In article . com, " wrote:
For what it's worth, it's vanishingly unlikely that *any* of the cables in
the
box containing the GFCI and dimmer has its other end in the box containing
the
4-way switch. When controlling a light from 'n' locations, you need two 3-way
switches and (n - 2) 4-way switches; e.g. for 2 locations, two 3-ways and
zero
4-ways; for 3 locations, two 3-ways and one 4-way; etc. -- and the 4-ways
*always* sit in the middle, with the 3-ways at the ends. Power goes to one
3-way, and the light is connected to the other. Since the power for the
island
lights is clearly routed through the GFCI, there must be a cable leaving that
box and going to one of the 3-way switches -- *not* the 4-way.

Well, I'm no so sure about that. The island lights are on a different
circuit than the dimmer/ recessed lights. I have to turn off two
separate breakers to disconnect both.

Perhaps they *used* to be on two separate circuits... but the fact that the
island lights don't operate *now* unless the dimmer is switched on, too,
proves that they are on the *same* circuit as the dimmer.

I assume they share a common neutral.

Perhaps -- and if their hot wires came from two different breakers, this could
be *very* dangerous. It's imperative that the two breakers be on opposite legs
of the 240V service, otherwise the neutral could be dangerously overloaded.
This is a potential fire hazard.

(yes, this is where I'm glad I have a voltmeter - there is no
way to tell whether the power is off for the island light switch
otherwise)

Sorry, I guess I wasn't clear -- your diagram does not appear to represent an
electrical circuit that can actually function. Maybe I'm misunderstanding or
misinterpreting it. Or maybe there's a mistake in it -- have you
double-checked it to be sure it accurately represents the reality of what's
in
the box? (Referring specifically to the diagram of the GFCI & dimmer wiring)

I've checked it twice, but like I said, there is a LOT crammed into the
box that houses the dimmer and GFCI. I've lost daylight (chicken and
egg problem....) but will re-re-recheck tomorrow.


Yes, the recessed lights turn on regardless of whether the island
lights are on or off.
Interesting. And somewhat surprising.
What about the recessed can lights? Do they also go out when the GFCI trips?

No, the recessed lights stay on.

Interesting. And *very* surprising.

Does anything *else* go off when the GFCI trips? Any outlets that go dead?

Yes, there is an outlet on the other side of the kitchen that goes
dead. One of the three switches that controls the island lights is
above that outlet.

(I'm betting on 'yes' answers to all of the above)

Almost. That is what I found so puzzling about my problem originally
(GFCI tripping when island lights were switched on). They're on
separate circuits, the only common element is the neutral.

That's actually *less* puzzling. This is an oversimplification, but it will, I
think, get the point across: the GFCI doesn't "know" about loads that belong
to other circuits, so when the second circuit puts current on the neutral that
the GFCI doesn't "know about" it says "Whoa!! I didn't put that there!!" and
shuts down.

Some more info.
1. When the island light switch is OFF, the voltage for both input
("black left" in second diagram) and output ("black right" in second
diagram) is ~80V. It doesn't matter whether the dimmer is on or off,
voltage is the same. I believe this is normal for a 4-way switch when
you measure using a voltmeter

Well, it's common, anyway, when you're using a digital voltmeter, or a
high-impedance analog meter. You're reading, in all probability, a very very
low-current (milliamp level) induced voltage that is of no significance at
all.

Seeing a real current (several amps) at 80V anywhere in a circuit that's
supposed to be 120V is a _major_ red flag. I'll bet if you put a test lamp
across the leads where you're seeing that 80V, it won't light up -- and the
80V measurement disappears, too.

2. When the island switch is ON, the voltage is 0V for both input and
output. Again, doesn't matter whether the dimmer is on or off.

3. If any of the other island light switches are switched on, the
voltage drops to 0V as well.

4. When the dimmer is OFF, the voltage between the two black wires
attached to the dimmer switch is 120V. When on, the voltage varies
between 12V (maximum light output from recessed lights) to 100V (almost
no light output from recessed lights)

5. The GCI outlets supply 120V, regardless of any switch position.

2 - 5 sound normal.


Not to confuse the issue, but here's another piece of information.
There is a second set of recessed lights in the kitchen. The second set
is also controlled by a dimmer (in an outlet box that also has a GFCI).
The first dimmer is located closer to the island switch, so I didn't
notice this before, but....if I turn on the second dimmer, then the
island lights work, even with the first dimmer off. In other words, to
turn on the island lights, I have to turn on the island light switch
and EITHER dimmer.

Pair of 3-way dimmers.

Shot in the dark (no pun intended): There are a LOT of neutrals tied
together. I can't easily get to it, it's wedged tightly in the back of
the box. If there is a neutral loose, could that explain the symptoms?

YES.

Thanks again.

Good luck! And let me know what you find when you're able to check the wires
again -- especially that neutral.

--
Regards,
Doug Miller (alphageek at milmac dot com)

It's time to throw all their damned tea in the harbor again.