Signal Presence Detectors?

[Harry Bissell]

I don't think the optocoupler will help... if you can light the LED in the
optocoupler
without interference, you can light a visible LED as easily....

It the audio signal is buffered with a high impedance, there should be no
direct
coupling... I agree dI/dT is the worst problem... Good power supply
decoupling
should solve that... even an RC filter to the display stages should drop any
disturbance
to sub audio...  A good method might be to drive a retriggerable one shot to
drive the LED... (which might make problems of its own...)

Have you seen bipolar (red-red) LEDs... I LOVE them... i even drive them off
the
AC mains (with a high value series resistor)... The forward biased LED
protects its
brother from reverse bias, they are "full wave" and light on either polarity.
Check them out.  I'm also using them as clipper stages for fuzz circuits...

H^) harry

Martin Czech wrote:

> :::I'm a stickler for audio quality though, and I'd be interested if
> :::there are known good ways to do this without creating any noises,
> :::distortions or spikes on the power supply lines.  So running the input
> :::signal through a gain stage, rectifying it, and directly drving an LED
> :::would not be a good approach.
>
> why? As long as you don't allow the gain stage to saturate (clamping)
> the supply current should only change very little with audio
> signal. So the gain stage allone should be no problem.
> Or any other op amp in your circuit will give you the same problem
> (ok, say gain stage, for all op amp haters ;-=>)
>
> However, as soon as you run a comparator this will be the case.
> The largest current thump will be of course by the LED.
> Do you remember the days of ECL logic? It is switching, but supply
> current is almost constant. Because it's differential.
> This would mean: make the LED driver symetrically, a dummy
> LED and the real one, one of them is always active.
>
> OTOH you could choose a soft turn on/ turn off circuit,
> it is not the current that bothers you, it is di/dt.
>
> But certainly separate power and ground for the detector will
> be the best. If you really have problems you could use an opto
> coupler, linearity is no requirement here.
>
> There are basically three ways of interference:
>
> -galvanic: separate all lines, ie. power and ground
> -electric: relatively easy to shield with metal, or use slow slopes
> -magnetic: difficult to shield low frequency, so distance or slow slopes
>
> Sorry for beeing academic.
>
> m.c.