PSN-L Email List Message

Subject: Grounds!
From: Mike Lozano mikel@.......
Date: Fri, 20 Aug 1999 13:57:07 -0500


I've been following the discussion on 'grounding' with a great deal of
interest.  What's becoming plainly obvious is that there are many
misconceptions which are being perpetuated in this forum.  One of those
misconceptions is that 'grounding' is a D.C. phenomenon, measured as a
resistance.  It's not!  When speaking of  'ground resistance' we're
actually speaking of 'ground impedance' or resistivity.  This
'resistance' can only be measured with special equipment, using the
'fall of potential' technique.  This type of equipment is manufactured
by companies such as Biddle and AEMC, to name but a few; and, it may be
possible to ask your local power company or lightning protection company
to check your ground rods for the proper 'resistance'.=20

The resistance issue has clouded the field of grounding, and is best
demonstrated by the 'more is better' school of thought.  These
proponents hold that if one ground rod is good, then multiple ground
rods are better =85. the 'more then merrier' school.  At first blush, thi=
s
technique seems to make sense.  The confusion stems from confusing
'ground resistance' with 'resistance' in D.C. analysis.  While in D.C.
analysis, resistors in parallel decrease the total resistance to less
than the value of the least one, this is most definitely not true when
multiple ground rods are considered!  In fact, when driving multiple
ground rods; they should be driven no closer together than twice the
length of a single rod.  Although beyond the scope of a short note, the
reason the 'more the merrier' is not a valid ground resistance reduction
technique has to do with mutual inductance between the rods.

As to the placement of signal leads in conduit, care must be taken that
ground currents flowing through the conduit walls don't couple currents
into the conductors encased within.  In short, this technique creates a
linear transformer!  As to the prevention of dangerous voltages reaching
a sensitive instrument, a multiple protection scheme must be used. T
he principal idea is to provide both 'common' and 'differential' mode
protection.  For example, take two signal lines referenced to 'ground'.=20
Common mode transients take both lines 'high' in respect to 'ground'.=20
Differential mode transients, on the other hand, take one line, or the
other 'high' in respect to the other two.

If it'll make it a bit clearer, consider the case of a common A.C. surge
suppressor.  It you take one apart, you'll find three Metal Oxide
Varistors, or MOVs inside.  One MOV goes from the Black wire to the
White wire. This is the differential mode protector. The two MOVs going
from the Black and White wire to the Green or 'ground' to provide common
mode protection.  The exact, same principle can be used to protect any
number of signal lines.

One fellow asked if replacement boards are to be considered a normal
expense.  The answer is, most assuredly not!   I've designed and
installed the common/differential  protection scheme to protect several
Doppler Radar installations in the Midwest, as well as in the South.=20
These radar systems are located on frequently struck television
transmitting towers.  At from one million to several hundred thousand
dollars a pop, my clients simply can't afford to replace these systems;
but it's simple to replace the protection circuits about every five
years, or so.


As to the use of zener diodes - they are fast, but not too sturdy.   In
the cases mentioned above, I've used  triple protection!   A current
limiting agent in the form of a fast-acting thermistor; an MOV and a
gas-filled spark gap.   I've found that the MOV (only slightly slower
than a zener) is more than adequate protection.  For most signal lines,
I use an MOV rated at 12 Volts, with the highest possible Joule
ratings.  As for the gas-filled spark gap, I use a CG75L manufactured by
C.P. Clare Co. This small device fires at approximately 60 volts, and
while slower than an MOV, can handle tremendous amounts of power.

Of course none of these protection schemes will work well without a good
quality grounding system.  For the industrial installations I've
performed, ranging from petroleum transfer stations, to 250 foot tall
water towers with telemetry  and two-way radio antennae at the top, I
insist on no more than five ohms of ground resistance as measured with a
Biddle ground resistivity meter, using the three lead, 'fall of
potential' method.=20

Just plunking a rod in the ground does not guarantee a good ground!  If
you want to construct a ' relatively good quality ground', dig a hole
about the size and depth of a five gallon bucket.  The ground wire
should be NO SMALLER than  #4 copper stranded wire.  Attach that to a =BD
inch to =BE inch thick grounding rod, at least 8  to 10 feet long.  Drive
the rod into the ground so that the top of the rod is level with the
bottom of the hole.  Then, dig a channel about three inches across, and
three inches deep around the circumference of the bottom of the hole.=20
Into the channel, pour ice cream salt until the channel is filled.=20
Then, fill the hole with dirt;  moisten it down and then tamp it down.=20
Every few days add water to make sure than the salt dissolves slowly. =20
The salt will enhance the ground's 'conductivity'!  =20

When connecting to the ground lead, make sure there are no kinks or
short bends in the cable.  Kinks and bends represent inductance, which
tends to lessen the quality of the ground by raising the impedance.=20

If there's any interest, I'll try to put a JPG diagram of the protection
scheme I use, on my company website at:

	http://www.sciencearea.com

I also have a small quantity of gas-filled spark gaps left over from a
project.  I'd be glad to make these available to the PSN members at
cost, plus shipping.

Regards to all,
Mike Lozano, Meteorologist / Engineer
NSLI Certified Lightning Safety Professional
Applied Sciences, Ltd.
-0-

_____________________________________________________________________

Public Seismic Network Mailing List (PSN-L)


[ Top ] [ Back ] [ Home Page ]

Larry Cochrane <cochrane@..............>