Probe type LWCO failure

cn30

When I tested my probe lwco (McDonnell & Miller PS-802) it did not shut off the burners on low water as it should have. The red light just blinked a while and then stopped. It's supposed to blink for 10 seconds and then shut the burners off. I tested it twice with the same results.

Anyone ever see this before? The boiler is a little over a year old. In the manual it also says "Conductive liquid required" and "water/glycol mixtures up to 50% concentration may be used." I'm not aware of glycol ever being added to the boiler water (I certainly haven't) and 50% sounds huge. Also, I would think--but don't know--that lack of conductivity would prevent the boiler from ever running rather than the opposite.

Does anyone know how these things work? I have a service tech coming tomorrow to clean the probe, which the manual says may also be an issue.

ChrisJ

Just a homeowner

But, if it was mine I would drain the boiler enough to remove the probe and then remove it and inspect / clean it.   I'm guessing your boiler needs a skim bad and the oil is also coating the probe causing it to sense liquid even during a low water situation.



As far as I know antifreeze is only used in hot water systems, not steam.

cn30

clean probe

Yes, I'm definitely going to skim after having the probe cleaned. The water in the sight glass isn't bouncing more than usual (plus or minus a half-inch, with occasional 1 inch dips or rises), but the glass is very wet above the water level. I'm going to let the service guy remove and clean the probe since the whole system's still under warranty.

Hap_Hazzard

Did you use PTFE thread tape on the probe threads?

The probe relies on the threaded joint for its ground connection. PTFE tape can insulate the joint. The LWCO might be seeing some conduction due to capacitance and assuming it's low because of the water chemistry.

ChrisJ

PTFE

Hap, I actually did use tape on mine but it works fine.

You know it kind of annoys me, in the electronics world we would always have a ground wire on something like this.  Seems really dumb to rely on a connection where you have to use some kind of thread sealant.



I also assumed though that a poor connection would create a false low water alarm though rather than the opposite.



At least I think I used tape....  Now I can't remember but I do remember being aware of the concern at the time.      I either used tape + dope or just dope, I want to say I used both though.

Hap_Hazzard

You should change that next time you clean your probe.

Sometimes the threads will break through, but it's not reliable. I use a thread compound that contains PTFE on mine, so it's easy to remove but makes good contact.



Re. ground wire, where would they connect it on the boiler?

ChrisJ

ground wire

They would need a simple screw connection threaded into a thick spot on the block. Same as the tappings except obviously much smaller and not all the way through.



Or, like on my EG-45 the base of the block has a few flanges, perfectly fine to clamp a ground to.

cn30

Update

The service guy came this morning and upon seeing the probe, said the problem was due to the use of pipe dope on the probe threads (dope was visible). He took it out, cleaned the threads, put it back in, and it then worked fine. The probe part itself has a loose piece of plastic on the end that bounces around with boiling and prevents any kind of build-up, and that part in fact was very clean.



The manual says "Sparingly, apply pipe sealant to the external threads of the probe." It also says not to use Teflon tape. The installer (not this guy) did in fact not use much pipe dope. Anyway, the service guy that came today said absolutely don't use any. I'm going to test it again today, then again tomorrow, then weekly, then monthly if all goes well.

ChrisJ

dope

He said not to use any pipe dope? How does he expect it to seal in the block?

cn30

sealing

My feeling is that npt threads make tight seals by design, but anyway, he said that brass (the probe) is soft and thus when torqued makes a tight seal.

 I'm also wondering if the pipe dope that squirts back when you thread it tightly made a barrier between the probe nut and the boiler block. The threads are of course integral with the nut, but the ground wire is attached to the nut, so maybe there's a better connection if that nut meets the boiler? No idea, just speculating--don't even remember if the nut gets that close to the boiler.

BobC

Test it

i can't see how pipe dope could block a contact on something that is torqed down, just use an ohm meter to make sure the body of the probe is connected to the block and the LWC) ground inside the box.



How often have you taken something apart and not found the compound or tape shredded on the interior threads?



Bob

Hap_Hazzard

It doesn't.

NPT threads do not make a fluid-tight seal without a thread sealant of some kind because the crests don't fit tightly against the roots, leaving a helical leakage path. NPTF, a variant used mainly in automotive fuel and hydraulic applications, is designed to fit more tightly, so if you are more familiar with those applications, I can see how you might have gotten that impression, but NPTF and NPT are not the same threadform.

cn30

Ok

I'm not familiar with those other applications, but I thought I had read that NPT threads were cut so that when tightened, the crests took up 100% of the depth of the valleys--but apparently not.



My main question now, though, is how a weak connection, whether through dope on the threads or maybe a loose ground screw, could have caused the failure. A weak connection should have caused the boiler to intermittently quit or not fire, no? Like a loose car battery terminal?

Hap_Hazzard

Weak connection

Actually a battery terminal can do very similar things. When there's a little corrosion you can get a connection where everything electrical seems to work but the car won't start because it can't carry enough current to turn over the engine.



There are good conductors, poor conductors and insulators. Some materials are such poor conductors that they behave like insulators until they become thin enough to allow current to pass through. PTFE tape is such a good insulator it can be stretched or compressed into an extremely thin sheet and still insulate, but it becomes polarized when current is applied--all the electrons jump to one side of the film--causing a spurt of current. If the charge is removed they jump back to their normal positions, causing a current in the opposite direction. If an alternating current is applied, this effect allows the insulator to behave like a conductor, but the amount of current depends on the frequency.



So, basically, it's not an all-or-nothing proposition, and the reason the connection wasn't intermittent was that the connection was tight mechanically, but the viscosity of the pipe joint compound didn't allow the two metals to make contact.



If the compound is applied sparingly, as the instructions say, there should be enough to fill significant portions of the helical leak path, making the joint fluid-tight, while still allowing the flanks of the threads to be in contact.

cn30

Second tests

Yesterday I tested the lwco probe 3 times. First time, the red low-water light blinked appropriately but the burners never cut out, even though the water had just disappeared in the sight glass. (This is 4-5 inches below the probe.) Second time, the light blinked for 2-3 minutes as the water level dropped and the burners cut out with about 1/2 inch of water in the sight glass. There's supposed to be just a 10 second delay on break. The third time, the light blinked briefly and then the burners cut out right away, with 2-3" of water in the sight glass.



I'm not happy with this. First, I'm thinking the circuitry needs exercising, which seems rediculous. Second, I realized that the water was progressively cooler at each test (with some make-up water by the 3rd test if not the 2nd), and I'm wondering if that's why it worked better. Finally I'm thinking maybe I dropped the water level too quickly at first, and that it would work better with a very slow drop.



On the plus side, I turned the power off after each test, waited a bit, and turned it back on. All 3 times, the device wouldn't let the boiler start in a low-water condition, which is exactly what it's supposed to do (one of its functions, anyway). It also let the boiler start when I reconnected the pump and the water level rose. Now I'm thinking it may be a switching or electrical issue. Does anyone know where the switching function takes place, and why the probe might work perfectly (as far as I have tested) starting from low water and going up, but inconsistently going down?

Hap_Hazzard

The only thing I can imagine

is that it might "think" it's immersed if it's still wet. Normally this wouldn't happen because it dries faster when it's hot.

ChrisJ

Foaming

Can cause a probe type LWCO to not shut burner off when water is below safe level.



Also I believe just because the gauge glass claims the level is too low, doesn't mean thats the real level inside the boiler while steaming.



keep in mind these videos below are about commercial size boilers, but it does give you some ideas of how things may not be what you think.



Part 1 : http://youtu.be/a0xZPl_bwHI

Part 2 : http://youtu.be/tI3tWuSsX7c





If I was you, I would probably check the LWCO wiring against a wiring schematic from your boiler manufacture, Chances are the installation manual includes one and is available online. Once I confirmed it was wired appropriately I would replace the probe. If that still fails then replace the entire LWCO.

Bob_Engelhardt

LWCO failure modes

There's been some back-and-forth on this, so here's a summary: the probe works by conducting when the water is up and not conducting when it's low.  Therefore pipe dope could not give a false high, only a false low.  I.e., by insulating the threads it could prevent conduction and make the control think the water was low even when it wasn't.  It could not create conduction when water is low, giving a false high, which is what the OP is experiencing.



A false high occurs when the probe is conducting without being submerged.  When there is something "stuck" in the probe that conducts in the place of water.  Foam, maybe - who knows.



The fact that the OP's problem was "fixed" by cleaning off the dope was a coincidence.  Handling the probe probably dislodged whatever was shorting it.  Maybe only somewhat, so the problem has come back, sort of.  If 'twere mine I would pull the probe and thoroughly clean it.  Are they replaceable, without replacing the whole unit?



BTW - if the LWCO fails high (it thinks the water's high when it isn't), is there a backup safety, or can the boiler run dry?



Bob

JStar

Lwco

Has anybody checked if this was wired correctly?

cn30

wiring

JStar, could it still be wired wrong if it works some of the time? I would think it wouldn't work at all, but I'll have it looked at. The thing is, it is functioning correctly when the burners are off and the water is low--it won't let the boiler fire even though the thermostat is calling for heat. And, sometimes it functions as the water drops.



I'm thinking it must be the control board, because if the red light is blinking, the sensor has done its job--it's signaling a low-water condition. But, the control board either won't signal for a cut-off or takes forever to do so. I saw the light blink for 2 minutes straight the other day. It's supposed to act within 10 seconds. I'm trying to get a replacement under warranty for both the sensor and the control.



What's interesting to me is that the manual insists that the boiler absolutely will not start at low water, but is not as forceful in saying it will cut out as the water drops. I wonder if they're banking on catching a problem between cycles. But, I have a condensate tank and a boiler feed pump, and my boiler gets every drop of its water from that pump.

JStar

Lwco

You may be on the right track. I would just verify it by doing a few more electrical tests. I see some very dirty LWCO's that still work.

cn30

Agreed

I think I'll have my electrician check it out--I don't have the equipment or the know-how. A couple more questions: does all the switching take place at the gas valve, or somewhere else? I take it the thermostat, the pressuretrol, and the lwco all send signals to the same place? Everything is wired one way or another into a 24 volt transformer, which has a wire to the gas valve.



And following up on Bob's point above, is it possible to have both the probe and the float chamber hooked up as separate and redundant lwco's? My float chamber runs the pump and also would be the lwco but was not hooked up due to the boiler coming with the probe.

JStar

Advice

Most electricians are terrible with HVAC controls. Find a boiler tech.



You can always add a second LWCO for extra peace of mind.

ChrisJ

LWCO

Brand new probe type LWCOs.

http://www.pexsupply.com/Hydrolevel



I personally prefer Hydrolevel, though others like Mcdonnell & Miller which is also avaiable at pexsupply.com.