# high voltage to audio

1. ## high voltage to audio

I have a signal which can be from 40-90v and the power audio amp takes up to 5v signals.

It seems logarythmically scaling the high voltage signal is the ideal way to get it to the amp.

There are very expensive high voltage and high power oamps out there but they are huge and very expensive, whats a minimalistic method to get the signal logarythmically scaled?

Perhaps there is a transistor logarythmic scaling circuit but Im not sure which is ideal.

•

2. ## Re: high voltage to audio

Hi,

For audio one uses linear voltage dividers (resistive voltage divider). If you use something non-linear like logarithmic you get lots of distortion.

Maybe you mix the phrase with "logarithmic audio volume control potentiometers".
With these pots
* not the audio_input to audio_output is logarithmic but
* the moving_angle to gain is logarithmic.

Klaus

1 members found this post helpful.

•

3. ## Re: high voltage to audio

Originally Posted by KlausST
Hi,

For audio one uses linear voltage dividers (resistive voltage divider). If you use something non-linear like logarithmic you get lots of distortion.

Maybe you mix the phrase with "logarithmic audio volume control potentiometers".
With these pots
* not the audio_input to audio_output is logarithmic but
* the moving_angle to gain is logarithmic.

Klaus
Wouldnt linearly scaling the signal ommit much of the fine details were as log scaling would preserve them?

This might be intended for other signal types aside from audio.

•

4. ## Re: high voltage to audio

Hi,

You could simply use Excel to simulate this.

Let's imagine a sine waveform input.
Now if you use V_out = A × log(V_in)
You will see that there is a problem with input voltages <1

You may choose another mathematical nonlinear formula.

Maybe square_root.
It will work with positive input voltages only, thus you need to solve this problem, maybe this way:
V_out = sqrt(|V_in|) × V_in / |V_in|

If you simulate this you see that the sine shape becomes more like a square wave. This means overtones = distortions, THD.

What you need is some kind of "audio compressor".

Klaus

1 members found this post helpful.

5. ## Re: high voltage to audio

Now if you use V_out = A × log(V_in)
This is a common problem; we use the conventional db (decibel for these messy problems) in such messy cases.

If the input voltages is less than 1 (V, I presume), you express the voltage in uV - but then other problems (like negative voltages) will certainly give headaches.

The problem is that log function argument must be a pure number and hence V or A are not proper arguments. We often use a ratio for this: log (test_V/ref_V)

The problem of negative voltages still do not go away and we use power (depends on the voltage squared) but it cannot be negative - helps certainly.

But in the log scale 0 to 1 is given the same space as 1 to infinity. So it is inherently more useful in the 1 to infinity range.

1 members found this post helpful.

•

6. ## Re: high voltage to audio

You should really tell more about the signal you want to process. How is the problem related to audio? I doubt that "logarithmic scaling" can be usefully applied to audio signals. The nonlinear characteristic will severely distort it. A pure log characteristic doesn't work at all, because it's undefined for negative half waves. You need at least a symmetrical completion of log function like y = sign(x)*log(abs(x)). If the nature of audio signal should be maintained, an envelope manipulation like a compressor function may be more appropriate.

1 members found this post helpful.

--[[ ]]--