Jim Brown wrote:
> On Tue, 18 Aug 2009 08:13:23 -0700, Michael Tope wrote:
>
>> BTW, I am not familiar with series suppressors
>
> Series-mode suppressors have been developed that are quite
> effective on power line branch circuits. They work by storing the
> strike voltage in a very large inductor, then discharging it
> slowly once the strike event has past.
One example would be a series L followed by a shunt C (yes.. a low pass
filter).. In a typical transient, the real problem is the high peak
voltage, not necessarily the actual energy in the transient. The low
pass filter spreads the big sharp peak out into a lower longer one.
Obviously, if you can set things up so the energy is dissipated, you're
better off.
However, consider this.. say you have a "300 Joule" surge suppressor..
You're running your electronics which is say a few hundred watts..That's
several hundred joules per second. Take that 300 Joule impulse and
spread it out over a couple seconds, and you're not really talking about
much.
NOTE... we're not talking about direct hits from lightning here.. that
is up in the kilo or megajoule range, and needs totally different
treatment. This is line transients of a few kV (UL 1449 says 6kV @
500A, 1.2x50 uS for voltage, 8x20 uS for current)
By the way, a new version (3rd) of 1449 goes into effect in a few
weeks. It's now called ANSI/UL 1449-2006
For phone lines and similar things (a rotator control would probably fit
this description), the appropriate standard is UL 497, 497A, or 497B,
depending on the application. The B standard is probably the most
appropriate for datacomm sorts of signals.
Eaton has a nice TVSS handbook on this sort of thing (googling might
find it) which talks about the waveforms, the 3 location categories
defined in IEEE C62.41 (branch circuits indoor receptacle, major feeders
& inside service panes, and service entrance & outdoor long lines)
The standards call out both a voltage and a current, because the
assumption is that the voltage will cause a breakdown or clamp and then
the transient source can produce substantial current. (i.e. testing
from a 10kV supply with a 10 Meg resistor in series isn't a very
realistic test...). The test equipment needed is not trivial, by the way.
A 1999 IEEE paper by R. Siegert and O. Mohammed describes one approach
(using a big capacitor, a triggered spark gap, and RLC networks for
pulse and impedance shaping)
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|