Why are gas furnaces designed to retry if ignition fails?

csigona

My pal has a Utica Boiler UB90-100. I suspect the make and model doesn't matter since my question is generic. In any case, he asked me (a non-professional, all-around fixer guy) to come over to take a look. He describes his problem as the gas igniting, then immediately shutting down, and repeating this a few times before taking a longer break. After the break, it goes through another set of retries, and is eventually successful. I suspect the igniter/sensor or poor ground, but that's not my question.

I'm wondering why these systems are designed to retry in the first place. It suggests to me that the process itself is prone to failure and the solution is to just repeat until it works. I understand why the system would shut down if it doesn't detect a flame, but why, at that point, doesn't the system just stay off until the service guy comes? After all, something's wrong, isn't it? And retrying is, in a way, just ignoring the failure and praying that this time will be different. But whatever tripped it the first time remains a problem, one we've masked by retrying, until it grows steadily worse and in the end fails despite a dozen retries.

GGross

Let's say the gas supply is interrupted for whatever reason but only briefly, without auto retry that would be a whole lot of nuisance service calls

I might add that there are other things that might be only a temporary condition that could cause the flame to be too weak for the control to recognize there is flame as well. I would tend to agree that many controls are a bit too lenient on retries though

Derheatmeister

So that the home does not freeze…

Have not seen this yet so i am calling dibs on a Ignition devise that is hooked up to your LAN and will send out notifications should there be a issue. 3/21/2025

mattmia2

https://forum.heatinghelp.com/discussion/comment/1849252#Comment_1849252

and most of the times that issue will come over the network…

GGross

That feature is built in to many boilers these days. remote monitoring, some have full remote control, all I have seen will alert when there is a fault, some apps allow notifications for faults on your phone as in not just a text but a straight notification through your phone. And yeah usually the most trouble I have with them is getting and keeping them connected. It would be nice to get something more universal because right now every company that uses this tech pays someone to keep their apps updated to varying degrees of effectiveness

Derheatmeister

https://forum.heatinghelp.com/discussion/comment/1849256#Comment_1849256

Full Remote control !

Just like the 50 "Dis"ConXus that we installed and have had many issues with for Years !!!!

Full Remote control !…One of the additional benefits is that it actually will turn the heat to the house »»OFF «« while trying to adjust the "Ramp delay" and the "Mod factor" remotely !

On two unrelated sites in Blue River Colorado i personally had to drive a half hour to the customers home and physically turn the boiler back on othervise the homes would have Frozen !…Your Welcome Lochinvar !

Yet Lochinvar cannot fix this issue or is being held hostage by a programmer that does not even understand the first thing about Hydronics maybe He/She/They/Them does not even care !

To top it off: even though we were very polite and hope that the conversations are recorded for training purposes.

On the Last call to Lochinvar the technician was giving us ill advise,told us that we are "Barking up the wrong Tree", Stated that Lochinvar does not manufacture the ConXus and is therefore not responsible for it not working as intended,Told us that this conversation is not going to go anywere and disconnected himself from the request for help…

Nice Customer Service LOCHINVAR !

mattmia2

I'm being pressured by management o remotely monitor av systems. the stuff that goes wrong is very difficult to monitor and thus far the monitoring has broken things. frequently. if you put the amount of effort that it would take to monitor the thing in to fixing the thing so it doesn't happen, no need to monitor it.

mattmia2

But back to the original question. When Chrysler first started making cars with microprocessor controlled combustion, they actually worked reasonably well. but they didn't think about what to do if part of the system broke, they didn't design it to still run with some default settings if it wasn't getting correct readings so the car would break down if some part of the system failed. This gave customers the perception that the system was unreliable. Add to that that most mechanics didn't understand how to work on it and it was seen as a failure and most cars got a conventional carb installed. They eventually learned to design the systems to run in a default fallback mode .

I think that is what is happening here. There are lots of transient reasons an ignition attempt may fail, brief power interruption at just the wrong time, air in the gas from some utility work, wind blowing just the wrong way, or parts that are wearing and sometimes failing to ignite. The preference would still be to light with the failing parts if it is possible, gives you more of a change to notice it, especially if it is smart enough to store an error.

109A_5

So you would call the tow truck if your car did not start on the first try ?

My old boiler with a standing pilot has started on the first try for maybe decades, can't remember the last time the thermocouple was changed. Apparently the newer technology is not as good (at least for starting and staying lit) since it may need a retry.

csigona

Thank you all for your responses. I appreciate the time you took to answer. I have, however, a reply to Mr. 109A_5.

I believe your comparison with a car that fails to start, in a way, shows my point. That is, a guy starting his car would himself notice that his car is getting harder to start, and as it gets harder and harder, eventually brings it to a shop and gets it fixed. The furnace, on the other hand, keeps this knowledge secret from the owner. I guess that's no longer the case for modern furnaces that report failures to your phone or computer, but it is the case for the furnace I'm looking at. It's quite possible it's been failing and retrying all winter long with the owner unaware of the problem. This unit has no error codes, or none that I can find. In any case, thank you, Mr. 109A_5.

Just to let you guys know, I visited the unit yesterday. The owner describes the failure as flame-on followed immediately by flame-off as if it were blown out. I know him, and know that his descriptions are often unreliable, so I wanted to see it myself. It behaved nicely the whole time. I ran some experiments and turned it off and on about 10 times. Too bad I couldn't see it fail to ignite. I did manage to rule out a faulty igniter/sensor and also ruled out a bad ground. I'll have to think about it some more. But given enough retries, it works again.

mattmia2

Sounds like it is either failing to prove flame initially or is losing proof of flame once the main valve opens depending on how this is controlled. If it is proving flame with an intermittent pilot and the pressure is dropping when the main valve opens then the pilot could no longer be covering whatever the sensor is in this design. The draft from the main burner could be moving it off the sensor too. If it is direct ignition it could be that it is intermittently failing to prove ignition altogether.

oh, i forgot transient drops in gas pressure causing the pilot flame to not cover the sensor in my reasons of why you'd want to try again. if it isn't doing it while you are looking at it, i don't see how locking out is going to help anyone.

usually sensors on a car either work or they don't so the engines on early designs would just die when that happened.

HVACNUT

@csigona. Gas systems that repeat the cycle (usually 3 to 5 times before lockout) are draft induced. Every safety in the line and low voltage circuits are operating properly to even allow 24 volts to the gas valve. The Igniter doesn't know if there's gas or not. It just knows it's circuit is complete. If there is gas to the manifold, but no ignition, the inducer fan then clears the HX of unspent gas and pushes it out. On the down cycle AND when starting a new sequence. There's really no danger in repeated (but not unlimited) trials with today's equipment.

Oil, of course is different. One trial for ignition. If no flame signal within 5,15,45 seconds, depending on the safety control, the burner locks out. Obviously fuel oil is not a gas, so if fuel is entering the combustion chamber without ignition, it's not going anywhere until it's burned out. Carlin used to have an oil primary that would repeat once after a failed TFI. They're gone for good reason.

109A_5

You did not say which Series the boiler is. The Series 4 has a whole troubleshooting flow chart that deals with the type of complaint your pal is describing.

Jamie Hall

Designing a control system which will fail operational is not easy. Nor, usually, is it inexpensive. A control system which is fail safe is much simpler — and what we usually have for both oil and gas burners. As has been noted, multiple tries for ignition on a gas burner, provided draught is available, is actually not all that dangerous (it's not great, mind, but not that bad).

A comment on the car comparison though. Yes the early Chrysler electronic systems weren't all that great; much too much faith was placed in the various black boxes working as they should. The fact of the matter, though, is that while modern black boxes and the associated wiring is better, it's not that great either — but what has happened is that there is almost always built in a fail operational mode of control. This is both good and bad. Fail operational means the thing will run, but in a degraded state. Most motorists won't notice, to be honest. How many cars are driving around with the MIL light (also known as the check engine light — or in the mechanics trade the money light)?