[sdiy] Protective Earth on PC soundcard GND?

[Magnus Danielson]

From: "JH." <jhaible at debitel.net>
Subject: Re: [sdiy] Protective Earth on PC soundcard GND?
Date: Thu, 27 Jan 2005 21:56:53 +0100
Message-ID: <003901c504b2$bc196d80$0200a8c0 at jhsilent>

Hi Jürgen,

> > But the other side, PC balanced TRS outputs to unbalanced RCA inputs on
> the
> > mixer, would only work if (*if* !) these TRS outputs would really be
> > "floating balanced" (as with a transformer, or with a special
> cross-coupled
> > opamp output). But I suspect "balanced output" all too often means just
> > complementary signals which are both referenced to GND, in which case a
> > tip/ring to RCA connection will be fatal (or at lest not break a ground
> > loop).
> 
> >There is ways to actively drive wich allows the drive to be diffrential
> while
> >he common mode may be controlled or indeed acidental as it would become
> with
> >ground loops and other imperfections in offset. Within the (little) range
> of
> >the op-amps it works, but the cost is headroom.
> 
> Yes that's what I menat with "floating balanced" (probably bad choice of
> words), the circuit putting out a well-defined differential signal whatever
> the common mode is. (The case of the negative output being tied to the GND of
> the receiving end included!) It's possible (withing the CM range of the
> opamps of course), but I have rarely seen it implemented!

Indeed. But there exist complete chips offering this. They are in use.

> *Mostly* you see just an inverted and noninverted output, both referenced to
> GND. Then I see no way to connect it without getting a voltage that exists
> between sending and receiving ground added into my signal.

I agree. But these buffers is used more and more since they are so darn usefull
and doesn't eat as much effort to get things right than standard components
leave you with.

> That's why im asking. I could buy the 1010 instead of the 1010LT, for twice
> the money, and get "balanced" inputs and outputs. Now it depends on the
> implementation of the "balanced" outputs whether I can break a GND loop
> (always talking about unbalanced I/O on the mixer!) or not.

Could be investigated, but I woul lean towards using the 1010.

Cheers,
Magnus