agc preamp distortion

I hv the following circuit assembled. I hv used 2N 5019 fet as per the instruction of the designer. I tried with replacement of LM 358 with njr4558 ic to reduce distortion.
I hv tested the circuit by connecting satellite receiver audio out as input & tv audio in connected to the output of the circuit mentioned.
The sound is distorted after connecting the unit.
waiting for suggestions of the fellow members.

Is your 2n5019 has linear resistance to Vds characteristics.
It seem that 2n3904 create some distortion.
You need one more amplifier (same input at pin 3) to isolate the output from 2N3904.
4580 is better option for low noise low distortion wideband but higher power needed than LM 358.

The FET is a very non-linear voltage controlled resistor. Its linearity can be made much more symmetrical if half the drain signal is fed to the gate as in my modification to your circuit.

The FET will cause additional distortion if the signal at its drain exceeds about 100mv peak. Add a fixed attenuator to the input of the circuit so that the signal is not as high.

Thanks audioguru,
I will try this circuit on Monday & will give u feedback.

This is the description of the project from the author:
Using the circuit presented here, you can construct a very inexpensive AGC amplifier with the following features: a dynamic range greater than 50 dB; negligible distortion to the output waveform; fast attack and slow decay; an adjustable output level from 0 to 1.2 V p-p; operation from a single 5-V supply; less than 1-mA current drain; and low cost (uses one half of a dual 8-pin op-amp package at less than $2.50 in parts). Better yet, if you need a second channel, the remaining half of the op amp can be used for that circuit.
Referring to the diagram, Q2 (a Pchannel JFET), coupled with R2 and the equivalent resistance of R3 and R4, form a voltage divider to the input signal source. With input levels below 40 mV p-p, the input is evenly divided between R2 (120k) and R3 ¦ ¦ R4 (120k). The output amplitude of U1A isn’t large enough to turn on Q2, which acts as a positive peak detector. The gate of the JFET is pulled to +5 V, pinching its channel off and creating a very high resistance from drain to source. This essentially removes it from the circuit.
At input levels above 40 mV p-p, Q1 is turned on at the positive peaks of the output of U1A, lowering the JFETs gate to source voltage. The channel resistance decreases and attenuates the input signal to maintain the output of U1A at approximately 1.2 V p-p.
The circuit, as shown, was tested with a sine-wave input ranging from 300 Hz to 30 kHz at 40 mV to 20 V p-p, a 54-dB range. It maintained the output level at 1.2 V p-p, ±0.5 dB, with no visible distortion when comparing it with the input waveform. With a 40 mV to 20 V p-p input signal, the amplitude of the signal across the JFET (VDS) measured less than 20 mV p-p.
Other JFETs with VGS(OFF) of 5V or under, such as the 2N5019 or 2N5116, should work equally well in this circuit, although they haven’t been tried. To use JFETs with higher VGS(OFF), such as the 2N3993 (it was tried and worked equally well), increase the supply voltage to 12 V.

Another serious failing of the circuit is that it only reacts to positive peaks, a precision full wave rectifier would improve it as well.

Thanks Nigel Goodwin,
Let me know where to put precision full wave rectifier in the circuit?
Is it to put after volume control & before 2N 3904?

In the base lead of the 2N3904, but you would need to add a good few more components as well - easier to find a circuit that does it already? - this one is really pretty poor, and like Audioguru already mentioned it's VERY bad to leave out the extra components round the FET.

thanks Nigel Goodwin,
i will appreciate if i can some good circuit.

Here is a good Peak Limiter circuit that cuts the level down to its setting.
It has a full-wave rectifier and its attack time is very fast.
**broken link removed**

Here's a mod of the original circuit which should respond to positive and negative peaks. I have not tested it, so it's worth what you pay for it.
Audioguru's circuit is probably better (I'm no audio guru).

Thanks audioguru & roff.
I really appreciate the way u r guiding me.
Just one question.
Is the circuit will just work as peak limiter for high audio output or will also boost the weak channel audio signal.
From the satellite receiver few channel audio output is very weak. So it needs to be boosted & high channel output to be reduced.

I doubt if anybody can hear the difference between a half-wave compressor and a full-wave one.

The modification I made to the first posted circuit reduces the distortion of the FET but causes the attack time of the FET to be much too long.
The "better" circuit I posted is not mine, it is from Rod Elliot. It has parts added to speed up the attack time of the FET.

palesha said:

Thanks audioguru & roff.
I really appreciate the way u r guiding me.
Just one question.
Is the circuit will just work as peak limiter for high audio output or will also boost the weak channel audio signal.
From the satellite receiver few channel audio output is very weak. So it needs to be boosted & high channel output to be reduced.

Click to expand...

The circuit isn't really a 'booster' it's an attenuator of loud signals, but this is really just a difference in description.

However, it's not really intended for satellite channels, you perhaps need to find why some are quiet?.

It appears to me to have a maximum gain of 71.2 (37dB) for outputs less than 1.2V p-p.

If the gain is high or if the input signal level is high on the limiter then it limits the output to its max output level, the same as a compressor. Since the gain is high then low level signals are amplified, the same as a compressor.

Thanks 2 audioguru, Nigel Goodwin & Roff 4 all the guidance. first i will try with my circuit as the pcb is ready. then i will make the rod elliot pcb.

I realise this post is quite old, but worth a shot.

The AGC circuit that was orignally posted at the beginning of the forum. Is it possible to implement a version of this circuit that implements an N type FET attenuator, instead of a P type to reduce costs. If so could anyone give an example of the circuit?

Many Thanks,

Jonathan

tufty88 said:

The AGC circuit that was orignally posted at the beginning of the forum. Is it possible to implement a version of this circuit that implements an N type FET attenuator, instead of a P type to reduce costs. If so could anyone give an example of the circuit?

Click to expand...

The original circuit uses an old low power LM358 opamp that is horrible for audio because it produces crossover distortion and has trouble with sounds higher than only 2kHz (audio goes to 20kHz). The circuit also has a very simple Jfet circuit that produces distortion.

In post #9 I provided a link to a much better circuit that uses a low noise low distortion wide bandwidth audio dual opamp, a full wave rectifier and a fast low distortion N-channel Jfet circuit.

Thank you for your prompt response Audioguru. In my application, I am passing the output signal of this amplifier through a low pass filter, and through more circuitry to isolate bass beats, and in turn drive an LED. So I don't think distortion is particularly a big worry for my application?

I do however, for the circuitry available in my design have only one op amp available for this part of the circuit- which is why I was interested in modifying the first post here- as uses one amp. Creating the one that you recommended, would drive up the cost through addition of one more amplifier- although I fully trust it is better performing- I'm trying to keep costs down!

Cheers,
Jonathan

Hello,


One of the other key points in audioguru's most recent circuit is that the FET is biased for DC, yet provides a changing AC impedance. This is necessary for the correct operation of the FET as an AC attenuator. So what this means is that there should always be a proper DC bias for the FET regardless what the level of the AC is, and that means the bias part of the circuit has to be right too. Note the first circuit in this thread does not have that kind of bias, just a couple resistors going to the AC input. This might work if the AC input is already biased for say 1/2 Vcc, but if it is a true AC input then it wont work without considerable distortion.
So check out audioguru's most recent circuit and note that the FET has DC bias but is connected to the AC through a CAPACITOR. This kind of connection regarding the FET is a necessity.