1

I've got this audio amplifier board, which contains exactly the following circuit:

enter image description here

I'm trying to understand this circuit, and the "AGC Control" part (highlighted in blue) has me the most curious.

What I had assumed is that AGC means "Automatic Gain Control", and that this AGC circuit will adjust the impedance of the input with relation to the output. Could this be correct? How? Is it always active? Is it only active when specific levels are reached? What if I remove this AGC circuit?

I've been analyzing it, but still don't understand it. I found it to be really strange that it takes from the +7.5V rail into the emitter of a transistor. The diodes present in the AGC circuit seem to have a reverse voltage of 35V, according to their datasheet, and 40V(peak) is around the maximum output voltage of this amplifier before it shuts itself down, according to my tests.

I tried to see what the effects of this AGC look like; however, with my tests I couldn't notice any attenuation anywhere.

Also, could "AC-IN" (highlighted in green) really be just to turn on and off some muting feature?

The diagram I posted above is from the service manual for the SA-AK37, made by Panasonic. Seems to be the only page of the manual available online.

Christopher H.
  • 31
  • 4

1 Answers1

1

This is an interesting circuit - thanks for posting: The speaker output signal is one way rectified and is charging the C551 with approximately the peak value of the L or R channels (scaled with the resistor dividers R574 and R571 and analogously for the Rch).

The DC voltage of C551 is divided by the R570 in series with R568 and the combination of R569 and D544. The diode drop and R569 add non-linearity to the transfer function: The slope gets smaller once the voltage at this node reaches the threshold of D544 (roughly 0.55V). The result is that the voltage at the base of Q522 is a function of the peak value at the speaker outputs. This was the easy part.

You correctly noted that the Q522 has emitter positively biased (3.3V) while the collector is at 0V potential. The transistor is in "reverse active mode" - basically with E and C terminals swapped. Since this is NPN transistor you get the same "palindrome" when you swap the terminals. The internal design of the NPN transistor is far from being symmetrical and the transfer function will be significantly different. That is what the designers wanted. Now - if you re-draw the schematic with C and E swapped it starts making sense. It becomes an emitter follower with it's "collector" biased at 3.3V. The "emitter" is biasing the base terminals of Q520 and Q521 and these transistors are with their collector impedance loading the input signals effectively attenuating the L_CH and R_CH signals assuming an output impedance of the circuit driving this board. This makes the gain of the amplifier decrease when the output voltage reaches certain level - protecting the speakers and your ear drums.

Martin Vana
  • 684
  • 1
  • 11
  • Thank you. From your explanation, I instead assume that the AGC is always active, and its attenuation effect increases exponentially as voltage levels increase, and not when reaching any specific level. I also assume that the attenuation is actually very tiny. Am I incorrect? – Christopher H. Sep 16 '21 at 18:17
  • I would not say the function is necessarily exponential or logarithmic nor whether the attenuation is small. (Small would probably have no utility). Since the transistor is in reverse mode there is probably not enough information to answer the question analytically. I suspect the purpose is to limit the output power at certain maximum. Can be probably measured and characterized. – Martin Vana Sep 16 '21 at 20:23
  • 1
    Q522, 2SD2144, datasheet advertises unusually high Vebo, suggesting it's intended for this unusual use. Some audio muting circuits did this, using an NPN as a shunt-to-ground for a bipolar signal line. Interestingly, the transistors Q520, Q521 doing the shunt function in this circuit seem ordinary, and it is the control transistor (exposed to the output) that gets the beefed up reverse Vbe capability. – Pete W Sep 16 '21 at 21:45