Re: [CAD_CAM_EDM_DRO] Re: Proper size of capacitors for transient suppression?

Andy Wander wrote:

>Tom:
>
>I agree with almost everything you've said(in case you care :)), but I have
>to take issue with the idea of
>
> "send voltage and signals down twisted pairs or wires where one is a
>ground. Twisting wire tends to cancel out noise since the wires act like a
>winding on a transformer wire with bucking coils"
>
>As far as I know(I'm an audio engineer), the only way this will work is if
>the input circuit that the twisted pair feeds is a differential input, which
>could be a transformer, or an active-balanced input, either of which
>compares the 2 input lines and generates a signal that is the DIFFERENCE
>between the 2 lines(differential). This acts as "bucking coils" only
>because the noise induced into the pair is the SAME on both wires, and the
>input looks for the difference. If there is some noise that's the same on
>both wires, it won't be seen by the differential input.
>
>If you have a twisted pair with one wire of the pair grounded, then any
>noise induced into the grounded wire will be shunted to ground, and I can't
>see how the receiving circuitry would know it was there at all, much less be
>able to use it to "buck" the identical noise in the "hot" wire.
>
>I am curious to know if you or someone else on this group has actual
>experience using this technique, and can say that it actually makes a
>difference to run a twisted pair with one side of the pair grounded.
>
>

I think it does work, to some extent. The idea is that most of the signal's
return current will flow through the closest ground conductor to the signal
conductor. This will make the loop area smaller than having the ground
conductor loosely associated with the signal wire. It totally defeats the
twisted-pair nature of the cable, of course. I am sure I have seen a
positive
effect when extra wires in the cable were used as additional ground wires.

There is no doubt that a truly differential signal sent down a proper
twisted
pair to a true differential receiver is better than any single-ended
foolishness.

>In theory, it should not work-but heck, I know that theory doesn't always
>explain everything.
>
>

Yeah, when you get into the arcane area of ground loops, and there are
25 possible
paths for a signal's return current to flow through, it gets wierd and
un-intuitive.
Modelling this sort of stuff can be done, but it is usually so hard to
know the real
impedances of all the connector pins, wires of different gauge and geometry,
etc. that it can be very hard to analyze it completely.

Jon