Re: twisting motor cables

> Would it be correct to say that signal line are more suseptible to
> noise because it is only 5V while the motor cable may be way above,
> say, 30V? What about motor cable as a source of noise?

Yes. The factor is the S/N (signal to noise) ratio. Also the input
circuitry is much more sensitive in most systems than outputs. Much
like an audio amp (poor example but best one I can think of) you have
to shield inputs but the speaker outputs are unshielded cable. In a
real rough sense we have "chopper" systems that operate in the mid to
upper audio range. Granted, the waves are square rather than sine
and contain a rich set of harmonics, but motors are real forgiving
about the shape of the wave and act as big inductors to help negate
the higher harmonics. Any time you have high current and voltage
moving a heavy load there is the possibility of noise generation.
Running wires (mini inductors) next to each other and pumping large
amounts of signal through one, will induct a signal in the other.
Wires have to be close to have their magnetic fields overlap and the
longer they run in parallel the better the cross coupling (and signal
transfer). Analog circuits that have amplification built in have
much lower S/N requirements. Digital circuits have much higher noise
immunity. Of note here: Wires in the same shielding (metal conduit)
have better coupling than in open air. There are other factors as
well but do not normally come into play at non RF frequencies.
Things get really weird up in the areas above 500MHZ where even short
pieces of wire become an antenna (like PC board runs)

> One of the pair goes to ground. Does the return current always

return

> via th other pair, or is it possible that the return current may go
> via any of the numerous other ground wires?

Current flows in both conductors in equal and opposite directions. If
the circuits share a common ground than the common ground carries the
sum of the currents. The most current will flow in the "closest
ground" wire meaning the one with the most direct path back to the
common ground. It can also flow in parallel grounds much like
parallel resistors. Using twisted wire around a common ground is
pointless (and may do more harm than good.) If you use a
conventional bipolar bridge motor controller (Gecko's and most
others), the motor wires do not share a common ground (at least on
the motor end). If you want to twist wires, then twist each motor
PAIR together; it does not matter that more than one of the wires may
be connected to the same ground at the power source. In critical
circuits the designer will often make it a point to have a central
ground at the point closest to where ground is established
(Transformer centertap, bridge negative, Large filter cap negative,
etc) and have seperate ground runs from different parts of the
circuit. Ground is relative. The further you are away from it the
more signal that can be induced onto it. If lightning strikes close
to a ground running into a building you can get hundreds of volts
inducted in the short run. Electrical equipment can be presented
with over voltages and/or reverse current flow...a BAD thing.

Confused? Ground loops, noise, and EMI are the bane of designers.