Arrived home in the early hours of this morning to a cold house. Boiler not working. Much as I would have liked to curl up in bed, I set about opening the boiler. It was stuck at the point in the ignition sequence where the fan is running, but no ignition spark. Looking back at my home automation logs, it had run for a short time when it was switched on remotely a few hours earlier. The logs showed the main burner cycling off and on repeatedly within a second. I have seen this before when the relay which feeds back the main burner state to my home automation suffers dirt on the contacts - it's a mains relay with coil in parallel with the main burner solenoid, but the contacts are not designed for ultra low power signal switching. They don't self-clean at such low current and need manual cleaning every few years.
Anyway, working through the ignition sequence, I find the air pressure switch is not detecting a pressure differential, even though the fan is running fine. I disconnect the tubes and operate it by mouth, and then the boiler ignites and fires up, until I release it from my mouth obviously. So the sensor works, but is maybe drifted out of spec. I open the boiler casing looking for leaks, but the seal is tight all the way around. Remove the fan and look into the air inlet chamber - all clear - can shine a torch out through the outer flue pipe and see the neighbour fence. Can't check the inner exhaust without going outside with a ladder, and it's pitch dark and pissing with rain, but that's probably OK too. I dive onto the web to find a next-day supplier for a new air pressure switch, but just to rule out anything else, I momentarily bypass it.
Boiler works, for a minute, but then the main burner starts cutting in and out. So this is what I saw on the home automation logs, and not just dirty relay contacts in my signalling relay. It cuts out for a short enough time that the main burner doesn't have time to go out, indeed flame barely noticably drops before it comes on again (initially I didn't see this and thought it was the pilot solenoid dropping out, until I put a test meter on it). This seemed like a component heating up with a bad joint.
Now I'm theorising two faults. As someone who teaches fault finding, I know this is incredibly unlikely and it almost always means the root cause has not been correctly identified. However, I took out the circuit board, and there was a power resistor there which had done the usual desolding itself over ~25 years. Proudly took it to the electronics workbench and resoldered it, confident it would fix the second problem.
It didn't. I had temporarily replaced the control board without its cover, and I now noticed something else - in the momentary solenoid dropouts, the ignition spark was triggered (flash visible from the gas breakdown device in the ignitor circuit.
So, was I now looking at a false flame failure detection followed by instant reignition (except burner never actually gets anywhere near going out)? Still worrying about the two separate faults - probably not actually identified the real root cause.
Gave up and went to bed (nearly 4am).
This morning, I tried it again and ran it long enough to get enough hot water for a shower with the air pressure sensor bypassed. An order for a replacement sensor was just waiting for me to hit return, but the two faults were still bugging me. What could link the two problems? What about reduced air flow through the boiler? Fan spun freely and was making same sound it has for last 15 years, so I didn't suspect that. However, lack of air flow might cause flames to not play on the flame failure (and ignition) electrode. The one thing I couldn't check last night was the exhaust flue because it was dark. Walked around to take a look - a bird's arse was sticking out, jammed in the grille. How it got in that far in I can't imagine, but on pulling it out and reconnecting the air pressure sensor, everything works properly.
So there was just one fault, when it was correctly root caused.