Analog + digital ground separation?

[stormbringer]

Hey. Getting further along in my preamp design. Currently looking into the channel switching circuit. I will have a couple of logic ic's here and would like some tips.

I know that digital circuits might create some ground noise, anything i should think about when designing to minimize this? Or should i even use a small separate transformer even and optocouple the signals between circuits?

[J M Fahey]

Not that bad, but "delay" switching a little, a common value is adding, say, 1M series + .05 /  0.1 uF to ground, at the switching FET gate or equivalent CMOS switcher.

Cleans/avoids ground spikes enormously and allows you to use "one ground for all".

Not possible in a real Digital circuit where you have tons of many MHz strong signals going on continuously.

Personal view: a switcher, even if conceptually a digital device, is an *audio* switcher for me and audio rules apply.

[Roly]

Yes, in this context signal path switching isn't really what I'd call a digital circuit.  If you are going to control these paths with an Arduino o.n.o. going mungemungemunge all the time then its a good idea to have the digital part confined in it's own pen (or in another building in another State if possible), but single transition stuff is really just "saturated analogue".

As a sidebar, I've seen it noted that the quad LM324 op-amp is a good choice for such control circuitry precisely because it has such a "leisurely" slew rate, and "slugging" the switching signals as JM suggests actually produces a better result subjectively than a sharp "digital" edge.  You don't need isolated power and ground returns to keep spikes out of the analogue if you don't generate the spikes in the first place.

If you are going to use actual logic chips (e.g. 4000-series) then go for low power CMOS and just make sure that they are very well bypassed by an electro and an HF cap right on the chip and this will contain the current surges to that local loop.

[stormbringer]

Alright. thanks again guys.  :dbtu:

[J M Fahey]

Here I'm posting a Commercial Amp solution to the problem.
This Hughes € Kettner (brilliant guys  :dbtu:) uses the very popular CD4053 switcher.

Why so popular?

1) It has 3 (three, count'em ) "SPDT relays" inside .

2) it has internal logic to drive them, so you simply need to either ground or apply +5 to +7V to each control pin to switch the corresponding relay one way or the other

3) it's fed +/- 7.5V @ less than 1 mA , *easy* to get from the +/-15V you already heve for the Op Amps

4) it costs *less* than 1U$S [/color]

What's not to like about it?  <3)

Look at this Attax 40 schematic, where I show the 3 relays (in fact they leave 1 unused; if they wanted they might use it to control Reverb, some boost, whatever), the mechanical switching (foot and panel button), the RC "time delay / slowing" so you have no spikes (so no special ground needs), and the CD4053 internal logic block which turns simple +V/0V (ground) voltages at the control pins 9/10/11 into relay switching.
The relays are called "SLC 1/2/3" .

[stormbringer]

Yeah, i have both 4053's, a load of low current relays, and ldr's that i'm experimenting with. Im going for one clean, and two high gain channels.

There will also be a built-in maxon od808-ish boost circuit which can be switched in before the high gain channels to boost signal, and a delay instead of the classic reverb.

Now, theres quite a few components, but i really build this to suit the way i play. The footswitch can be used in two ways, either as a plain channel switch, but also programmable to activate "banks", ie. Button 1 = channel 2, button 2 = channel 3, boost and delay.

[Kaz Kylheku]

3) it's fed +/- 7.5V @ less than 1 mA , *easy* to get from the +/-15V you already heve for the Op Amps

4) it costs *less* than 1U$S [/color]

What's not to like about it?  <3)

That it does not run on the +/- 15V you already have for the Op Amps.