Microwave thermal switch trips after several minutes on full power

[sanchaz12]

My microwave seems to overheat and trips the (bimetallic) thermal switch mounted on top of the magnetron lately. It heats for about 5-6 minutes on full power and then I hear a faint click which is different from the normal relay click, and the magnetron/transformer stops buzzing and heating. Everything else continues to work fine (timer, fan, light, turntable rotates) but it stops heating for about 30-40 seconds before the magnetron kicks back in. The magnetron also does not turn back on when I stop the timer and restart it, it only starts heating again after the thermal switch has reset itself. On lower power levels the magnetron does not trip the thermal switch.

I already removed the thermal switch and did a crude test by taping a thermocouple to the metal body and heating it up with a hot air station, making sure to only heat the metal body and not directly on the thermocouple. The markings on the switch indicate 90°C, and I found the switch opens at around 92°C and resets at around 71°C. I don't think the thermal switch is bad given it opens at the correct temperature. There was no significant dust inside the microwave or on the vents and the fan works fine, so I think that is not the issue. I cleaned what was there for good measure, but it still overheats (see attached picture before cleaning).

At first I suspected that it might be the magnetron starting to fail, but since there is also a transformer, fuse and a capacitor which all need to work properly I wondered if one of these might be the reason why it overheats and not the magnetron itself.

Is there a way to measure each of these components to find out if one is misbehaving causing the magnetron to overheat, or is this indeed a sign of a failing magnetron? Like, can a faulty transformer/capacitor/diode cause the microwave to work fine initially but overheat after time, without blowing the fuse or other symptoms? I can find a replacement magnetron, diode and capacitor for a reasonable price, only the transformer isn't.

I know how to safely work on a microwave, I discharge the capacitor with an insulated screwdriver and even keep it shorted with a clip lead before touching anything.

[Gyro]

Three things I can think of, all Magnetron related (assuming all your other checks, eg. freely spinning fan)...

1. The Aluminium cooling fins are coming loose, and so not achieving good thermal transfer from the Copper Anode.

2. Fractured magnets resulting in inefficient Magnetron operation.

You can see an example of both, and the internal melt-down effects on the Magnetron in this post... https://www.eevblog.com/forum/projects/will-a-magnetron-draw-480-watts-without-radiating-energy/msg3210292/#msg3210292

3. Burnt antenna on the magnetron (often caused by arcing from running empty or with insufficient moisture load, foil etc.). This is the most hopeful cause. You will need to remove the Magnetron from the wavequide to check this (making sure that the mesh gasket is in good condition on reassembly). Here's an example...



You can find replacements on ebay [Edit: search 'magnetron cap'] - they have different shapes, square, triangle etc. holes on the end, I'm not sure if this is important but best to stick to the same. They are a push-fit replacement.


P.S. Before anyone starts again, consumer microwave Magnetrons do not contain Beryllium (that one has been done to death many times previously). Obviously don't crack the insulator though because it will trash the vacuum seal!

[shakalnokturn]

No beryllium oxides, still beware of the high voltage supply, it is frame referenced and deadly.
Always check the 2kV capacitor is discharged before touching, bleeder resistors can fail.

[amyk]

The magnetron is essentially a vacuum diode. If it's gassy, that can cause it to draw more current and overheat.

[sanchaz12]

Thanks for the suggestions, I disassembled the microwave again, took the magnetron out and did as many tests as I could this time.

The magnetron cap looks intact and does not show any signs of arcing or other damage (see pic 4 & 5), also not inside the waveguide or on the waveguide cover.
There is some discoloration on the inner part of the magnetron next to the cooling fins and some kind of pattern is visible on the magnets (see pic 1 & 2).
There is no short/continuity between the microwave terminals and its case, and between the terminals it reads almost zero. I read that this is good.

I tested the diode with a 9V battery in series, where I measured a voltage drop of 4.13V (9.4V - 5.27V) one way and no continuity the other way. I mostly read to expect a 3V-ish drop, this is slightly higher but I think it's still good.

Testing the capacitor out of circuit with my multimeter, it reads 1.12µF and the capacitor should have a value of 1.10µF, so it seems to be within spec.

Also I did another temperature measurement, running it at full (900W) power with a bowl of water in it. I ran it for 3:30 minutes then quickly took the cover off to measure the temperature with an infrared thermometer. The magnetron was 72°C, the transformer was 31°C and the capacitor was 24°C. The magnetron is definitely getting pretty hot within just 3 and a half minute.

Power draw from the wall is 1530W, the sticker on the back cover states 1500W microwave input and 900W output, that seems fine as well.

I do not intend to measure voltages while the microwave is plugged in as I do not have the proper equipment to do so.

Is it my best bet to just replace the magnetron at this point?

[Gyro]

You do look to be running out of other options at this point.

Apart from your (very valid) safety concern about measuring internal voltages, there isn't much point anyway as a I think any fault (eg. loss of capacitance) would result in lower than normal anode voltage rather than higher. The filament is powered by a dedicated winding on the transformer and so fixed by design. Shorting of the HV rectifier or capacitor will have been ruled out by your dead continuity tests.

The Magnetron anode is always going to look a bit crusty as it is pure Copper and oxidises easily in that environment, loose contact with the cooling vanes can be detected with a thin blade (being careful not to actually cause the problem).

There is a Mica type window where the waveguide enters the main chamber. This can become dirty but there's a very low limit on the amount of power that it could absorb without disintegrating and, as you found, there is no sign of arcing. You can measure the actual power delivery by putting an exact volume of water in the microwave and measuring its temperature rise in a given time (this is normally included in the service manual), but if the Magnetron is overheating, there doesn't seem much point.

I think you are almost certainly looking at a tired Magnetron, where the magnet strength and / or internal anode structure are beginning to fail. Luckily they are relatively cheap these days.

[sanchaz12]

Thank you for the helpful information. After my last post I realized I forgot to put the waveguide cover back before heating the bowl of water and the magnetron heated up the same as before, that also can't be it then.

I do remember having used it to dry those silica car dehumidifiers about a year ago, I can imagine they absorb a decent amount of energy while full but as they dry there will gradually be less water which absorbs the heat. I dried them until they didn't feel humid anymore, so the microwave may have spent some time basically running empty as there was little to no water left in the dehumidifiers. That may also have not been so kind to the magnetron, and could very well be why it is acting up now.

Since the transformer does not sound strange and stays pretty cool, and the thermal switch, capacitor and diode check out ok I do still think the magnetron itself is the issue here.
A replacement magnetron is about €30-35, for me it's worth trying to replace it as the microwave is still in good condition otherwise and has been used 2-3 times a week for 7-8 years at most.