> > Would this be a recommendation from the broadcast industry? The code
> > (NEC-2002 at least) requires a somewhat larger conductor as a buried
> > ground (i.e AWG4, 20 fet long), The low voltage part of NEC requires
> > AWG10 (copper) for the "grounding conductor" (i.e. the wire from
> > whatever to the "grounding electrode")
> I don't think that they mention multiple radials. If you are using only
> one conductor as a buried conductor then that conductor is being asked
> to carry the full current.
> Don't get hung up on NEC codes. They have blanket codes that try to
> cover many situations. Often they are way overkill. Sometimes they fall
Until your house or building burns down, and someone asks: "Was the wiring
to code?" or if you try to collect on one of those surge suppressor "$50K
equipment replacement" guarantees.
I agree that code compliance is no guarantee that it will work, however, in
a lot of cases, there's a reason for the code, based on practical
experience. (There's a huge running discussion about this on one of the PE
discussion forums: If cost sensitive clients ask why are you doing anything
more than the bare minimum the code requires, what do you do?)
> >> It is even better with radials than just parallel wires as the radials
> >> afford more dissipation to ground being spread out. The ground does
> >> not get a chance to saturate as it can with only one or a few ground
> >> rods.
> > Soil is a resistor. It does not "saturate". There is a recommendation
> > that the maximum current density per electrode be limited to avoid
> > "smoking rods". For 8 ft rods, it works out to a maximum current of
> > around 500-1000 Amps in typical soils. Lightning is short duration
> > compared to other grounding requirements, and there's a square root of
> > time factor in the recommendation which might result in a factor of
> > 100-1000 increase in a lightning kind of application.
> Oh the ground is much more than just a resistor. It has capacitance and
> inductance as well. It also has propagation delays.
But, your oriignal statement was "the ground does not get a chance to
saturate", which implies a non-linear behavior. I maintain that (except as
you noted with respect to glassification below) this does not occur. There
is no timevarying behavior of the soil.
> I have never heard of a ground rod being smoked but it is common for the
> soil around a rod to turn to glass during a large strike because of
> arcing in the ground.
> I don't know where the 500-1000 amps per rod comes from. What do you
> think happens to the current in a 10ka strike if there is only 1 rod?
That's the IEEE recommendation on maximum design fault current to avoid
smoking rods. That particular requirement stems more from sustained faults
(shorts in distribution equipment, for instance). There is a note about
increasing the limit for short duration pulses.Lightning would be... 3-4
strokes lasting 50 microseconds or so over less than a second will put a lot
less energy into the rod/ground system than a sustained high current fault
(like a 480 V 1000 Amp feeder shorting to the ground).
> Yes the ground does saturate around a ground rod! A given area of earth
> around a ground rod can only dissipate so much energy in a given amount
> of time. That is one of the reasons for rods being spaced twice there
> lengths. The effective area of a ground rod is a diameter and depth
> approximately equal to its length.
No. The ground around a rod can absorb unlimited amount of energy (sure, at
some point the soil will melt, just like in an arc furnace). The "effective
area" you refer to has to do with the electric field around the rod in the
soil and the resistance of a rod (or combination of multiple rods). Rather
than have people exhaustively calculate electric fields for various
arrangements, there's a table in the IEEE spec and some rules of thumb to
aid people in selecting arrangements of rods that are economical and
effective. The basic statement is that the rods need to be twice their
length apart in order for the resistance of the parallel combination of the
rods to be near half the resistance of one rod. If you have two 8 foot rods
6" apart, the resistance is going to be almost exactly the same as one rod.
Put those two rods 16 feet apart, and the resistance will be half.
> > > Good lightning grounds do NOT necessarily make a good RF ground. The
> > requirements are totally different. No practical RF ground is going to
> > be asked to carry a current of kiloamps. A lightning ground might have
> > a DC (low frequency AC) resistance (defined in kind of a funny way, I
> > grant you) of 10 ohms (NEC allows 25 ohms), and be perfectly good for
> > lightning protection where the goal is to conduct the stroke current
> > somewhere "safe" (i.e. it doesn't result in high induced voltages or
> > flashovers to neighboring conductors).
> The requirements are not totally different.
> If a lightning ground is not a good rf ground then it is not a good
> lightning ground! A lightning ground may be a marginal one that
> satisfies the "code" but that doesn't necessarily mean that it is a good
> lightning ground.
> Lightning can not be thought of as just a DC current with a little AC
> component in it. It must be treated as DC and rf. The rf portion of it
> is very substantial at 1mhz and extends up into the vhf region.
But you could have a fairly high impedance (in the inductance sense) ground
for lightning and have it be perfectly safe (in the not burning down the
building sense), especially if the inductance is spread out over a large
distance and is relatively even, so there's not a huge voltage difference
along the conductor. Consider, for instance, the typical lightning rod on
top of a 2 story wooden house. The grounding conductor is a 25-30 foot
single conductor of wire from the air terminal to the grounding electrode.
The inductance of that would be huge (about 10 uH) and with lightning di/dt
of 20kA/usec, there'd be a potential difference of 200kV from top to bottom.
However, because the voltage difference along any small section is fairly
reasonable, it doesn't spark over, and there's no fire hazard. I don't
think anyone would say that that same conductor would be a good RF ground
for a transmitter on the second floor.
> > Soil does not saturate. The voltage rises because of resistance and the
> > stroke current and/or inductance and stroke current rate of change
> > (di/dt). The overall system is basically a big RLC... C in the
> > cloud/earth, R in the stroke itself and your grounding system, L
> > likewise. Your contribution to the system is basically the bottom tiny
> > part of a giant voltage divider. The lower the impedance, the lower the
> > voltage. Changing your impedance (either R or L) isn't going to change
> > the time waveform of the stroke a bit.
> Lightning is modeled as a current source. A certain amount of current is
> available in a particular strike. It does not matter what the resistance
> of the path is, The stroke is still going to develop that strike current
> amount in the path.
Yep..I'll agree about modeling it as a current source. Just like a big
voltage divider where the load is much smaller than the source impedance.
> By the way the cloud/earth path is not a capacitive one. It is a plasma
> path that actually has negative resistance. Thus the current source.
Nope. The resistance is positive (in that, everywhere, Voltage is of the
same sign as current). The slope of the V/I characteristic is negative,
just like in all high current arcs. Whether the source has a positive or
negative dynamic resistance (another way of talking about the slope of V/I)
doesn't make or not make it a current source. The appropriate model of a
lightning stroke is a BIG capacitor charged to a fairly high voltage,
discharged through a BIG resistor and middling inductor, into a RC load.
(this is the stroke model, not the leader model). To a measurement at the
"ground" end of the stroke, it will appear as a current source (in that the
source impedance is so high compared to the load impedance that changes in
the load impedance don't make much difference in the current).
> Yes the ground system has R and L. But it is also a transmission line.
> It has time delay. That is why Tom has large differences in potential
> between his ground systems.
I imagine just pure resistance would account for a lot of the potential
difference, at least as far as the gradient across the soil is concerned.
The soil is pretty low inductance (it's a big, wide, flat conductor).
> >> In a common lightning ground system installation it is recommended
> >> that ground rods be placed around the tower and separate radials run
> >> out to each ground rod from the tower. Additional ground rods would be
> >> installed at approximately the distance of twice their length on each
> >> radial to the same wire.
> > Is the recommendation a "generally accepted industry practice" or an
> > actual recommendation from a standards body (like NFPA or IEEE or
> > EIA/TIA??) I am curious if there is an actual published standard (I've
> > been looking for one, but haven't found it, but that doesn't mean that
> > it's not out there).
> A generally accepted industry practice. At least it is becoming so.
> > The recommendation that ground rods be spaced at least twice their
> > length IS embodied in several standards (IEEE 142, for instance) and is
> > based on both analytical models and field measurements (closer spacings
> > don't provide as much reduction in ground resistance).
> Closer spacing allows for ground saturation and the second rod is of
> less use.
Nothing special here. No saturation. Just application of simple electric
fields. The resistance of two rods close together is higher than two rods
far apart. It applies at 1 mA as well as at 10kA.
> > The recommendations that I've seen talk about a ring around the base of
> > the tower and several (not 60) ground rods. Perhaps half a dozen. And
> > the rods spaced twice their length apart.
> A "ring" connecting ground wires is a waste of wire. Think about what
> happens during a strike. The energy travels out away from the tower in a
> straight line. It does not make bends at the ring to go over to another
> ground wire or rod. All ground leads leaving the tower are at more or
> less the same potential as the stroke propagates. So the points that the
> ring is attached to are at the same potential already whether the ring
> is there or not.
I believe the ring recommendation is to reduce the possibility of high "step
potential" (voltage gradient on the surface) in the direction around the
tower. It's an inexpensive way to attempt to create an isopotential surface
inside the ring.
> > There's also a requirement that "every down conductor must be connected,
> > at its base, to an earthing or grounding electrode. This electrode needs
> > to be not less than 2ft from the base of the building" (p118, IEEE
> > 142-1991)
> Are we grounding buildings here or towers. Building down conductors are
> also required to be bonded to anything near by on their way down. Pipes,
> staircases railings etc.
Grounding anything. The requirement to bond to neighboring conductors is to
reduce the relative potential difference between conductors that can carry
appreciable current and might support an arc that is hot enough to start a
fire. The code requirements are designed to make sure the building doesn't
burn down, not to protect the equipment that might be inside. From that
standpoint, you don't care if the voltage rises or falls on a particular
piece of metal, as long as it rises and falls with the adjacent pieces.
> Also you mentioned before about corroding buried copper.
> Most ground rods are copper coated to prevent the steel rod
> inside form corroding. Copper oxide is a good conductor.
Very few (if any) metals have an oxide that is anywhere near as good a
conductor as the base metal. So I think, copper oxide (Cupric oxide, CuO)
is actually not a good conductor, at least not compared to most metals,
although it might be a better conductor than soil. I don't have any
references handy to look it up. Cuprous Oxide (Cu2O) it is a semiconductor,
and was used in rectifiers before the common use of Germanium and Silicon,
but is formed at high temperatures unlikely to occur in a wire buried in the
ground. I'd also venture to guess that the corrosion on copper is some
complex combination of oxides, sulfides, chlorides and carbonates and such.
See: http://www.mscomputer.com for "Self Supporting Towers", "Wireless Weather
Stations", and lot's more. Call Toll Free, 1-800-333-9041 with any questions
and ask for Sherman, W2FLA.
TowerTalk mailing list