On Thu, 2010-10-07 at 12:04 -0700, Jim Brown wrote:
> On 10/7/2010 11:14 AM, Don Moman VE6JY wrote:
>
> >The ISP installation has 4 runs of CAT5 shielded (drain and foil grounded at
> >bottom only) to their equipment at the 140' level.
>
> That's a problem -- to be effective, the shield MUST be bonded to the
> shielding enclosure at both ends. A shield bonded at one end only on a cable
> longer than about 1/20 wavelength is no shield at all.
I am the CTO for a wireless ISP and we ran into an issue of interference
to low band VHF on one of our shared tower sites that turned out to have
an easy remedy
We always use shielded cat5 to tower mounted kit, Our Tower mounted kit
is always mounted in shielded metal housings of our own design and
mounted using metal stand-offs, nuts and star washers to the metal
backplate however the cat 5 enters the housing via gasket and is then
grounded via the RJ54 Jack body.
Unknown to us at the time one of our (now former..) vendors had decided
(for reasons I could never get a sensible answer on ) to leave out the
zero Ohm jumpers that ground the RJ45 Jack shells and the ceramic SMT
capacitors that decoupled the two cat5 pairs used for Power over
ethernet on an updated version of their product. As a result there was a
high level of trash present as common mode on the outside of the Cat 5
runs connecting the affected boards. The source of the trash was the
Switch mode powersupply/regulators on the boards themselves, not the
100M Ethernet data
The fix was rather simple and consisted of a short piece of braid
soldered to the metal body of the RJ54 Jack and the grounding point next
to the mounting hole. This reduced the noise at low band VHF by around
30dB. The addition of 2 clip on ferrite chokes (one inside the housing
and one outside reduced the noise to the point where I could no longer
measure it along the outside of the Cat5 runs even with a sensitive
receiver and near-field probe
Putting the cat5 runs inside conduit clamped to the tower may (or may
not) work but I would be fairly certain that there are easier and
cheaper cures
almost all (all in many cases) noise associated with cat 5 runs is due
to common mode trash coupled onto the runs by the equipment at one (or
both ends) and resolving this is the best way to solve the problems.
Consumer and pro-sumer routers and switches use noisy wall mounted
Switch mode supplies along with further switched mode regulators inside
the units , there is often little thought given to good EMC friendly
design, perhaps because EMC compliance testing may not require
measurements to be made with Ethernet cabling actually attached to the
unit
--
73
Brendan EI6IZ
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi
|