Re: Electrical Wire Safety

[John Denker <jsd@AV8N.COM>]

Quoting "Edmiston, Mike" <edmiston@BLUFFTON.EDU>:

> Fred Bucheit claims that thousands of lives would have
> been saved over the years if the power company had not
> chosen to ground one of the transmission wires.

Right.

The official name for this general idea is "redundancy".
A colorful metaphor is "belt and suspenders".  The point is
that it takes _multiple_ faults before you get into trouble.

If you think there is a one-in-a-thousand chance of belt
failure, and also a one-in-a-thousand chance of suspender
failure, and these events are roughly independent, then
you have roughly a one-in-a-million chance of a joint
failure.

This general approach is very, very widely used in the real
world.  For example, a small airplane engine has two spark plugs
in every cylinder, driven off two separate ignition systems,
even though only one would suffice if it could be made 100%
reliable.

> I think it is difficult to know if that is true.

I think Fred is exactly right, and I imagine it would be
straightforward to prove it beyond any reasonable doubt,
by analyzing accident reports.

> One thing we gain by grounding one of two wires is
> predictability.  If the power company did not ground
> one of the wires, who is to say that your neighbor
> won't ground one of them.  What happens when a device
> (like a clothes washer) goes defective and one of the
> wires makes electrical contact with the case?

This raises an important question ... but answers it wrongly.

Yes, it is true that in any fault-tolerant system, failures
are hard to debug by the time-dishonored method of waiting
until there is a gross breakdown and then taking action.

SO DON'T DO IT THAT WAY.

The airplane is equipped with a little switch that you can
use to shut down one or the other of the ignition systems.
You use it during preflight to check that each of the systems
is working normally.

Similarly, in the electrical system, you should engineer a
system for detecting when there is a fault in one side of
the circuit.  It would be absurd not to.

If you tolerate single faults, it is just a matter of time
before you get a double fault.  So you *must* detect single
faults and eradicate them before they progress to double
faults.

> If you assume neither wire is grounded, and therefore
> there is no danger to touch either wire plus ground,
> you are in big trouble if the other wire has become
> grounded either accidentally or intentionally by
> someone in your neighborhood.

Again this misses the point.  The point is to have *layers*
of safety.  You shouldn't lightly give away any of your
layers of safety.

Right now there is only one layer of safety, namely
  a) the instruction to not touch the wire.

Just because we add another layer of safety, namely
  b) transformer isolation

doesn't mean we need to throw away the previous layer (a).

Buying a belt doesn't require you to throw away your
suspenders.

> I think the predictability of the grounded system
> probably makes it safer than the ungrounded system.

That's somewhere between implausible and preposterous.

=================================

While we're on the subject of safety, let me remark on the
huuuge advantages of putting the wires underground.
 -- Less chance of people getting zapped
 -- Less chance of outages (e.g. storm damage)
 -- Less chance of death and injury due to cars hitting poles
    by the roadside (not trivial!)
... Not to mention esthetics.