[TowerTalk] TowerTalk Digest, Vol 127, Issue 27

[Jim Lux]

On 7/11/13 5:53 AM, David Robbins wrote:
> actually.... you can have great grounding and still get this kind of
> damage if you don't equalize the incoming line voltage with ground at
> the entrance.  most of this damage was probably from what is called a
> back flashover.  this is where the ground potential changes relative
> to the power/signal wires due to a lightning strike to a tower or
> ground or power line ground.  in these cases the power/signal wires
> can be looked at as being at a constant voltage while the ground
> potential changes until it flashes over to the signal wire, in the
> case of homes this is often through your equipment power supply or
> modem on a phone line or cable tv box.  the fix for this is to apply
> mov's or other voltage equalizers between the power/signal lines to
> ground at your service entrance/single point ground so all the wires
> rise and fall at the same time.

Indeed.. a decent voltage clamp at the service entrance will insure that 
your power lines never go more than, say, 500V from the chassis ground 
(unless you've got big lightning currents flowing "through" the house or 
ground).  Almost all modern electrical appliances and such can take 
1500-2000V between line/chassis.

Your 1950s All-American 5 clock radio, perhaps not.
Your homebrew GaAs FET preamp, perhaps not.


>
> Low voltage stuff can still be damaged by another factor, induced
> currents from nearby strikes.  the longer the wiring attached to the
> device the higher the voltage.  preventing this type of damage
> requires mov or other equalizer on devices inside if they don't have
> them built in.  this is where surge arresters in power strips or
> ups's can help out since the energy is 'relatively' small compared to
> other events.


In order to induce voltages/currents from a changing magnetic field, the 
victim has to be a "loop" of some sort. Most often, the loop is formed 
by the "greenwire ground" between the two boxes in combination with the 
signal wires in the cable.  e.g. boxA chassis to threeprong plug ground, 
through house safety ground wiring, to three prong plug, to box B, to 
signal wires referenced to chassis, through cable back to Box A, where 
termination is referenced to chassis.  It's pretty easy to get a very 
large area inside that loop, especially if you're looking at things like 
network or phone wires running across the house.

If the only interruptions in current flowing in that loop are reverse 
biased junctions in the components at either end of the cable, you're 
looking a recipe for damage.

Transformer coupled Ethernet has the transformer insulation in the loop, 
which helps a LOT.

To a certain exent, double insulated wall warts (which don't require a 
ground pin on their AC supply) help with this.  Now the loop path has to 
traverse the parasitic C across the transformer windings, etc..

It starts to get more complex when you consider the propagation time for 
an impulse, but for lightning, the primary energy components are around 
1 Mhz, and at that frequency, even a house sized loop is still pretty 
small compared to 1 microsecond propagation delay.

A overhead power line to a building a mile away, though, with the other 
half of the loop being the ground between the buildings, and you've got 
a nice big loop.  But also one where the wave propagation needs to be 
considered.