N4CC and I have finally completed our remote Internet station in Hilliard,
FL. Probably half our time spent before and during construction was
studying and implementing lightning abatement. Ground rings were created
around each tower, and the communications shelter. All grounds, including
the adjacent electrical service ground are brought together at an external
ground bus (EGB). In total, 55+ ground rods are used with four rods each 24
ft. Lines are bonded at the top and bottom of the towers. We do not
disconnect anything. Everything stays up and running 24/7 regardless of the
WX.
Poyphaser rotator MOV protection devices are installed at the base of each
tower. At the EGB, we're using an Array Solutions model. In looking at all
the failure modes, I missed one: If the MOV on the rotator return lead was
to short to ground, it creates a disastrous situation where the prop pitch
will turn freely beyond the electrical stop point. The MOSFET is controlled
by a PWM circuit and the duty-cycle is what varies motor speed. A shorted
MOV will cause the lines to break apart up the tower at the coax loop. The
prop pitch would stop at nothing and keep turning. Do the MOVs in these
units generally fail open or closed?
We're using a pair of M2 PCX2800 controllers. To help protect the MOSFET
device from a similar "short-to-ground failure" during a lightning event, I
designed a circuit that engages a vacuum relay such that the MOSFET is only
exposed to the outside world during rotation. It's not a 100% guarantee
against MOSFET failure, but should help to mitigate damage. This is an
add-on that's similar to what K7NV did with the Green Heron prop pitch
controller model.
Here's my thought for a potential fix: While another vacuum relay could be
added on the +48V supply side to the prop pitches, I could isolate circuit
ground from chassis ground on the secondary side of the controller's power
transformer. The primary would still be safety protected and meet UL.
Isolating circuit ground from chassis ground would inhibit rotator turning
in the event of a MOV failure on the return line.
Anyone been through this? I welcome comments on any better ways to manage
this.
Paul, W9AC
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|