Be careful with single point grounds.
If they are too far from the power source you can inadvertently create a
loop. What you really want to do is keep electrical loops small or
non-existent. The power and return lines to a device should not create a
loop. You can do this by twisting them together and making the physical
distance between the power and ground leads as short as possible. Signal
wires are especially prone to inducted noise since the receivers being used to
sample the signal are usually very sensitive. Keep those loops small,
using twisted pair really helps if it is common mode noise (i.e. both the
return and supply lines are radiated approximately the same).
The ground bus should run as close as
possible to the supply bus. Doing these things helps keep inducted noise
off the bus. Conducted noise is a different matter and is taken care of
by using some sort of filtering (capacitors, etc). Keep in mind that if
conducted noise is on a pair of wires that forms a loop, it can radiate and be
picked up by other leads that also make a loop.
Keep your loops small, just as small as
you can make them (twisted pair). Shielding helps too, but don’t
ground the shield at both ends or you could create a loop and the shield is less
effective (if at all).
My first impression about the problem
below is that it looks like it was a victim of inducted noise. I don’t
know what the probes look like, and I don’t know what the wiring was
like, but could it be that the wiring to and from the probes was acting like an
antenna more than the probes themselves?
Kevin Stallard
Legacy #300
From: Lancair Mailing List [mailto:lml@lancaironline.net] On Behalf Of James Cameron
Sent: Monday, December 11, 2006
2:02 PM
To: lml
Subject: [LML] Electrical bugs
These electrical bugs sound like grounding
problems. Running all ground connections to a
single point ground would probably clear up 90% of them. I like Bob
Nuckolls' "forest of tabs" ground system (sold by B&C), and since
using it in my last couple airplanes, have had very clean systems in terms of
noise and crosstalk. Once good grounding has been taken care of, routing
of wires is perhaps the next thing to look at. If antenna wires can be
separated, and high-current wires, especially alternator leads, can be kept
away from others, that may help. Finally, some equipment is just not very
well designed to filter out noise. Some years ago I had an E.I.
fuel quantity guage that would go nuts every time I keyed the mic to
transmit. Apparently the fuel probes were acting as antennas, and the
E.I. device had poor (or no) input filtering.