On Sun, 11 Jan 2004 02:12:43 GMT, Gunner
wrote:

Ive got a problem with a water heater at work. This is a natural gas
unit used to boil water out of oil for reusing the oil

Controller is a Honeywell S89F controller. Everything checks out
normally..sorta

Unit is powered up, gas on.
35 second prepurge cycle starts

At this point, for 4 seconds the ignitor fires up, the gas solenoid is
supposed to open, and then run for a few seconds to allow the flame
sensor to stablizes (DC rectification through the flame) then run. If
no flame, its supposed to recycle another 3 times then lock out if no
burn is detected.

My problem

When the solenoid voltage for the gas valve comes on..it makes and
breaks 3-4 times, or simply comes on then quits within a second. If
the flame actually does fireup, it may trigger the solenoid on and off
click click click for the 3-4 seconds, then shuts down sometimes it
only clicks once. It may or may not go through the following 3
recycles.

Ive checked the ground, cleaned all the grounding connections, double
checked the ignitor via jumper and still I cant figure out whats
causing it to not come on and stay on. Input voltage is constant, the
24vts to the controller is constant, etc etc.

I put a DC milliamp meter in line with the flame sensor rod, but was
unable to get any reading in the second or two that the flame actually
did come on.

Do I have a bad flame sensor or a bad controller? The flame sensor
appears to be in the proper position from the burner face, and the
flame sensor lug does appear to have 110vts on it, measured to ground.

Any ideas? My gut instinct is a flaky controller, but the entire
blower unit was removed and taken to the dealer where it did work
properly, its alleged by the staff.

This is the manual for the unit
http://hbctechlit.honeywell.com/tech...0s/68-0066.pdf

Gunner


Well, you've gotten some good advice so far. Gunner.

I read thru the responses to your posts. Seems as if folks have
covered all the probable things.

So I doubt I can offer much of value.

However, a couple points stick out in my mind and cause me to go
Hmmmm.

You say entire blower unit was removed and sent off for a bench test
and it supposedly worked there, on their test bench.
AND
You say it doesn't work on site.
AND
You describe the sequence of operation of the controller as being
somewhat flakey on site. Sometimes multiple cycling of the solenoid,
sometimes not. Sometimes does the 3 retries, sometimes not.

The above bothers me. Makes me sorta curious.

The thing about those controllers is that internally, they have
digital circuitry. All digital circuits absolutely HATE differences
in ground potential, any differences, between this point and that.
Some are a little more tolerant than others, some less.

In my experience, that particular series of Honeywell controllers is
one of the top haters of any differences in ground potential anywhere
within it's internals, or anything connected to it.

Differences in ground potential which are minute, which would make no
difference in electrical or electronic controls systems, can confuse
and confound that line of Honeywell controllers.

So I wonder to self.

Hmmm. What is different between what you're calling the "blower
assembly", which worked on a test bench. And what you're working with
when you have everything put together on site, and now things aren't
working.

Did that "blower assembly" (I have no clear picture of exactly,
precisely all the components you mean when you say that), as tested on
the test bench include the very same transformer? The same solenoid
valve? The same gas valve? Etc.

Do you see what I'm getting at? What's different that's connected to
that controller now, from what was connected to it when it was bench
tested.

First thing which jumps into my mind is that often when somebody says
"blower assembly" ... that doesn't include the burner itself. May, or
may not. ---- ALL---- grounds that system, anywhere in it, is supposed
to have, MUST be solidly connected to that burner. You don't want
even 2 ohms of difference between this leg of ground and that. Check
the schematic. See all the points to be ground. Make absolutely sure
that when testing from this point in ground to another, that your
meter is as flatlined as possible. Personally, I'd make sure I wasn't
measuring anything over one ohm anywhere. Less if I can make it
happen.

Next, I'd check connection from burner to cabinet. Now, sometimes
folks rely on simple metal to metal contact between the pipes and
other metal parts to establish that electrical connection to ground.
And there is no separate ground wire bolted to burner, then bolted to
the cabinet. This can work fine on a brand new unit.

Can suck, and drive yah nuts trying to figure out the problem on an
older unit. With age, joints and connections corrode. Resistance
increases. Or ground path included a section of pipe, and somebody
took it apart, and upon reassembly used some sealant on threads. And
what used to be a good ground path is now not such a good one.

I'd make sure. I'd add a damned good ground wire from burner to
cabinet if there was not one. If there is, I'd investigate. Conductor
intact? Both connections good, and solid and clean? Etc.

I'd then double check. Power shut down. Do I measure zero ohms, or
damned close to it, from grounded leg of transformer to cabinet? From
gas valve grounded terminal to cabinet? Ground terminal(s) on
controller to cabinet? Ground of ignitor to cabinet? Etc.

That everything looks okay, doesn't hack it. Measure, verify. No
more than 1 ohm, preferrably less. And I'd personally go for less.

After verifying that. My next step would be to verify, beyond any
doubt, that the cabinet is grounded to EARTH ground.

One problem when dealing with digital circuits in particular arises
here. In many installations, particularly older installations,
cabinet ground was made by the electrician with him relying on a
ground path running thru the conduit (if there is some) itself. In
other words, he used the metal pipe as a ground path. This kinda,
sorta works. Most times, most applications, works fine.

At first.

Problem is, the passage of time. All those conduit fittings that used
to be tight, clean, nice metal to metal contact, after the passage of
time often become not so tight, not so clean, etc. And if cabinet
itself was grounded to earth ground via a path thru the conduit
material itself, over the passge of years you might find there is
significant resistance now built up (and thus potential, voltage,
differences) between the cabinet and the actual earth ground of the
building. As measured at the electrical service entry.

Not a good thing. Couple volts aren't gonna hurt or even be
noticeable to regular electrical devices. Some types of digital
devices, and that Honeywell controller is one of em, aren't gonna like
it much.

BTW, do NOT consider a connection to AC common the same as a
connection to ground. Period. I don't care that back at the breaker
panel common may be hooked to the same ground bar as the ground wires.
Do NOT consider a common wire as ground. Ground is ground. Common in
normal operation carries current. Yah don't want the ground for that
controller hooked to anything carrying current. Will confuse the
devil out of it.

The above ground related things, I'd be inclined to check, and double
check first.

Then I'd check whatever else is hooked to that controller one way or
another, that did not accompany it to the shop for bench test. What
else, that wasn't attached to it during bench test, might be shorting
to ground, drawing too much current, etc.

Sometimes a transformer will seem fine. Measure output voltage, she
looks good. Hook her up to something that's drawing too much
amperage, i.e. a relay or solenoid which does a big spike draw upon
energizing, voltage droops. Sometimes very quickly then comes back to
normal because relay coil dropped out and field collapsed. And what
you hear is chatter. Often yah can't see that rapid fluctuation on
many meters. But yah can usually hear it. Problem might be relay or
solenoid coil is drawing more than it used to, particularly on
energizing and making something move (inrush current), or windings are
starting to fail and after a moment get too hot. Or sometimes it's
the transformer. Age. Might say 40 watts on the label, but she can't
handle half that any more. And particularly on inrush current draw,
she starts putting out erratic, drooping voltage. Digital circuitry
does not like this sort of thing.

Just some thoughts.

Best of luck.

Bob