(2) dirty/clean ground again [sdiy]

Neil Johnson nej22 at hermes.cam.ac.uk
Tue Aug 20 01:17:18 CEST 2002

> >zero impedance :-).  Copper has a finite, if small, resistivity, and its
> >that which generates the voltage difference.
> Around here we call that Ohm's Law :)

Actually, people often forget the constant temperature bit about Ohms Law.

> >Now, look at op-amp circuits.  They draw (or dump) current through their
> >supply pins *only*.
> Depending on the circuit, they can sink or source significant output
> current, and off course the inputs wouldn't be inputs if they didn't...

Indeed, into/out-of the load.  As for inputs, yes a tiny current does flow
there, but e.g. in a FET-input op-amp those currents are picoamps or less.

> Well, monolithic caps are hardly expensive, and if "seeing" half the value
> were a problem, doubling the values would solve it.

Cheap??  12 pence for something the size of a small, round bogey.. bloody
ripoff!!  And my catalogues don't stock 220n miniature ceramics :-(

> Nor do I see what sort of signal would flow from one rail to the other.

Think of each rail as a low-impedance path to the PSU ground.  Sure, they
have some DC on them, but at AC they look like a path to ground.  Except
they only join ground back at the PSU, via the output decoupling caps.

> >Unless its for safety (in which case you have a separate circuit) ground
> >should only be used for reference, not carrying current.  That's what the
> >supply rails are for.
> Since this is synth diy, that's not very realistic. Let's look at something
> as basic as an envelope generator.  It could have a single-sided timer
> chip, a single-sided mux chip, voltage dividers for the pots, discrete
> semiconductors, and the timing cap, all connected to ground.

Oooohh *shudder* bloody awful!  Keep the ground for reference, not supply
current!  Better still (if you have no option), have a noisy ground for
unipolar chips, and a clean reference ground for analogue signals.
That's why wide (12+ bits) ADCs have separate analogue and digital ground

> >... Hence we put hefty local decoupling (10-100uF) to keep the
> >supply noise down.
> While that's a good value for a filter cap local to an entire module or
> PCB, a cap that large would not provide the high frequency decoupling
> recommended for the specified performance of many op-amps.

Indeed---a large tank per board, then local 10n or 100n beads per chip.

> >There's also the issue that placing them further away from their
> >respective device reduces their effectiveness.
> That's a separate, although valid issue.  Considering lead inductance it's
> another argument for using two caps instead of one between both rails.

Bzzt!  Wrong---now you have two sets of lead inductance in series, so
you've doubled the total inductance, upping the impedance even more.

> >Ideally, a multilayer board with a single, massive ground plane is heaven,
> >and then it doesn't matter so much where the decoupling caps are located
> >as there's such a low impedance from any supply pin to a decoupling cap.
> According to some, a board trace longer than 2mm is too long.  A small cap
> is useless if it is too far from the supply pins.

Designing ECL circuits is fun...every track properly terminated at both
ends, even if its only an inch long.  Oh joy, was that a fun trip if
ever...they brought me in to do the Philips ECL PAL (yes they do, or did,
make ECL PALs...what fun we had) on a mobile phone base station system.

> By "single-sided" I meant one of the supply pins is connected to ground.

Yes, I realised that, thanks.  I try to avoid the LM324, unless its for a
toy application with unipolar supplies, or I need that ability to swing
down to the ground rail.  And avoid that 555.  Use the lower-noise 7555
CMOS version.  And even then there's an argument for running it across the
rails rather than contaminating the ground.

Ah, dear me, time for sleep!

Night all,

Neil Johnson :: Computer Laboratory :: University of Cambridge ::
http://www.njohnson.co.uk          http://www.cl.cam.ac.uk/~nej22
----  IEE Cambridge Branch: http://www.iee-cambridge.org.uk  ----

More information about the Synth-diy mailing list