On 8/8/15 10:32 AM, JVarney wrote:
Grant Saviers wrote: >>I would like to be enlightened why #00 wire is
needed to attach to two ground rods which are likely several ohms
resistance to the "remote earth"? Especially when #6 was ok in the last rev
of the standard and when from an I^2*R and fusing current it would do the
job.<<
I'm only quoting the TIA-222-G code; I'm not aware of the facts or reasons
that motivated the committee to upgrade the grounding standards from
222-F. However, I will say that because the committee is composed of
experts and engineers from leading tower designers, tower erectors,
broadcasters and cell owners, I generally accept their judgment. Just as
the committee toughened standards for ice loading in response to tower
failures, I presume they toughened the grounding standards in response to
lightning-induced problems.
Beyond that, my personal speculation is that going with a big #00
essentially makes the radial wire act as a horizontal ground rod, thus
increasing the total volume of earth available to lightning to dissipate to
and of course lowering resistance of the system as well.
Also, I misquoted TIA-222-G: #00 is not specifically required, it says
connections between tower and rods shall be "not smaller in surface area
than 2/0 solid." I interpret that to mean that the spec could be met with
multiple strands of #4 or #6 in parallel.
I'm curious about why "surface area". That's important for skin
resistance, and would be real important if you were concerned about your
AM transmitter antenna efficiency.
Further, it is a bit of a mystery why Ufer grounding is not included
...This forum has opined extensively about why Ufer's beat ground rods. <<
Again, I'm not aware of the committee's reasoning. Just my personal
opinion: I don't see how the math checks out for Ufer grounds for towers. A
multiple radial and ground rod system might offer on the order of 10,000
cubic feet of earth for lightning to dissipate to. With an Ufer on a tower
with a 6x6x6 footing, that's only 216 cubic feet of volume available to
dissipate one trillion joules of energy. If I'm building a tower I'd rather
direct the lightning energy away from the tower versus concentrating it in
the footing, IMHO.
No where near a Terajoule of energy. Most of the energy in a lightning
stroke is dissipated in the air (somewhere between 10 and 100kJ/meter).
Typical peak stroke current is on the order of tens of kiloamps, and
with a ground resistance of, say, 10 ohms, that's a high peak power:
30kA@ 10 ohms is 9 GW. However, it only lasts for a few tens of
microseconds, so the energy is pretty small: around a megajoule. That's
big, but not huge (a car battery is 3-4 MJ).. about the same as 1/4 lb
of high explosive.
And most of the energy dissipation of a Ufer ground occurs in the soil
surrounding the concrete, just as the dissipation in a driven rod system
is in the soil. The important thing in all this dissipation stuff is the
current density: you want that as low as possible for a variety of reasons.
So a 6x6x6 footing has 180 square feet (about 26,000 square inches) of
surface area, through which 100 kA might flow (in a really big stroke):
That's a current density of 3-4 Amps/square inch, which isn't very high,
so you're not going to be getting boiling water, melting sand, burning
peat, etc.
Or, another way, shifting to metric units because it's easier.. if you
assumed that the concrete dissipates nothing, and all the energy is
dissipated in a thin layer 1cm thick around the footing: You're
dissipating a megajoule in about 170,000 cc of soil. If we assume that
soil is the density and specific heat of water, that's about 6
Joule/gram, or 1.4 calorie/gram, which would heat about 1.4 degrees C.
In reality, the energy is dissipated over a MUCH larger volume, both in
the concrete and around it, so the temperature rise from a single stroke
is pretty small.
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|