Re: Electrical Wire Safety

[John Denker <jsd@AV8N.COM>]

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

> I have often heard that the grounded distribution
> system was developed from a safety viewpoint.

I doubt it.

> If, in fact, it is not safer, what happened?
> Was the grounded system developed for some
> other reason?

Speaking specifically about the _distribution_ system i.e.
the ~17kV stuff that killed the local pool-maintenance guy,
the answer is simple: less cost, i.e. one wire instead of two.

> Was the safety analysis done incorrectly?

Making cost/safety tradeoffs is an art form.  I suspect
if it were being redone today, the answer would come out
different.

The earth-return system is not just a cheap trick, it is a
dirty trick.  In many locales, the conductivity of the dirt
is not large compared to the conductivity of underground
pipes.  So the power companies are "borrowing" everybody's
pipes to use as the return conductor.  This commonly leads
to bizarre voltaic corrosion.  (It would be worse with DC,
but it's bad enough as it is.)  As a result of this (plus
other factors I suppose) nowadays people use electrically-
insulated pipe for many applications.

> Additional note...
>
> Today we can detect faults with GFCI devices.
> These work independently of whether the system is
> grounded or not by comparing the current in the
> two wires to the device, and breaking the circuit
> if the two currents are not equal-opposite.

Not quite.  That would work with a two-wire system
(more precisely, a wire-return system, as opposed to
an earth-return system) whether or not one side of
the system was grounded.

A GFCI cannot possibly work with an earth-return system.

> Prior to GFCI availability, faults were detected
> when a live wire contacted the properly grounded
> metal case of the appliance, and thereby caused
> the fuse or circuit breaker to blow.  It seems to
> me this was the main method of detecting single
> faults as they occurred as John has suggested we
> should do.

A grounded case "detects" and deals with one class of single
faults, namely faults *within* the grounded case.  But that
approach doesn't suffice to deal with other types of faults,
e.g. involving wires, cords, and things like lamps that don't
have closed grounded cases.

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

Also note that a GFCI is not a panacea.  It, too, is expected
to function in a scenario where there is a gross fault to
ground, whereupon it trips the circuit.  You should *not* test
or demonstrate a GFCI by grabbing a hot wire ... even if the
thing works perfectly you may discover that
 -- the "trip" level is set shockingly high, and
 -- the response time is not negligible.

In general, whenever you get a "new" layer of safety, you should
not let it tempt you into giving up the "old" layers of safety.
In this case the layers include
 -- insulation
 -- closed grounded case
 -- turn it off before messing with it
 -- isolation transformer
 -- GFCI
 -- etc.

and you should not lightly give up any of them.

For more on the general topic of layers of safety:
  http://www.av8n.com/how/htm/decision.html#sec-layers-safety