[Chris Morriss]

In message , Eeyore
writes


Arny Krueger wrote:

"Eeyore" wrote

You can in fact very easily have a 'hum loop' like that *inside* a piece
of equipment. I've measured potential differences of the order of
tens of uV

between different parts of a steel chassis caused by electromagnetic
induction from the stray field of a large power transformer inside the
equipment for
example.

That's one reason why we have separate signal grounds, even inside the
chassis.

This used to be very popular but is inherently incompatible with good EMC
practice.

I use chassis ground as my reference and I'm VERY careful about
stuffing current
into the ground ( i.e I avoid it like the plague ). As a last resort, I also
make some internal connections differential/balanced, which of course entirely
eliminates the complication of questionable ground potentials.

Graham


I have to agree there Graham. If you're going to have a ground in a
piece of equipment, make sure there's only one, and make sure that
everything that connects to it has the lowest impedance path you can
manage.
--
Chris Morriss