K8RI's assessment of the grounding problem is right on. I would like to
add a few comments.
Current waveforms for lightning simulation is not always the same.
Different waveforms are used for different purposes, depending on the
maximum waveforms expected in a particular application. Since towers
are right out there in the open, they see a variety of possible
waveforms. Cables coming off the tower have modified waveforms.
Here is some data on some measured lightning waveforms.
http://plaza.ufl.edu/rakov/ICLP2000/positive1.16.pdf
As you can see the pulse width measured can vary a lot, and also there
can exist a continuing current after the initial big pulse that may be
hundreds to tens of thousands of amps for periods up to 10 ms. So,
although you do have a lot of energy at 1 MHz and higher, there can also
a lot of energy present even down in the audio frequency range. Not all
lightning strikes have the same waveform.
Inductance of long cables does very little for these long continuing
currents. Hopefully most of this will go into the ground at the tower,
but there is going to be some traveling down the cables, no matter how
long the cables are. It is the low frequency components that are
responsible for most of the burning associated with lightning. Coax
shields won't take very much current before they melt. A large ground
wire will lower the impedance between the tower and the common point
ground, as well as reduce the voltage drop for these low frequency
components.
As most people know, but tend to forget at times, don't believe
everything you read. In some cases it may even be written by someone
who should know better. The Polyphaser book has several errors in the
lightning area. The most glaring one is in their calculation of current
distribution on guyed towers and in cables. By their calculations two
runs of RG8 down the tower would have the same inductance as a 35
inch-face tower. That leads to the conclusion that, to half the voltage
drop down your tower all you have to do is run two coax lines down it.
This should be obviously incorrect, even to a casual observer. You can
end up with some large errors if you neglect mutual inductance.
Also you can get some information about voltage drops by only using
di/dt, as Polyphaser did, but that doesn't give enough information to
draw real conclusions, because it considers only one part of a
waveform. As a minimum, a SPICE simulation will produce closer results,
but that too is only an approximation, because it also only considers
one portion of what is happening. To calculate something closer to
reality it has to be done by calculating the fields involved, probably
using the method of moments. Here is one article using that approach.
http://www.lightning.ece.ufl.edu/PDF/Kordietal2003.pdf
Notice you will have to modify what was done in this article by
substituting the appropriate waveform for the stimulus.
Jerry, K4SAV
Roger (K8RI) wrote:
>
>
>
>>According to literature from Polyphaser, if the ground systems are
>>separated
>>by more than 75 feet, they don't "see each other", so it doesn't make sense
>>
>>
>
>I think they are exagerating the 75 feet. Remember that key word
>"business".
>I think they make a good product and I use some, but I also think some of
>their information is a bit misleading.
>
>It is true when looking at rapid rise times they (house and tower) will see
>less of each other the farther apart they are located, BUT they still do see
>each other. Typical rise times equate to a frequency containing the major
>portion of the energy of roughly one MHz. Also if the cables tying the tower
>to the station ground consist of a series of ground rods and side branches
>it will bleed off or "sink" a lot of the energy into the ground before
>reaching the house/station. Therefore to me it makes good sense to tie them
>together.
>
>I've mentioned a number of times how often the big tower/array gets hit. It
>averages 3 verified hits a year. Tht is hits some one has seen. How many
>times it's actually been hit I don't know. I do know there are no
>unprotected coax connectors up there that have any plating left on them. and
>you can no loger get a PL-259 to screw onto the unused ports of the remote
>antenna switch. Next time there will be protection over the connectors.
>
>
>
>>to bond them. I believe the reason is that the inductance of the connecting
>>wire or strap is too great at that length to allow the systems to float to
>>the same potential.
>>
>>
>
>It's not so much the inductance but rather the wave length. If they are far
>enough apart the rise and fall times are are such that even were the path
>strictly resistive the delta T makes them appear as seperate entities for
>all practical purposes. That distance is one whale of a lot farther than 75
>feet. OTOH with a major strike, not one of the typical ones, the voltage
>difference across a few feet can be substantial. So, depending on how you
>view it, or the impression you wish to create, you could make the same
>statement for 10 feet as you can certainly get enough voltage difference to
>wipe out solid state equipment.
>
>
>
>>I hadn't read that tidbit when I installed my first tower system, so I laid
>>265 feet of 1/0 wire at the bottom of the trench to connect the tower
>>ground
>>
>>
>
>"To me" I'd say 250 feet is approaching where I *might* consider not tying
>them together, but more than likely I'd push to at least 300 feet. Even then
>I'd most likely make it a run consisting of a network to bleed off the
>charge which insulated coax in PVC conduit will not. From the base of my
>tower to the entrance groundging bulkhead is over 75 feet. I have two bare
>copper cables (one on each side) paralleling the conduit run from the tower
>to the house. Those cables tie to the grounding plate which is the single
>point ground for everything. There are runs perpendicular to those cables
>that run out into the yard as well. Most of the back yard is one big
>grounding network with 32 or 33 ground rods CadWelded(TM) to well over 600
>feet of bare copper.
>
>
>
>>to the single point ground at the house. I'm pretty sure that was
>>unnecessary (I like to think of it as a really long horizontal ground rod.)
>>When I installed my new tower this year, which is about 225 feet from the
>>shack, I omitted the bonding wire. I was convinced not only by Polyphaser's
>>argument, but also by the fact that the outer conductors of the two 1-5/8"
>>
>>
>
>Are those outter conductors bare?
>
>
>
>>runs of heliax between the tower and shack would undoubtedly present a much
>>lower impedance path back to the shack than any wire or strap I could lay
>>in
>>
>>
>
>Impedance is only a part of the equation. I think we rely far too much on
>just the "impedance" of that connection when for longer, well designed runs
>the majority of the energy is disipated into the ground.
>Do those outer conductors have the current carrying capacity of the wire or
>strap?
>Do they provide a way to bleed off the charge?
>
>
>
>>the trench. I'm hoping the tower ground system is robust enough to
>>dissipate
>>any surges that may occur.
>>
>>
>
>It will (In most cases) if the system is well designed and covers a wide
>enough area.
>
>Just remember a coax with an insulated jacket is a conductor straight into
>your house while the bare copper is a path to ground that also goes to the
>house.
>
>
>
>>73, Dick WC1M
>>
>>
>
>73
>
>Roger (K8RI)
>
>_______________________________________________
>
>
>
>_______________________________________________
>TowerTalk mailing list
>TowerTalk@contesting.com
>http://lists.contesting.com/mailman/listinfo/towertalk
>
>
>
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|