one such means. While the patent shows a variable gain network using diodes – today we should use field effect devices in a manner shown in the patent. The circuit described in the patent was used in the first Double Sideband Synchronous Detector receiver – generally known as a Synchronous Detector and invented by Dr. John Costas...
The bias equation shown in Figure 1 is actually relatively easy to obtain by using integrated circuits or discrete components. My patent describes one such means. While the patent shows a variable gain network using diodes – today we should use field effect devices in a manner shown in the patent. 

The bias equation shown in Figure 1 is actually relatively easy to obtain by using integrated   circuits or discrete components.Tubes require a negative voltage for their bias.The radio detector rectifies the RF signal and the detector is arranged such that it provides a negative DC voltage. The size of the DC voltage is directly a result of the size of the carrier received by the radio. The output from the detector is filtered by using a large resistor (normally about one megohm) with a capacitor bypass of around 0.05 microfarads. This filter removes the audio signal and only allows the carrier DC component to reach the tube grids as AGC bias. The AGC system is actually a servo system and would be classified as a type one servo. This system is stable if only one “low pass” filter is used. Using two such filters in series would cause the system to become unstable . As is seen in Figure 2 the larger the DC bias the lower is the tube gain. When a small signal is received on the antenna of the radio there is only a small negative bias from the detector and the RF gain of the radio is large. When a large signal is tuned in - a greater negative bias is sent to the grid of the tubes resulting in a smaller gain from the RF stages that have the AGC bias.

Figure 3 shows the block diagram of a radio with AGC. The algebra is shown in the diagram and as you can see it is straightforward. It uses the equation of the straight line shown in Figure 2 instead of the actual tube curve equation which would complicate matters. Note that the block diagram of Figure 3 includes a filter block F which is nothing but a large resistor and a bypass capacitor and the equation for this block is not included in the equation except by denoting it as F (t) which infers that it is some function of frequency or time, and we leave it in that form for simplicity.

Note also that the straight line formula uses a (1-Eo) equation for the straight line, again for simplicity. The numeric (1) influences the gain of the system because it determines the steepness of the tube bias line – in our case the line simulates the tube bias curve.The block diagram of the AGC system shown in Figure 2 shows the radio input term as Asin wt (where w = 2Pi f radians) for the received signal from a radio station and would in the case here, A(t) represents the carrier but in general A(t) is actually a function of time that includes the sidebands with the carrier. In our case we will ignore the modulation and define A as the carrier because we are only interested in the affect the carrier has on the automatic gain control system.Figure 2 shows the antenna input as Asin wt and C is the mixer with a gain of C but C does not have to be a mixer it could just as well be a tuned circuit with a gain of C.  Included is another gain circuit which would be considered as the IF gain, but again this could just be more gain from a tuned circuit – in other words the block diagram will cover either a superhetrodyne or a TRF set. The block D is the detector and is considered to have a gain of unity even if it has a gain slightly less than unity. The block F is the filter and will not be included in the output function for E0. F(t) is the filter and is assumed to have unity gain for the detected carrier DC which is sent to the tube grids.We will neglect the block F since it has unity gain for DC. If the bypass in F were to open, then the AGC loop will become unstable because F is actually a stabilizing factor for the AGC servo loop. The system is a type one servo which means that the loop has only one important lag function –the capacitor and resistor which give the loop a six dB per octave drop in loop gain. Type two servos can be difficult to stabilize.We will simplify the analysis by setting G equal to one. The curve of the input signal (esinwt) versus the output E0 is shown in Figure 2. In actual systems the AGC ceases to operate properly for very large signals, and the output of the radio typically increases because of the tube non-linearity. The system operation as described is somewhat optimistic because of the simplifications made. The system would be more realistic if the constant in the denominator were to be increased to a larger value and this would then degrade the system performance. What would appear to be very small insignificant changes such as the availability of the carbon potentiometer are found to make important changes in the final product.  Changes such as the multi-grid tube have made great changes in the design. Not to mention the great improvements in the speaker magnets or the ferrite inductor advances that have made possible the reduction in radio size such that they can now be made in the size of a one-eight inch thick  card which is referred to as a card radio, and they are presently made in an AM/FM radio- both on the same card. I found my first card radio in Japan and it was just AM but later got an AM/FM card radio on eBay, the size of the card is about the size of a credit card. .  It is rather amazing to see that we have gone from the wire wound resistor/antenna radio gain control to a miniature card radios in such an interesting industry. It took about one hundred years, but it is important to remember great strides are often made in very short time periods. These great strides appear to go along with all the technologies that I can think of. I never cease to be amazed in the advances that we have made in our technical areas. The vacuum tube provided a foundation fundamental to many of these advances. It seems that the infrastructure that  is close to being in place for a new industry before it is gathered, and put together by a person or some group. When all the pieces for an idea are available it is not unusual for a number of groups to have the same thought, and many attempts are made to begin a new industry before someone (or company) gets it right. The radio is a good example.


Questions or Comments? Please e-mail me at Thanks!


Popular posts from this blog

The Costas Loop - Dr. John P. Costas

Signal Generators.

Consciousness - 1