[TowerTalk] Fwd: Re: Re: lightening strike

[Jim Lux]

On 7/11/13 6:08 AM, Bill Weinel wrote:
>
> Hi Skip,
>
> Theoretically, you'd think ground potential should be the same everywhere.
> This is not true, because a point current source in the ground at some
> distance between two separated ground points will cause a different voltage at
> each ground point caused by the resistance between the voltage source point
> and the ground point by Ohms law [E=IR]. (Unless the distances happen to be
> exactly the same, and the ground resistance is uniform, the voltage measured
> at each point will be different.)
>
> Differential ground voltage rises are caused during a lightning strike by the
> surge arriving at two separately grounded circuits at slightly different times
> caused by the distance of each ground from the point source. This causes
> significant voltage differences (sometimes on the order of hundreds of volts)
> between the two circuit grounds for a number of milliseconds and thus causes
> damage to any loads connected between those circuits.

Considering that the overvoltage transient/fields propagate at close to 
the speed of light, the difference in time is more like microseconds for 
any reasonable sized house. (Even the Spelling/Ecclestone 50k square 
foot mansion  is only a few hundred feet across, and that's <1 
microsecond light time)

Difference in ground potentials can result from a couple of causes:

current flow through a resistive medium (e.g. the soil under the house). 
   Lightning hits or a power line falls in the back yard, and there will 
be a potential difference between front and back yards due to the 
current flow and the IR drop. This is often called "step potential", as 
in the potential difference when you take a step.  No time difference to 
speak of.

If you've got ham gear in the back of the house "grounded" to a stake in 
the back yard, and the electrical system is "grounded" to a stake at the 
front of the house, then there's a potential difference between the two 
stakes.

if the only path between the stakes is through your gear, then it fries.
If you have an alternate lower resistance path (e.g. a wire that's not 
something like 30 AWG magnet wire), then the current will flow through 
that wire, rather than through your delicate semiconductor junctions.
Note that the current isn't going to be all that high (no kiloamp 
currents), so you don't need BIG wire (the inductance limits the current 
and rise time).  Your goal here is to make sure that the voltage on 
either side of something fragile (semiconductor junctions in the 
equipment) is low. Not that your ham rig on the back porch is at the 
same voltage as the ground rod in the front rod.



>
> In your case, differential ground voltage rises is likely what took out all
> your gear.
>
> If you do some research on the subject, you'll find that lightning prevention
> experts recommend that all grounds in a building be bonded together to a
> common ground point. This point should also be the entry point for all
> services coming into and out of the building. This is known as a single point
> ground. The purpose of this is to prevent differential ground voltage rises
> during nearby a lightning strike.

Not so much "single point" as "equipotential".  You want everything in 
the house to move up and down together, and if there is a difference in 
voltage between places (due to finite propagation speed), that the 
current resulting from that difference goes some place "safe".

Lots of buildings don't have a single point ground, but DO have a ground 
grid.  Imagine if your house had a solid metal floor. A piece of 
equipment is placed somewhere on that floor, and the "exterior wire" 
comes in and is bonded to that floor next to your equipment (or the 
transient voltage protection is between the exterior wire and the ground 
"next to the equipment").  If this is the case, one side of the building 
might be at a different potential than the other (as current flows 
across that metal floor), but there is a small potential difference 
between a wire and the equipment, at any given location.

>
> While almost impossible to do unless your dealing with new construction, the
> goal in an existing building is to minimize the ground resistance between
> separate circuit grounds by bonding them together with large gauge, low
> resistance conductors.

Actually, there's not much need for LARGE conductors.  In the transient 
case, the inductance dominates and results in potential differences, and 
inductance doesn't depend very much on the size or shape of the 
conductor.  It's all about length.

In the steady state case (60 Hz power line falls on the antenna), the 
conductor has to have "low enough" resistance to keep the potential 
difference low enough to be safe.  The code required green wires and 
bonding wires (variously 14, 12, 10, 6, etc. AWG) are more than 
sufficient for this.   Note that the code requires different sizes more 
for mechanical reasons than electrical.

This is typically either copper strap or large gauge
> copper conductors. This means interconnecting the electrical, cable, and phone
> grounds together to a single grounding point with a heavy gauge low resistance
> conductor.

Strap is fine for RF grounds: a vertical antenna for 20 meters should 
use strap, because the AC resistance is much less than for the same 
amount of copper in a round wire.

Strap doesn't have any electrical advantage for lightning impulses or AC 
line safety. Strap might be more convenient mechanically: you can bolt 
stuff to a flat strap a lot easier than to a round wire.

The size is more determined by mechanical concerns: a ground bonding 
wire outside by itself needs to be bigger than one inside a conduit. 
Larger wires are also less likely to "flashover" to adjacent conductors 
because the radius of curvature is larger: a reason to use round wire, 
rather than flat strip, for lightning conductors, by the way.  A 1/4" 
diameter copper wire has a higher breakdown voltage than the edge of a 
0.020" strip of copper flashing. (5-6 kV vs 1kV)

The bonding wire needs to be big enough to carry the maximum expected 
fault current without melting.  The worst case for this is not 
lightning, but a local medium voltage power line shorting to your 
wiring, because it could source several hundred amps for seconds, before 
something trips.  Lightning has really high peak currents, but the pulse 
only lasts 50 milliseconds, so there's not much energy dissipated in the 
grounding conductor. (AWG 10 is big enough to handle all but the largest 
lightning strokes without melting)



>
> By doing this you can minimize the effect of differential ground voltage issues
> during a lightning event.
>