Atenuate analogue signal using analogue logic

[casemod]

Hi all,

I have an analogue signal, sinusoidal shape, whose amplitude is continuously changing. The signal is always greater than the input. What I need to do is attenuate this signal, without causing distortion so that the output has a constant amplitude.

I first thought about using a digital pot, however this requires the intervention of a microcontroller. Is there a way to achieve the same purpose using only components such as an opamp or passives?

 

[Warpspeed]

Simplest way is to use a fet opto isolator such as an H11F1.

The fet goes from a few hunderd ohms to hundreds of meg ohms, its fully isolated and perfectly linear for audio signals up to +/- 30v.
It can either be placed in series or shunted across the audio signal path.

Varying the current through the LED will vary your audio amplitude over an extremely wide range.

Its then just a case of using the peak audio output level to vary the current through the LED in the opto isolator. That will require a peak rectifier charging a capacitor, which in turn modulates the current through the LED in the H11F1.

 

[Dan Mills]

Depends on the details but limiting amplifiers have a long history.

If it is in a more or less audio signal then a search for 'jfet audio limiter diagram' may yield results, or various VCA based approaches.

Regards, Dan.

 

[Audioguru]

Simplest way is to use a fet opto isolator such as an H11F1.

The fet goes from a few hunderd ohms to hundreds of meg ohms, its fully isolated and perfectly linear for audio signals up to +/- 30v.
It can either be placed in series or shunted across the audio signal path.

Varying the current through the LED will vary your audio amplitude over an extremely wide range.

No. It is not perfect.
The datasheet says 2% non-linearity. The graphs in the datasheet do not show a straight line.
I tried it in a Wien bridge oscillator and the distortion was bad unless the signal level across the Fet was almost at very low noise levels. I think with a second Fet for pre-distortion then most of the distortion can be cancelled.

 

[D.A.(Tony)Stewart]

1. What causes the signal amplitude to change? and how fast and how much does that vary?
2. Desired Gain & dynamic range of input?
3. Frequency?
4. Impedance? or current?
5. Disturbances ( Temp, Supply etc)
6. Any other important details?

 

[Warpspeed]

No. It is not perfect.

Nothing is absolutely perfect.

But I suspect for the intended application here, it will be more than sufficient.

 

[crutschow]

Why are you asking the same basic question in two different threads?

 

[FvM]

crutschow is right. Thread closed, please go to the same topic thread for possible open questions.
https://www.edaboard.com/threads/341068/