I routinely do something completely different to confirm
before I grab a wire - because I so hate electric shocks. I
even brushed the hot wire on safety ground wire. No sparks?
Then it should have been dead. Safety grounds were not
connected together elsewhere. I was 'had' by a multiple
failure. In another case, when the switch was flipped, then
two circuit breakers became connected. Flip the breaker, and
radio would go silent. Why? Defectively wired switch was not
in right position. Change that switch and circuit breaker no
long quieted the radio.

That wire tracer is a great tool. It says something is
seriously wrong. But no tools by themselves are sufficient
information. Other tricks and tools are required to explain
the anomaly or to find a wiring defect. The wiring tracing
buzzer is reporting something weird. No way around that. The
question is now, "Why is it suggesting that three breakers are
interconnected". If that question is not answered, then a
potentially serious problem still exists.

Even brushing the hot wire on safety ground AND measuring
with a volt meter was not sufficient. I still got shocked
because it was a multi part problem. If all tools do not
agree, keep looking for the problem. My problem? I assumed
there was only one fault. Literally got burned by the
assumption.

Chip C wrote:
These plug-in circuit tracers are handy enough but I have no doubt
there are lots of situations in which they indicate the wrong cables
back at the panel, so I'd never trust my life to them. Anything they
indicate needs to be verified by other means. For Roland's situation
it sounds like the tracer's signal is being induced into other cables;
this is pretty common (though three at once is worth bragging about).
Probably all the cables are close together somewhere, going through a
joist or a sill. In fact, they may be *too* close, for too long a
stretch, by current code, but it's hard to know if this is worrisome
enough to start tearing up walls.

Now, as to why the signal is not appearing on the right cable, I
dunno. But if you *know* that the breaker kills that circuit, then
clearly the signal *is* being blocked, so the reason why is really of
academic interest only (on a hunch, I'd try it with the fridge
unplugged). I would never discourage you from having it checked out,
but my gut feeling is that all is ok and the tester is just giving
wrong weird results.

As an aside, it's only a matter of time before your toaster and
electric frying pan gang up on your fridge and they all knock each
other out. If you're going to spring for an electrician, have a
dedicated line run to the fridge, and as a side issue have the current
circuit checked out.

But J.Doe has stumbled over true evilness. I have some knob-and-tube
in my house that I haven't completely figured out yet, and I have a
sneaking suspicion that the neutrals, if not the hots, are mixed from
different circuits. I have a little inductive tester that's my first
tool to check for hot wires; its negative results in low-sensitivty
mode (ie, direct contact) are mostly pretty accurate but it gives a
lot of false positives owing to induced voltages. I also have a
lamp-type device. I test them both on known-live circuits lots before
I trust their negative results. Both of these are cheap at most
hardware stores. Prior to touching a bare wire, *after* I'm "certain"
by all other means that it's cold, I hold it by the insulation and
carefully brush it on a neutral or a ground wire. This is a test of
last resort; the arc from a hot wire is serious business (the flash
could cause vision damage if you're looking at it, and start a fire)
but in my own opinion it does rank as slightly more survivable than
grabbing it with bare hands with your feet on the ground.

Chip C