I note that many popular kit and QRP radios do not include any AGC capability. Thus it must be possible to operate successfully without AGC.

  • Are there difficulties involved in manual gain control? And if so, what are they?

  • What are the primary benefits of AGC?

  • Is AGC most beneficial if implemented in the RF stage (LNA?), IF stage (if any), or audio output stage, or for multiple blocks/stages?

  • If AGC is implemented across multiple receiver blocks (say both RF and audio), how is loop stability accomplished?

  • 1
    $\begingroup$ Applying AGC in the IF of course isn't the only way to do it, but it was fairly easy. There's only one frequency you need to concentrate on (often 455kHz) regardless of what you're receiving, the signal is still RF, so you can rectify it and get a DC voltage that corresponds to the strength of the signal, and then use that DC voltage to control the gain of a prior (or later) stage in the receiver. IIRC, many rigs use that voltage to drive the S-meter too. Of course Direct Converstion receivers don't have an IF, so that doesn't apply. $\endgroup$ – Duston Oct 19 '20 at 13:42

The dynamic range of (shortwave) received signal strength varies from noise equivalent field strength (in the order of 250 nV/m in 3 kHz SSB bandwidth) to 1 mV/m (for radio ham transmissions via the ionosphere). Local signal can go over 1 mV/m. In short: dynamic range is in the order of 72 dB. AGC doesn't have to cover that range, but 50 dB is "nice to have". The slope of the AGC has an effect on how the difference between weaker and stronger signals are perceived. With rigid AGC, zero dB output variation for 50 dB input variation, you loose the "band feeling". Good choice is 10 dB per 50 dB.

Your questions:

  1. Manual control requires continuous action for this control.

  2. Benefit of AGC is an automatic gain control; that is in the name.

  3. AGC at the antenna looks like a good solution to prevent audio level variations, but results in a very limited maximum signal to noise ratio. Beginning at the start of the AGC, the S/N will not grow further. On the other extreme: AGC after the receiver demodulator leads to excessive requirements for large signal handling up to the demodulator. Gain variation for AGC must start at the last IF stage, just to prevent overload, and then anywere in front of this, in the mixer section, and then, for the strongest signals, in the front end LNA, or even at the input of the LNA: the antenna input.

  4. Stability of such a multi-AGC loop is best obtained when the gain variation is a FACTOR per volt AGC control voltage, I.E. a constant decibel per volt gain control. Only then the AGC feedback loop is constant, with a fixed time constant. Any other attempt results in instability or variation of the time constant (and thus the frequency response: a disaster for AM envelope distortion!). Best AGC is a single time constant or a PID controller.

Experience obtained in 40 years radio development at Philips/Magnavox and Catena (NXP). Brightnoise, PA0FSB

  • $\begingroup$ So you are saying that if the radio designer picks 10 dB per 50 dB, then a strong signal sounds 10 dB louder than a signal that is 50 dB weaker? This sounds very reasonable to me. I don't like radios in which a loud signal and a weak signal are exactly the same volume in the headphones. $\endgroup$ – rclocher3 Oct 19 '20 at 16:44
  • $\begingroup$ That's it, yes. The "feeling with the band". $\endgroup$ – F. Sessink Oct 20 '20 at 11:20

There's a huge variance in signal levels — the lowest signal that's readable above the noise floor and the strongest signal that won't overload the receiver are probably at least 60dB apart. To get that into a comfortable listening range without AGC, the operator has to adjust the receive level. Maybe that doesn't sound like a big deal (can't you just set it for your QSO partner and leave it?) but:

  1. If you're scanning the band, there will be lots of different signals, and you would like to be able to hear the quieter ones without blowing your ears away when you tune to one that's 30 or 40 dB higher.

  2. Maybe you're listening to a net, and different participants have vastly different signal strengths. Do you want to have to spin the volume knob every time someone new starts talking (and do it quickly, to avoid missing the beginning of a sentence, or getting ringing ears)?

  3. Even in a single QSO, signal levels can change rapidly due to fading.

While AGC isn't perfect, it greatly reduces all of these effects, making life much more comfortable for the operator. That could be accomplished by AGC near the end of the chain (AF or last IF), but doing it earlier in the chain might be necessary to keep later stages in their best (most linear regions) and reduce overall distortion and noise.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.