[time-nuts] GPS antenna and lightning

kevin-usenet at horizon.com kevin-usenet at horizon.com
Mon Oct 5 02:59:11 UTC 2009

```> I disagree - call me an old fart or whatever.  Your points are valid
> and proper logic, but every situation is just not the same.
>
> Bonding everything is proper, but without an effective ground, lightning
> will still seek a path to dissipate itself.  We are taught if the ground
> is not good to set up a  larger ground field to dissipate it.  But there
> comes a cost factor and trade offs.
>
> On a lot of mountain tops there is no way to get an effective ground.
> You cant drive ground rods in rock.  Laying cable on top of the mountain
> is also very ineffective.  Bringing ground up with power is ineffective.

Actually the principle is still the same: you have to make sure that
all alternate paths have something like 1M times the impedance of your
"main" ground path, or enough current will flow to damage something.

If your main ground is a few ohms, that's not too hard.

If you have a really lousy earth resistance (hard dry rocks), you can have
over a kohm of resistance to ground through your rod, and then you have to
have over a Gohm of impedance on alternate paths, which is a lot harder to
arrange.

Unless you arrange for a low-impedance shunt around any potential arc
paths.

If the earth ground sucks, the same principle applies, but this time you
have to do it to the whole building: it doesn't matter if the building
bounces to 1 MV, as long as it all bounces together.

You need to lay a conductive ground mat under your equipment shed, made
of some seriously heavy conductors.  I generally use a perimeter and
radials from the mast of 2/0 cable (256 uohm/m) to limit the voltage
drop of a 60 kA stroke to 15 V/m.

And remember that 15 V/m number when you think about grounding different
pieces of equipment to different points on the floor.

But this keeps the entre floor which your equipment could arc to at
a close enough potential that an arc can't form.

If you are laying a foundation slab, you can use the reinforcing steel
for the job.  This is called an "Ufer ground" and works very well.

> Large transmitters will also take higher levels of energy because the
> components are larger and able to dissipate more energy.  Most large
> transmitters use tubes and a high voltage device stands a better chance
> of surviving.
>
> Stopping the arcing is the mission, but unless you have a place for it
> to go, shunt, series or etc, it still will jump.

Yup, but it's going to head *down*, so if you can put a highly conductive
"equipotential layer" under your equipment, you can avoid having sparks
jump off your equipment.  Because there's a better path to that point on the
floor via the ground system.

The fact that the resistance to distant earth ground under your
equipotential layer is significant just means that the current will tend
to spread out to the edges.

One thing to watch out for is that any wire heading up the mountain will
need *multiple* overvoltage devices along its length.  This is a problem
telephone people discovered.   If the soil conductivity sucks, the entire
top of the mountain, including the top end of your wires, can bounce up
many kV.

That voltage is dissipated in the form of IR drop in the soil as you
travel away from the strike site.  But that dropoff is much faster than
in the highly conductive copper.  At a certain distance from the strike,
your difference betwene the two will reach the failure point of your
insulation, and zap!

But that's only the beginning.  That arc will equalize the potentials,
but that can still be many kV above distant ground, and as you go further
along the wire, the IR drop in the lousy conductor soil can reach the
insulation failure level *again*.

```