MrC1 wrote:
I have a 50 gal gas water heater, which is, of course, in the attic where it
can do the most damage if I spring a leak. I recently lost a 1'x1' patch of
ceiling texture and paint below the heater, so I went up to check it out.

I found that it had a very slow dribble-leak from the steel nipple fitting
at the top of the tank on the 'out' (hot) side. The reason for the ceiling
damage was that the morons that installed the drain pan barely tightened the
PVC to the pan and didn't use any sealer whatsoever.

I sealed up the pan and then got to work on the leak. I unsoldered the
copper so I could work with the steel fitting. I removed the fitting and
the threaded copper adapter, which were pretty corroded, so I replaced them.
I used plenty of teflon tape on the steel pipe threads and cranked it down
to spec: "dang tight". I then soldered the copper back together and let
'er rip. No problems for several days, but...

Last night I checked the ceiling and it was dry. I then went up to check
the water heater and sure enough, it had an identical leak! (ARGH!) It's
dribbling from the base of the steel nipple fitting right where it threads
into the top of the heater.

Any ideas how to stop this freakin leak? I've got the fitting tight enough
that I fear if I try to put another turn on it I might strip or split it,
that is, if I'm able to get it that tight at all. It's really tight now.

Thanks!!!



I wonder if that galvanized fitting has a pinhole leak right through to
the root of the first thread which is exposed above the tank boss?

(Excuse me if I'm repeating myself and told this tale here before.)

I had one go like that on me a couple of years ago, about three months
after I installed a brand new heater and used a couple of dielectric
unions between the copper supply and discharge pipes standing on 2-1/2"
long galvanized unions because some "expert" told me that was the "best"
thing to do.

A leak similar to what you described occurred and I went through a
disassembly, retaping and reassembly to find the damn thing still
leaking from the same spot.

When I pulled things apart for the second time I found both nipples were
already about half clogged with rust and badly corroded on their insides.

I sliced the "leaker" in half and here's what it looked like after I
wirebrushed off the soft rust:

http://home.comcast.net/~jwisnia18/temp/nipple.jpg

That's when I put my brain back in gear and realized that those galvanic
couplings don't do anything other than make money for the sellers if the
copper water piping is grounded (per code) and the steel water heater
tank is also grounded (again per code) because the two ground
connections provide a low resistance galvanic current return path across
the insulator in the dielectric unions.

I searched around the web and found that Rheem, maker of zillions of
water heaters, had already figured out that dielectric unions were
dinosaurs and wrote a tech bulletin on it:

http://home.comcast.net/~jwisnia18/temp/Rheem_union.pdf

They missed the "two ground connections" part I figured out, but their
conclusion was the same as mine, dielectric unions don't help, but the
newer dielectric insulated nipples do, because the galvanic current has
a longer water path to traverse which has a higher electrical
resistance, significantly reducing that current's density.

It was a learning experience for me. I yanked out the dielectric unions
and the steel nipples and replaced them with copper unions and copper
stubs screwed right into the tank. Two lears later, and no leaks yet.

Jeff

--
Jeffry Wisnia

(W1BSV + Brass Rat '57 EE)

"Truth exists; only falsehood has to be invented."