[sdiy] Re: stranded or not stranded

[jhaible at debitel.net]

> > One reason for using twisted wires (stranded, of course) might be
> > stability. But I see an even more obvious reason: to keep noise from
> > inductive coupling (magnetic fields) as low as possible.
> 
> This only works for balanced lines, where you want the interference to be
> coupled into both lines with equal magnitude.  And even then you want an
> even number of twists to ensure maximum cancellation.

In balanced Lines the effect may be *optimised*, but this doesn't mean
you don't get it in other systems, too.

If you have an alternating magnetic field, every loop will pick up
a little of this. The operation is a vector multiplication, so
by periodically changing the orientation of the area inside your
loop, you can reduce the induced voltage along your loop.
With the "even number of twists" you have the best cancellation,
because for every partial loop you have another one with opposite
orientation. But even if you have an uneven number of twists,
like 99, then 49 will find other 49 to cancel, and the remaining
single twist / partial area will not be compensated, but will
have much less noise induced because of the much smaller area
that is enclosed, compared to one big loop of a non-twisted wire.

I have no idea if this is *necessary* in the application we were 
talking about. But I presume it has some effect, and the effect
is positive, and it cannot hurt to have it, especially if it comes
almost for free.

> If its just signal and ground, the only benefit is keeping the wires neat
> (which is a good thing in itself).

In general, if you have the lines balanced towards GND (one going positive
while the other goes negative), you have the *additional* advantage of
a noise reduction from *electrical* filelds. But noise reduction against
*magnetic* fields comes from the twisting, electrically balanced or not.

JH.


-------------------------------------------------
debitel.net Webmail