EMRFD Message Archive 12828

Message Date From Subject
12828 2016-05-16 07:11:10 KENNETH CHASE AGC help
Hi All

I am looking for some help.

I would like to try the AGC circuit in EMRFD page 191 Fig 6.55.

Before I build it, I would like to understand how it works. I don't have any opamp experience except soldering them.

What I'm actaually looking for is a circuit to give an outout of about 4 volts (max gain) DC with no RF signal. Then drop as the RF signal increases.

This DC level will control the gain of a RF preamp.

An info would be much appreciated.


12831 2016-05-16 08:27:54 Bill Carver Re: AGC help
In Fig 6.56 the left op amp is just an audio amplifier with an AF gain of 2. The middle op amp is an inverting audio amplifier with a gain of 1. It turns the audio "upside down". So the first op amp changes C1 on positive audio peaks, and the second op amp flips negative peaks over to an equal positive peak which also charges C1. The DC voltage on C1 is buffered by the third op amp so C1 only "sees" the 10M resistor, discharging the rectified peaks with a 10M x 1 uF = 10 second (!!!!) time constant.

All the op amps are biased to +6V by the two 10K resistors off the +12V supply, so the DC output from the third amplifier sits a diode drop lower with no signal. But the diode current is very low so I'd expect perhaps 5.7VDC with no signal. The output will increase with signals, which is the opposite direction from what you say you need.

Op amps come 1, 2 or 4 to the package, and LM324 has four. So what I'd suggest is you use the fourth op amp to invert the DC output and shift it to +4V with no signal. Put 10K from the third op amp output, the "to controlled stages" output, to the "-" input of the fourth op amp. And a 10K from the output of that fourth amplifier back to the same "-" input. That inverts the DC so it will go down with increasing signal. Then wire the "+" input of that fourth amplifier to the wiper of a pot. The pot will hang off the +6VDC produced by the two 10K resistors and bypassed by the 22 uF...those are already there. The other end of the pot should be about +4V. So if you used a 100K pot, you would put a fixed resistor of around 180K from the other end of the pot to ground. The pot will set the no-signal output of the fourth op amp. I'd expect a pot wiper of about 4.85 volts will give you 4.0 volts DC with no signal, and you can vary that quiescent voltage around that 4V you want.


12832 2016-05-16 16:59:43 Ken Chase Re: AGC help
Hi Bill

WOW thanks for the explanation. While I was supposed to be working, I did read up
12833 2016-05-16 22:26:41 Bill Carver Re: AGC help
Yes, the voltage divider and bypass capacitor will make about +6V, and that biases everything to the middle of the supply voltage. The 4.7K to the "+" pin is about equal to the two 10Ks that connect to the "-" pin, so in addition to biasing the amplifier to half the supply voltage, it also compensates for the bias current of the "-" input of the amplifier. This trick is useful/necessary when the impedance at those terminals is high....but in this circuit I don't think you would see a change of even one millivolt if the 4.7K was replaced by a piece of wire.

The suggested use for the 4th amp will produce an output voltage that is solid as a rock as long as you don't draw more the 5 milliamps or so.

Op amps are analog magic. Once you figure out the basic concept of what an op amp is and how to use it, you can come up with your own circuits with a pencil and a napkin. And the vast majority are dirt cheap so you can get wild and crazy designing things!

In studying examples to understand what's going on, it might be useful to stick to application notes and sample circuits that use + and - dual supplies. It's easier to understand/analyze what's happening that circuits biased to half of a single supply voltage. For example, if you had a -12 volt supply for the LM324 of that Fig 6.56 circuit, the two 10K and 22uF disappear, and the absence of signals the output is zero volts, rising to a positive voltage when signals are applied. For me, at least, it was easier to get a solid understanding of how op amp circuits work. Once you understand the basic concept of an op amp then you can cope with the bias to half supply easier. My opinion anyway.


12835 2016-05-17 06:00:30 KENNETH CHASE Re: AGC help
Hi Bill

I did some more reading and I fully understand what is going on.

I built up the cct last night with LM358's dead bug style. I used a 1 khz signal and it works. Even though the 1 khz isn't strong enough to drop the o/p down to anywhere near zero volts, it is strong enough to change it down to 3 volts. So it does work. One change I did was the 100k pot and 180k resistor. I didn't layout the cct conveniently to use the 6 volt bias. So I used a 10 k pot between ground and 12 volts. The wiper goes to the + input of the 4th opamp. I adjusted it for 4 volts with no signal. So the cct is working. Some more experimenting is in order.

Thank you so much for your help, it is much appreciated.

Another satisfied customer!



12836 2016-05-17 08:06:57 AncelB Re: AGC help
Regarding the LM358 and LM324 series opamps.
There's a gotcha in the middle of the supply rail as the output of the
opamps use class B bipolar transistors.

It's called crossover distortion and it's a nasty surprise regarding
signal fidelity.

Either bias the opamp so that it's output doesn't crossover the middle
of the supply rail., or place a load resistor on the output (4.7k?) to
the low or high rail to force the opamp into class A operation. This
eats a bit more power as the cost.

Also, if you can select the higher grade 'A' versions you get better
voltage offset specs etc. when doing differential amps or canceling DC

With these tips you can keep the output precision to within 20mV using
reasonable gain in practical circuits without trimmer resistors etc.

12837 2016-05-17 08:47:42 Bill Carver Re: AGC help
And NOW, the op amp magic begins............it sounds like you need more
gain so the 4V drops lower with less signal input. In that fourth
amplifier increase the 10K resisstor from output to the "-" input. Make
it 20K and the output will drop twice as fast: this resistor becomes an
"AGC GAIN" resistor. This will change the "4V SET" ("AGC ZERO" ????) pot
setting to get 4V quiescent output slightly. Or make it a 100K pot and
you can vary the gain over a 10:1 range (too much gain and maybe your
AGC system will break into oscillation)("motorboating").

But there's something lurking here that might trip you up. If the AGC
starts with an output of +4V, and goes down.....how FAR does it have to
go down to get full range out of the stage(s) you are controlling? The
output voltage of many op amps will only swing within a few volts of the
positive supply (12V in this case) and within a few volts of the
negative supply (which in this circuit is ground). So if you needed the
AGC to drop all the way to zero, or something like 0.7 volts, the output
of some op amps won't go that far. In which case you might want to
search for an op amp that's said to have "rail to rail" output capability.

I looked at the LM324 data sheet. It says with a 2K load the output
stage will swing to the positive supply voltage minus 1.5 volts (+10.5V)
and within millivolts of the negative rail. Which is better than I
expected. But if you need to approach zero volts output I'd be checking
the "Vo" range for the op amp you're using, and just for good measure
check that you get the desired output swing in your application.

12839 2016-05-17 17:38:18 Ken Chase Re: AGC help
Hi Ancel

Thanks for the info. I will be looking at this gotcha, when I get to that stage. At this point, I don't see the output going over 4 volts with the middle of the rail at 6 volts.



12840 2016-05-17 17:45:31 Ken Chase Re: AGC help
Hi Bill

I'll have to do some experimenting as to how low the voltage has to go on an overloaded signal. The amp being driven with the AGC, has a lot of gain, so I am guessing it won't have to go down to zero. I'm hoping somewhere around 1 volt or maybe more. But some testing I need to do.

Once again, I thank you for your input. I will be back later with results.



12841 2016-05-17 20:04:28 Bill Carver Re: AGC help
Good luck and have fun!