[Top] [All Lists]

Re: [TowerTalk] Conduit for cables to the tower?

To: "'Jim Brown'" <>,"'Tower Talk List'" <>
Subject: Re: [TowerTalk] Conduit for cables to the tower?
From: "Daron J. Wilson" <>
Date: Thu, 14 Sep 2006 18:51:57 -0700
List-post: <>
> Twisted pair cable DOES reduced noise coupling in unbalanced circuits. The
> fact that adjacent twists exposed to nearly the same field are out of
> polarity and cancel is a VERY significant factor, and the tighter the
> twist,
> the better the cancellation. An input stage will respond to the difference
> between the two conductors, and the fact that it happens to be unbalanced
> does not change that.

My point was taking 12 'control' wires of unknown state and voltage and
blindly applying them to twisted pairs does not mean you are going to keep
signals where you want them.  For example in Ethernet, the twisting of the
pairs is definitely essential, and splitting pairs is death to the signal.
So, again, just the fact that the signals are on twisted pairs does not mean
that they are they are paired correctly.

> Virtually all big power amps that don't include 70V output transformers
> (these transformers are only used in industrial paging applications) are
> unbalanced, and it is well known that zip cord couples FAR more RFI into
> the
> output stage than twisted pair. The most common internal part coupling
> mechanism is the feedback loop around the output stage.

Yes I see your point, however, that is a logical pair to keep together.
Twisting two grounds together on a pair or two DC power control wires on a
different pair may not yield the same desirable result.  Again simply
changing the type of wire used from standard to twisted pair without regard
for the signals and individual pairs is not necessarily valuable.

> >As far as shielded cable, I strongly disagree.  We run literally miles of
> >shielded twisted pair for current loop controls.  There is a reason they
> >make it, there is a reason they sell miles of it, and that is basically
> for
> >the improved isolation from other signals by having a (grounded) shield.
> Ignorance of the basic physics is a major reason. The cable shield on a
> balanced pair works ONLY on the E-field and the EM field. It is of no
> benefit
> on the M field. Below about 200 kHz, virtually all coupling is magnetic.

Frankly I don't have a BSEE and I'm not a PE, so I'm not really in a spot to
tell the engineers I work for that they are wrong.  I can certainly pass
along your concerns to them, however, it won't change the engineered
solutions that we install as they design.  Oddly enough, they work, and more
interestingly often removing the ground causes erratic and dysfunctional
signals.  Go figure.

> I suggest that you study the papers by Bill Whitlock on the nature of the
> balanced interface.
> Also Henry Ott. Google on his name to find his excellent text, which is
> considered the bible on EMC. Henry comes from Bell Labs.

Excellent information, thanks for the references.
> >Having wired a few AM studios where the transmitter is in the building,
> we
> >exclusively used shielded twisted pair wire and even then it was a
> challenge
> >to keep the RF out.
> That's E and EM field. Noise coupling into control cables at baseband is
> all
> magnetic. And some of what you were doing (and some of the equipment you
> were
> using) may have been wrong. See the publications on my website that
> address
> "the pin 1 problem."  Also, take a look at my papers on Shield Current
> Induced Noise, which clearly show that by virtue of their construction,
> some
> shields significantly INCREASE the coupling of RF onto the signal pair,
> especially the type of cables like Belden 8451 that have a drain wire
> wound
> with the same twist as the signal pair!  SCIN is a very strong RFI
> coupling
> mechanism below about 10 MHz, and the pin 1 problem is a big deal from low
> audio frequencies up to at least the GHz range.

Also very interesting, not particularly applicable to control cabling, but
interesting none the less.

> >Your mileage may vary.
> No, physics is physics.

And yes, physics is definitely physics, but I suspect you don't know
absolutely everything there is to know about physics.  And unless you know
everything about physics and everything about a particular installation, you
cannot accurately predict the outcome.  So, it will vary.



TowerTalk mailing list

<Prev in Thread] Current Thread [Next in Thread>