According to :
In alt.home.repair wrote:
I will call the utility in the morning and see what they have to say.
I'll post back with further developments.
Short version: Utility claims they fixed something and the test results
are better (3.1 V difference as opposed to 9.4 V), but I'm still not
sure what exactly they fixed.
You did exactly the right things.
For future reference: power codes require that the voltage varies no
more than approximately 3-5% from a nominal 240V at the panel under
worst-case conditions (ie: max power on one leg, zero on the other).
Another 3-5% are permitted at the end of the circuits.
[Precise values of permitted variance will change from jurisdiction
to jurisdiction.]
[Note all of the below is assuming you're measuring the voltages on
the main bus bars. If you're measuring at a circuit-end, double them.]
What that means, for example, is that the leg voltage at the panel
on a given leg should drop no more than 4-5 volts even when max power
on that leg, and zero load on the other. Given resistance in the
neutral, _half_ of that would be the neutral "going towards" the
maxed out leg, and half of that would be the maxed out leg "going
towards" neutral.
In otherwords, (100A panel right?), drawing 100A on one leg, zero
on the other, the fully loaded side could drop as much as 4-5 volts
(hot-neutral), leg-to-leg drop by as much as 2-3V, and unloaded side
to neutral voltage could _rise_ as much as 2-3V.
If the neutral-leg voltage varies more than half of the leg-to-leg
does (see (1) below), you have a loose neutral.
All that said, there are several caveats:
1) Extreme caution should be taken when taking the voltage levels
off your voltmeter. There are a number of ways they can misread,
and many voltmeters are simply not that accurate at measuring RMS
volts. So, don't take a few volts either way seriously.
2) While voltmeters can be fooled by back EMI (line noise), for
the most part it will be minor. To minimize these affects, use
resistive loads (rather than motor or electronic). Ie: heaters.
3) While it may seem that you were approximately "on" those
numbers, remember that your test loads were MUCH less than 100A,
so the problem you were seeing was much worse than the allowable
variance. Frankly, if your wiring was good, given the loads
you were able to test with, I'd expect the voltage variation to
be a volt at most, but, generally speaking, you can't rely
on your voltmeter to that high a degree of accuracy.
Most times the variation will be extremely obvious. You were
on the low side of "obvious" [+]
While your "fixed state" is _still_ outside of what I suggested
above, it's a considerable improvement over what it was, and I'd
be doubtful of taking it too seriously. That said, keep an
eye on it.
[+] The high side of obvious is having a lightbulb on one
leg expire with a BRIGHT flash when you try to turn on
something on the other leg - and that "something" doesn't work.
--
Chris Lewis, Una confibula non set est
It's not just anyone who gets a Starship Cruiser class named after them.