AD603 Data Sheet Circuit - Figure 49, page 18


I am making an amateur radio transceiver.  I have based the IF strip of the receiver section of the transceiver on the circuit shown in the data sheet page 18 (figure 49).  My circuit is as shown in this figure except that instead of connecting pins 5 and 7 with a 2.5 k resistor, I used a wire link so as to decrease the gain of each AD603.  The amplifier section of this IF strip works very well and I am presently using it in a working prototype receiver.

I have a question regarding the AGC circuit shown in the data sheet, however.  I have been unable to get the AGC circuit shown in figure 49 to work.  I think what is happening is that Q2, a 2N3906, is biased so that it is saturated whereas transistor Q1, a 2N3904, conducts only on negative peaks of the amplifier output (The emitter voltage of Q1 averages at about 5 V, very nearly the same as the base voltage).  Consequently a constant voltage of about 9.5 volts is maintained on the capacitor marked Cav.  Since this voltage is the AGC voltage it is much higher than it should be for proper operation of the AD603 ICs in sequential gain mode.  Based on the data sheet information, its maximum value in this circuit should be 7 volts, I think.

I have made a different AGC circuit that works with my IF strip but I would still like to be able to try different AGC circuits for evaluation and, if possible, I would like to get a circuit like that shown in the data sheet (with suitable modifications) working.  I would certainly appreciate any suggestions as to how to modify the figure 49 AGC circuitry.

Thank you,

Allen Wootton 

  • Hi Allen,

    Congratulations on taking on quite a challenge, one could write a book on building a high gain IF strip and have plenty of material left over to write sequels. I get a sense of what you're doing but I take it you may have a lot of access to test equipment for troubleshooting. For starters, you might tell us how you are going about the project. If there are folks this inquiry you might benefit from additional and perhaps even better advice than mine. Have you built this on a printed circuit board? Do you have a schematic? Have you taken voltage measurements at all the nodes?

    Most issues like you describe are most often due to construction errors. Although I haven't built the circuit (this was done long before I arrived at ADI) I've analyzed it in the past and it checks out as a valid circuit. If you're seeing saturated transistors you should look for wiring errors anywhere in the signal chain. The error could be due to a lot of things such as wiring errors, loss of or error in power supply voltages, open grounds, oscillation, exterminated input, etc.

    I'm out of the office at the moment but will be back for a short while in December. Are you a member of ARRL? I used to build all kinds of equipment from their publications and found their QST magazine, annual handbook, etc. very informative, especially for a 15 year old kid (then, not now ).

    I Hope this works,


  • Hi James,


    Thank you for your reply to my question regarding my AD603 based IF strip.  I have a reasonably good set of test equipment - oscilloscope, signal generator and spectrum analyser.  All of it is pretty old, but it works ok for me.


    I built the circuit exactly as shown in the data sheet except for using a wire link between pins 5 and 7 of both AD603s in place of the 2.5 k resistors shown in the data sheet circuit diagram.  This is to decrease the gain of each of the AD603s.  As it stands, my IF strip seems to have all the gain it needs.  I presently have it connected to a mixer, post-mixer amplifier, crystal filter, post filter amplifier, then the IF strip, product detector and audio amplifier.  All these parts are constructed in “dead bug” style on pieces of unetched printed circuit board where the pcb serves as a ground plane.


    In the IF strip, the surface mount AD603 ICs are soldered to “surf board” circuit boards that allowed me to work with them more easily than would be possible if I used straight “dead bug” construction with them.  I used surface mount capacitors for bypassing where it made sense and leaded components otherwise.  The IF strip itself seems to work very well and, when I have an antenna connected to the mixer I am able to receive CW and SSB signals in the amateur bands without any problem. 


    I tried connecting a potentiometer between the +10 V supply to the IF strip and ground with the slider connected to the AGC line connected to pin 1 of both AD603s.  This works very well as a volume control - varying the pin 1 voltage between about 5.5 and 7 volts produces the expected change in gain as outlined in the explanation provided in the data sheet where a +/- .5 V variation between pins 1 and 2 produces a 40 dB change in gain in each AD603.  I have the two AD603 pin 2s connected as shown in the figure 49 with the first AD603 pin 2 at 5.4 V and the second one at 6.4 V so that the combination gives a sequential gain as the pin 1 voltage is changed.  Maximum gain occurs when the slider voltage is about 7 V.


    The problem I’m having is not with the IF strip itself, but rather, with the AGC circuit shown in the data sheet.  I think that the average output voltage of the IF strip (at pin 7 of the second AD603, A2 in the diagram) should be about 5 volts and this is about what I measure with my digital voltmeter.  When the oscilloscope is connected it shows the expected AC waveform varying around 5 V (I’m using an 8.19 MHz IF based on my homemade crystal filter that uses inexpensive crystals).   The data sheet diagram shows the base of Q1 should be at 5 V due to the biasing produced by the series resistors R10, R11 and R12.  I measured 5 V on the base, as expected.  I measured 9.52 V at the emitter of the 2N3906, Q, and 8.87 V at its base.  The 0.65 V difference between the emitter and base seem to me to indicate that the 2N3906 is conducting and providing a low resistance pathway to the collector of this transistor.  I measure 9.51 V at the collector, about what I would expect if the 2N3906 is fully conducting.  I think that the 2N3906 is conducting whereas the 2N3904 is not - at least, not enough to diminish charge on the capacitor Cav shown in the diagram as the capacitor that sets the AGC time constant.  Thus 9.5 V is applied to the AGC line and the IF strip is always at maximum gain; there is no AGC action.


    So far I’ve built the AGC circuit twice to try to make sure that I haven’t made an error.  I don’t think I have a error in the construction (but if there is a error it would not be the first time I’ve made one, unfortunately!).  As it stands, however, my measurements and my understanding of the circuit both seem to me to match one another. This is the main reason why I’m asking about this question - I think that perhaps there is an error in just the AGC part of the circuit (or in my understanding of it) and I’m really trying to gain an understanding of either how to modify the AGC circuit or my understanding of it.


    I am a member of the ARRL and certainly find the “Handbook” a wonderful reference, as is QST.  I’m quite excited about trying to build a transceiver - it is something I’ve wanted to do for a long time and now have time to do in retirement. 


    Thank you again for your response to my question.