Circuit board starts only when warm air passes over it.

[stj]

your seeing the other circuitry it's connected to.
my money is on a cracked solder joint that expands as it warms at this point

[ron350]

Why do those two pictures of the board look so strange.
No way to see cracked solder joints, cracked component or cracked copper trace.

[silentrunning]

Why do those two pictures of the board look so strange.
No way to see cracked solder joints, cracked component or cracked copper trace.

Struggling to get a decent picture. Is this any better?. I removed the thermistor and resoldered it. It was 14k - both ways(!) - whilst off the board. Still the same behaviour.

[CapLeaker]

Reminds me on a circuit board with an audio issue, No audio until I pressed a certain spot on the board. Whatever regular soldering I tried didn't fix it until I reflowed a section of the board. And to date I still don't know what exactly was wrong, but it still works.

I wonder if something is wrong with the PIC IC.
On the bottom right connector there are some crusties, The caps sleeve got a bit hot too, maybe worth taking it off and looking underneath.

[silentrunning]

Well I have had a lot of fun deciphering what the board does and learnt a shed load of stuff!

I was looking at buying a replacement thermistor but did a quick calculation with something I've never heard of called the Steinhart-hart equation (thermistor characteristic equation) plugging in details from a data sheet that seemed to match some of the resistance measurements I've been taking and figured I could replace it with a 5k resistor which would be the value of thermistor at about 20c when the unit should most definitely be running - and actually was if I gave the board the hair dryer treatment! I'm guessing the thermistor is there to stop the unit running at very low temperatures but either it or some of the surrounding circuit have become faulty. However, it seems that the processor is doing nothing fancy with the thermistor, just reading it's value so replacing it with an appropriate resistor will hopefully work, keeping it on at all temps which is what I want!

Having done that it's working but the house is much warmer and I'll need to wait until it's gets cold again to see if the weird behaviour returns. 🤞

[silentrunning]

NOPE!!!!! Spoke to soon. Same problem. So looks like something else is the issue!

[silentrunning]

I replaced the 3 big caps and put the thermistor all to no avail so the bin is beckoning for this board!

[CapLeaker]

Reflow it in the backing oven or hot air.

[silentrunning]

Reflow it in the backing oven or hot air.

That's beyond my pay grade! I'm just a beginner when it comes to this game. I've beginning to think it maybe a voltage regulator issue. I've identified this chip as a voltage regulator and the two resistors in the middle right as a voltage divider that sets the output voltage. I've noticed the output voltage is only about 0.9volts when the board is failing to start. Does the resistor with no markings look dodgy to anyone, I'm not familiar enough at the moment with failed components. It seems weird to have no markings. According to the data sheet of the chip it would need to be about 32k to give 3.3v output along with the marked 10k one although tracing it through it ultimately hits another 3 pin regulator that outputs 3.3v and feeds the MCU and other chips. When the board is running the output from the chip in the picture is the supply voltage of 24v which seems a bit strange, why use a chip like this if the supply voltage will do?

[silentrunning]

Here's another view

[CapLeaker]

I was thinking about the voltage regulator too. If the resistor is bad, it should start and then clunk out as it gets higher in resistance.

[harp]

So, why you not investigate in which section is problem, when you know that the slightly heat can trigger problem?

You can wrap a one piece of pcb into aluminium foil to slower rising temperature at this area, or cut a hole on it to expose only few components what you test it.

When device is in working state, you can use a ice cube in plastic bag and cooling down exact one by one component to bring some conclusion.
Also, test if the device is trigered by heating on back side of pcb.

https://m.youtube.com/watch?v=aPAeZWeJ-nw

What is resistance of this no-marking resistor?
Have you check for shorts in non-working condition, fuses are ok?
What is DC and AC voltage on electrolitic caps when device working/not working?
What is LED near top right corner status, is it ok?

[petehall347]

main things that start working with heat are capacitors and solder joints . ic chips mainly start with cooling . you could try giving the board a good clean i see corrosion and flux .

[silentrunning]

Thanks for the advice. The board did look a right mess under the microscope so I gave a good clean with IPA and a toothbrush and it's WORKING!!!!!!! Thanks Pete for the suggestion.

Blimey, what a journey! Harp - I did try everything you suggested apart from the ice cube trick - cost of living on the UK has gone through the roof so the refrigerator is a thing of the past for me! Nothing I did identified the temp sensitive component. Heating the back side of the board with the thermistor on it did nothing.

I'm guessing it was one or more of the components associated with the voltage regulator chip as that was failing to produce an output voltage above 0.9v unless the board was heated and some of the pins and joints around that area were a complete mess with crud and corrosion.

I've got to know that regulator chip well after reading it's datasheet numerous times. One thing I don't get about this board is that the feedback pin which should be set via the external resistive divider to the internal ref of 0.8v is reading 3.3v and the output of the regulator is about 22v ( input supply being 24v). This 22v goes to another small 3 pin regulator that brings it down to 3.3v which drives the MCU etc. Can anyone explain why you might go to 22v and then 3.3v rather than just have the main regulator output 3.3v?

Thanks for everyone's input on this. I think it's fixed, the house is freezing and it's turning on immediately and consistently.

[redwire]

It's fixed? It's winter and nobody needin' A/C unless south of the equator, like Australia.
Usually these control boards have higher DC voltage like 12 or 24VDC to power relays.
So if you see 22V that is the main DC rail, and then a regulator to drop that down to what the MCU and sensors need 3.3V or 5V.
MP4558 is the power supply buck-converter.

[silentrunning]

It's a positive air system not AC as such, it blows air into the house at very low adjustable speed and runs at about 16w. It is majorly effective at reducing (read completely eliminating) condensation in the colder months, something that plagues UK houses especially older ones like mine.

It's a 24v Power supply. The MP4558 takes that as input and pretty much puts 24v (about 22.8) out as output. it then goes through a tiny SMC voltage regulator that brings that down to 3.3v to power the PIC MCU etc. I don't know why the 24v supply doesn't go straight to the 3.3v regulator or have the output from the MP4558 go to 3.3v. The 22.8v output is not used by anything - the fan motor is directly connected to the 24v main input voltage with PWM from a pin on the PIC MCU controlling the speed.

There are a number of empty slots on the board so I figured maybe the 22.8v output is used for different variants of the board and is just left in place for them all for convenience and/or cost. Having said that the 3.3v going onto the FB pin of the MP4558 doesn't make any sense as far as I can tell. That seems like a mistake to me that somehow forces the chip to output what is essentially its input! To get an output of 3.3v the FB should have about 0.8v - the chips internal ref voltage - on it.