SV: [Elecraft] KIO2 noise problem

Sun Dec 21 02:28:00 2003

Ok Don,
I remember when working for a computerfirm that RS232 cables were
made the way you suggest.After some years it changed so the outer shield
was connected,soldered, to the metalhousing of the the DB25 at both
ends.
The other way makes it uncontrollable and you may end up in a endless
loop of job trying to get rid of radiation from big computerinstallations
which may give us hams QRM.
Sometimes if a computer to printer has a long cable-run and there is
something upsetting the communication so the printer misprints it
can be a good idea to gnd the shield of the cable at points along its
route and also put ferriterings around it here and there. Simular with a
long coax for TV which gives you ghost signals can be cured by gnd-ing the
coax here and there along its route.Was the screen not tight ?
73
Rolf SM0CLS

-----Ursprungligt meddelande-----
Fr�n: Don Wilhelm [mailto:[email protected]]
Skickat: den 21 december 2003 02:20
Till: rolf heidenborg; [email protected]
�mne: Re: [Elecraft] KIO2 noise problem

Rolf,

The ground loop problem has more to do with creating additional paths for
the signal return than it does with circulating ground currents due to
chassis potential differentials.

It certainly is not dead and gets a lot more attention in computer fields
when switching levels are low and also in data transmission arenas where it
can create a DC level offset problems as well as couple noise to unexpected
places.

Most communication cables contain one or more signal return wire(s) that is
not the shield.  Proper device design and adherance to accepted signalling
conventions determine which end of the cable can tie the two together.

If part of the signal return path uses the shield of a cable (not
necessarily by intent), that signal can radiate and defeat the purpose of
the shield.  I guess in ham radio terms one could relate this condition to
common mode RF current on the outer shield of a transmission line which will
cause the transmission line to radiate - one good cure is to use a balun (or
other RF choke) at the end of the coax to electrically decouple the shield
at that point - such decoupling is the electrical equivalent of
disconnecting the shield (at least at the frequencies of interest).

73,
Don W3FPR

----- Original Message -----

> Hi,
> I think the talk of just connecting one end of a shield
> was a myth and dead by now ?
> I think shields shall be connected at all ends all the time
> and one shall try to keep the two ends gnd potential
> so close to each other that the shield does not have to curry
> gnd current due to the difference in chassis potential ?