[TowerTalk] Ground resistance measurements
Eric Gustafson Courtesy Account
n7cl@mmsi.com
Fri, 5 Jun 1998 10:58:14 -0700
>From: "Dick Green" <dick.green@valley.net>
>Date: Thu, 4 Jun 1998 20:25:19 -0400
>
>Coincidentally with all this talk about lightning protection and
>ground systems, I recently bought a used ground resistance
>tester. It's a professional model, AEMC #136.100. It appears to
>be in good condition, complete with documentation. It was such a
>deal...
>
>The manual does a good job explaining how to measure the ground
>resistance for a single ground rod using the Fall of Potential
>and 62% methods. However, these tests seem to be designed to
>measure the resistance to ground for a single isolated ground
>rod. You have to disconnect the ground wire from the rod, and
>pound in two test electrodes to do the measurements. The
>Polyphaser book shows a similar test setup.
It all depends on what it is that you are trying to measure. The
above is fine for determining the connection resistance of one
ground rod to earth.
>
>First, the manual doesn't say how long the test electrodes
>should be (i.e., how far into the ground they should be
>pounded.) Does anyone know? Can I use a couple of those short
>TV-type rods from Rad Shack for the electrodes, or do I need to
>use some 8' rods?
These rods are used to characterize the earth conductivity
constants so the ground system connection resistance can be
accurately determined. They do not need to be overly long.
Eighteen to 24 inches is typical. But they do need to be long
enough to get down into the "bulk earth" material. For example,
if the field where the reference electrodes are to be placed has
an 18 inch thick covering of pea gravel or decomposed granite,
etc. on it, 18 inches won't be long enough for the electrode to
actually reach the conductive earth.
However long they are, both should be driven an equal distance
into the medium being measured.
>Second, since all my ground wires are Cadwelded to the existing
>rods, I can't disconnect them. There's one rod attached to the
>tower where I could disconnect the lead at the tower base to
>isolate the rod, but that would be rather a chore.
>
And probably not necessary. It might be educational to
disconnect all feedline connections from the tower to ground and
measure the _system_ connection resistance to ground. Be careful
to observe the instructions about placement of the reference
electrodes more than 50 feet away from the nearest approach of
the ground system being measured. Also don't place them near
(especially not parallel to) another nearby unattached buried
conductive system (underground power, water pipes, etc.) as the
reference measurement will make an erroneously low estimate for
the ground conductivity. This would cause the connection
resistance measurement to be erroneously high.
Don't worry if you have to lengthen the conductors to the
reference electrodes or to the measured electrode to get the
reference far enough from the measured ground field. Thousands
of feet of wire won't materially affect the measurement if it is
in the reference electrode circuit.
>I'm wondering if any meaningful measurements can be obtained by
>taking readings on a ground rod that is connected to the rest of
>my grounding system. Anybody know? I'm guessing that the
>readings would be lower than for a single rod because the rods
>and radials would provide additional paths from the test
>electrodes to the rod under measurement. True? Would that number
>still be meaningful?
It should be lower. But it can be menaingful (see above). But
more interestingly, you can use the technique to do periodic
inspections of the subgrade conductors and bonds. Simply go the
the various above grade parts of the system and measure the
system connection resistance at each place. They should all be
equal. If they aren't, the connection between parts of the system
is suspect. The cause for this should be determined and
corrected before a big strike points it out for you.
At my last house, I was measuring the service entry ground and
the water service pipes for lightning ground resistance. Where
the water service came from the street to the house there was a
riser, and a "T" connection. The water came up into the "T",
then went from the "T" into the house and into an outdoor
fawcet. Once in the house, the pipe went back subgrade and ran
under the slab to the opposite side of the house where the
bathrooms and kitchen distribution were. The pipes were all 3/4
inch galvanized steel water pipe with threaded pipe fittings. I
thought the pipe from the front of the house to the city water
main about 85 feet distant would be a good ground since it was
deeper and longer than the one under the house. But I thought it
would be contributing the the connection snce it was in parallel
with the house from the place where I was measuring the system
resistance.
Well, I measured the resistance by connecting to the riser. I
was surprised to find a connection resistance of 35 ohms. Ten
ohms higher than an 8 foot ground rod I had just driven nearby.
Then I connected on the house side of the "T" connection. That
measured 1.1 ohms. The same as the service entry ground on the
other side of the house. Investigating the cause of this
revealed two things.
First, the previous owner had been forced to replace the feed
pipe from the city. He chose to use PVC plastic pipe for 90% of
the run. The only metal part was the riser sticking up out of
the ground.
Second, the riser pipe - even though threaded tightly enough into
the steel "T" to hold water pressure without leaking, was not
making connection to the house pipe system. Doing this
measurement saved me from depending on a disconnected part of the
water system for a ground.
>
>I have some spare ground rods. I suppose I could just pound one
>in and take some readings on it, but I'd have to try it at the
>various grounding locations on my property. That means pulling
>the darned thing back out or wasting a bunch of ground rods. Any
>ideas?
>
>73, Dick, WC1M
>
Don't waste your rods. Drive them where you'll need them and
leave them there.
73, Eric N7CL
--
FAQ on WWW: http://www.contesting.com/towertalkfaq.html
Submissions: towertalk@contesting.com
Administrative requests: towertalk-REQUEST@contesting.com
Problems: owner-towertalk@contesting.com
Search: http://www.contesting.com/km9p/search.htm