Samsung NQ50H5537KB microwave failure and a simple Arduino PWM fan controller.

First off, read the red disclaimer to the right. I'm about to go inside a microwave oven. Don't do it.

Saturday afternoon.

There's a rather unusual heatwave in the UK. The weather's great. My wife is cooking a jacket potato for lunch in our combination microwave. I'm sat in the lounge, watching a bit of telly..

I notice a loud 50 Hz hum coming from the kitchen. I go to investigate.

It's coming from the microwave oven. Something's heavily under load, and there's a smell of warm transformer.

It stops, and appears to carry on cooking. I go and sit back in front of the telly.

So, later on the wife has gone to work, and I go to make some beans on toast (living the dream, eh?)

Beans in the microwave... and they come out cold :(

Now every microwave oven has a label on it... "Microwave energy - do not remove cover" ... so don't remove the cover. This means you.

I remove the cover.

I check to ensure there's no remaining charge in the high voltage capacitor, and proceed to check the high voltage fuse. It's open circuit.

So, remembering the smell of warm transformer earlier, and the sound of a transformer under heavy load, I check the rectifier diode....

It measures dead short circuit. Now with these high voltage diodes, you can't check them with the normal diode setting on your meter, as it requires a few more volts to start conducting. Normally a silicon junction will forward bias at about 0.7V or so, but these high voltage types are actually a stack of diodes, so the 2-3 volts your meter provides will not be enough. You can use a bench supply to check them, but a short circuit is a short circuit, and this will show up on a normal test.


A new diode is ordered and fitted and normal operation is restored. Thankfully the transformer appears to have survived it's ordeal.





It does, however, bother me that it's failed. The unit is only a few months outside of it's guarantee, and it replaced a previous oven that was also short lived.. I wonder if it can't get rid of the heat generated quick enough. You can see the big "cage" fan in the unit, and there's also another fan that forces air through a duct across the fins of the magnetron.

The oven is a built-in type, and sits above a conventional oven, and there's a vent up the back of the cupboard. Perhaps it's not enough.

Let's add some forced air cooling...

I order 5 off 80mm fans from eBay, the sort you use to cool a PC, and a small switched mode supply, capable of 12V at 2A.














The fans are mounted on some aluminium angle, so they will sit over the vent at the top of the cupboard, sucking the warm air out.

Now, we don't want the fans on all the time, and we want to keep them as quiet as possible, so a circuit is designed, based somewhere between the Simple Thermostat project and the fan controller on the Arduino Audio Wattmeter

I don't need a display, or adjustable set-point, so this simplifies the design, and the required code.

Here's the schematic...

Power comes in to JP1, and is regulated by IC1 to provide 5v for the ATMEGA328. A DSB18B20 sensor connects to JP2. The micro determines the required fan speed, and outputs a PWM signal on pin5, which is used to control the fan via the STP55N60L logic-level MOSFET. The FET should run cool, no heatsinking is required.

A board is designed...

Modelled... 

.. and turned into reality.

The code, and eagle files, can be found on my GitHub page here.

The fans arrive, and I set about making a frame out of 10mm aluminium angle...

I've found an excellent eBay seller of metal profiles like this... they're called mwprofiles (click here to link to their eBay page)

And it's all set up and tested. Power is supplied by a plug-in 12V 2A supply.

Sadly, I can't get a shot of it installed, as there's not enough room to get the camera in!