Start
• It's All
in the Grounding • Transient
Voldatge Supression • Modem
and Power-Line Protection • Unplug
the Computer! •
Automatic Thunderstorm
Switch
• Black Box
It
•
Waiting for Summer
•
Sources
IT’S
ALL IN THE GROUNDING
Scientists
still don’t fully understand lightning. Basically,
it’s a big discharge of static electricity that
flashes toward the earth along a pilot leader.
This
leader rushes down from the clouds in a series of
discrete steps, ionizing the air as it goes. The final
point is usually some elevated object on the ground.
The bright lightning discharge we all see is the return
stroke flowing back up the ionized path.
Any
protection scheme doesn’t prevent lightning from
striking. It merely provides a low-resistance path
for the lightning energy to ground. This path is the
real issue.
A
typical lightning bolt is 10–30 kA. The big strikes
are as much as 100 kA (the power industry uses a 100-kA
stroke with a rise time of 1.2 µs as its standard
stroke).
Even
if the path to ground were as low as 1 W, E = I ´
R tells us the DC voltage drop is 30 kV. If the resistance
to ground is greater, then the voltage potentials
are significantly higher.
Unfortunately,
less technical discussions on the subject don’t
include the disastrous effects of inductance in this
conduction path. Even with the massive lightning conductor
used in the typical building lightning system, the
inductance is on the order of 15 µH per foot.
The
inductive voltage drop on a 20' run of straight conductor
with the industry stroke applied is on the order of
one million volts! A conductor with lots of bends
and twists has significantly higher inductance.
If
the ground rod has a resistance of 10 W to ground,
that adds another million volts along the path. Together,
the total voltage floating around the building during
the lightning strike is two million volts!
The
voltage necessary to jump a spark through air is ~13
kV per inch, or 156 kV per foot. During a one or two
million-volt strike on a building, you have to be
careful about side flashes to any conductor that is
grounded but not connected to the lightning system.
That’s why conducting bodies like equipment cabinets,
machinery, metal rain gutters, and the AC electrical
system have to be physically connected to the building’s
main grounding system.
When
we casually speak of lightning taking the course of
least resistance, we’re talking about the flash-over.
When lightning hits the cable TV line and isn’t
shunted to ground via a lightning arrestor and surge
protector, the next stop is anybody’s guess.
Short
of putting up a tower to provide the proverbial zone
of protection, defense comes by providing a conduction
path with a lower overall impedance than alternative
paths through your computer and fax machine.
The
techniques are limited. The typical approach is to
space lightning rods on the top of the building and
use a heavy copper cable as a down conductor. Depending
on the slope and area of the roof, there are standards
regarding placement of the rods along the ridge versus
around the perimeter (flat roofs are the most difficult).
If
my house is any indication (part of the roof is shown
in Photo 1), overkill is the typical installation
choice. Counting the outbuildings, I have more than
25 lightning rods. The stranded copper cable is about
a half inch in diameter, and 100' of it weighs 20
lbs.
 Photo 1—Any effective lightning
system starts with a good array of lightning rods
spaced about every 20' along the peak. The insert
shows one rod in a little closer detail. |
The
fact that this is far from an exact science was illustrated
by the professional installer’s response to my
observations. I pointed out that I’ve seen systems
that employ sharp pointed rods as well as those that
use large spheres. I’ve also seen light-gauge
wire used as much as the heavy cable. His explanation
was fascinating.
Apparently,
there are two schools of thought in the lightning
business. Most follow the convention that lightning
hits the ground at a particular point because of the
charge built up in that area of the ground.
By
using very sharp points, the charge density at the
point becomes high enough to leak off this accumulated
energy, and it never gets high enough to attract a
leader stroke. Because this happens over a reasonable
period of time and at relatively low current, it also
reduces the need for heavy stranded cable.
The
other school of thought suggests that fate can’t
be deterred. If you’re going to get hit, so be
it. Just provide a good path to ground, and you’ll
be all right.
Round
spheres handle the high energy density of a direct
hit, and the heavy wire channels the load to ground.
OK, so why is he installing heavy copper conductor
and pointed rods, I ask? Insurance!
The
points still supposedly reduce the target potential,
but the heavy cable is there just in case that concept
doesn’t work quite as well as planned. I laughed.
The
rods and down conductors are only half the system.
It’s the total impedance to earth ground that
determines the voltage drop.
The
building ground should have a resistance from 20 to
50 W. For most applications, this level is achieved
simply by driving an 8–10' copper-plated steel
rod into the soil. Of course, the more conductive
the soil, the better the ground.
But,
my installation was at the other end of the spectrum—nonconductive
rock without a lot of deep soil. The only solution
was to create an artificial ground plane by burying
cable around the perimeter of every building, attaching
25' radial cables and ground rods (wherever they could
be driven) every so many feet, and connecting all
the building loops as one large grounding system.
Jeff’s
a little luckier. His house has a steeply sloped roof
and he only needs a few rods along the ridge. He also
lives a hundred feet from a lake, so he also doesn’t
have the ledge or the grounding problems I have.
However,
we both have a lot of sensitive electronics.
