From: John Larkin
Subject: Re: Grounding advice
Date: Sat, 12 Oct 2002 15:03:30 -0700
Organization: Posted via Supernews, http://www.supernews.com
References: <3D9DFA4D.firstname.lastname@example.org> <3DA726DB.email@example.com> <firstname.lastname@example.org>
X-Newsreader: Forte Agent 1.91/32.564
On Fri, 11 Oct 2002 23:19:50 -0400, Keith R. Williams
>In article <email@example.com>,
>> On Fri, 11 Oct 2002 19:30:47 GMT, Ryan Gammon
>> >I happened to come across the following in AoE (pp853):
>> >"... we suggest as many redundant ground connections as possible...
>> >... Don't worry about about "ground loops" in a digital or RF circuit;
>> >that's a microvolt audio circuit issue"
>> Now *that* is one of the few sensible statements ever made on the
>I took a course (professional, rather than academic) in EMI
>abatement, measurement, and standards a couple of years ago (I'm
>now an official EMI Guru (tm) ;-). One of the points was that
>ground loops are meaningless above audio frequencies. At RF
>frequencies the induced currents will cause loops even if
>connections don't. The theory then was to maximize ground
>connections, even in metal parts (unconnected gaps in metal,
>a.k.a. slots are antennas as much as wires are).
I may have to hire you to say that in front of one of my customers.
They're building a pulsed-power widget, and I'm doing the timing box,
and their Pointy-Haired-Boss (tm) insists that we ground nothing
because 'it would create ground loops'. He thinks cable shields should
be grounded at the source end only! By deftly (mis)managing a couple
of design reviews, I managed to ground *everything* to my metal case
and present a fait accompli (French for 'fooled you, Jack!')
prototype, which we don't have time to change.
Vision: a couple of dozen metal boxes, all isolated from the nice
solid-metal baseplate, connected by a spider's nest of crisscross
cables carrying power, logic, and low-level analog signals, all
ringing like the Bells of Saint Mary's every laser shot... nightmare!