> First we need to think of "ground" as a "system" of wire and ground rods.
> Unless it's the only ground rod, each ground rod doens't represent a "ground
> point" as it is no more than an *extension* of a grounding network or
> system. By themselves, the service entrance, phone line(s), antennas, and
> cable each represent a ground point where they enter the home which in this
> case would be 4 ground points. If allowed to remain seperate the equipment
> tied to them can rise to different potentials (sometimes drastically
> different) from a nearby or direct lightning strike.
Generally, codes
> require all of these be tied together to form one single system. Even metal
> water and gas lines are required to be tied into this system.
>
But not for lightning impulses... they're required to be tied together
so there's negligble potential difference at line frequency. 100 feet
of wire to connect them is perfectly OK, from that standpoint.
What you want for impulse protection is slightly different. you want
the chassis of all the equipment that is physically close together to be
referenced to a common point, and also, that an operator is also
referenced to that point. It's convenient that that common point also
happen to be connected to the power system ground, but not necessary.
For instance, you could have you and all your gear in a faraday cage,
powered by an isolation transformer with very high breakdown voltage or
off batteries.. I know someone who has sat inside the top terminal on
a very large (40 kVA?) tesla coil, while it's in operation, using an
oscilloscope and other equipment to make measurements.
http://www.lod.org/Projects/electrum/testing/pages/electrumeric.html for
instance
So what you're doing with that common point entry panel is trying to
make sure that everything inside the shack is referred to the same
potential (which may go up and down depending on the voltage drops in
the wires connecting to it "outside".
The tricky thing is that since it's unlikely your shack is a full
faraday cage (like the topload in the photo linked above), you could
still couple significant impulse energy into the equipment in the shack,
and this is where paying attention to loops comes in. For instance,
having that copper bar along the back of the bench and tying your signal
grounds to it, then dropping the power cords to the floor and running
them to a plug strip in the middle of the bench on the floor creates a
nice opportunity for a loop, which can pick up magnetic fields. Now, if
you had also bonded the chassis of the equipment to the big copper bar,
you've reduced some of the problem (if only because the AC line to
chassis insulation in most equipment is fairly good.. HiPot tests in
manufacturing are something like 2-5kV)
The typical ham station, though, presents some other interesting
problems. A lot of gear runs off 12V. So how do you connect up the 12V
power. Is one side of the 12V power supply hooked to your common point?
What about all those wall-warts? What about the circuit ground in
your PC?
Folks who do this for a living spend a fair amount of time working out
grounding diagrams for just this reason. The "bible" is the IEEE Emerald
Book (formerly the "recommended practice for grounding of sensitive
electronics", but now the word sensitive has been removed)
A couple weeks ago I was doing this at work for interconnecting a couple
racks full of test gear to a piece of equipment that will be part of a
space probe. We spent quite a while discussing the merits of isolation
transformers, whether the load side should be balanced (60-0-60) or not
(120-0), whether or not to bond the ground across the transformer,
whether or not you want isolation blocks in the RF cables, and if so,
should they just break the center conductor or both center and shield
(break both center and shield at 36 GHz, and the "isolator" makes a
pretty good slot radiator)...
Making some drawings and looking for where the paths are helps a lot.
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|