The fan will sometimes spin up as normal and run for a while, then simply stop for no apparent reason at no defined time limit. Sometimes the fan will fail to spin up at all. This fault is totally unrelated to 'the fan stays on whilst in standby' issue, There are several reasons why that 'fault' happens and I will write a seperate piece about that another time.
First, confirm that the fan is ok. Its exactly the same fan connector as used in most desktop computers, so really easy to either plug an old PC fan into the PVR-9200T, or plug the Humax fan into a PC to test if its working correctly.
The Humax uses a standard 12 volt 60cm fan that is very easy to get from many online sellers. A noisy fan should always be replaced even if it appears to be spinning correctly. In the 9200T the 12v fan is run at around 3.8v to keep it quiet.
I could not find any schematics for the Humax PVR9200T, so I reverse engineered the fan circuit:
Transistors Q586 and Q585, marked 2X (MMBT4401), are both found to occasionally break down under load. They are general purpous NPN's, and can be replaced with pretty much any other general purpous NPN transistors. I used 2N2222 as that's what I had in my box, and the fan works great.
Before replacing the transistors, you should confirm that the power MOSFET U5 (IRF7303 HexFET) is working correctly, as if the MOSFET is bad it can give similar symptoms. To test the MOSFET, connect the positive probe of your volt meter to +12v (yellow wire on HDD power connecter is easiest) then probe pin 3 of U5 with the negative probe. When the humax is powered on you should see about 9.5v, and in standby (wait until the time comes up) this voltage should drop off to 0v. If you are getting around 9.5v when the fan is stalled, it confirms that U5 is not the cause of the fault.
The Q585 and Q586 are located right by the fan connector, so can be easily found. Be careful when removing the old ones, as it is easy to destroy the pads or tracks. If you choose to use through hole parts instead of SMD as replacements, consider using something to help support them. (Ok, I admit to using 'heat glue' in my unit - yikes! lol)
I have added some of my notes here to help diagnose your own unit:
(all voltages taken with positive probe at +12v, probing with negative probe)
Fan working correctly = 3.6v, Fan Stalled = 1.7v
(3.9v with new transistors fitted)
Voltage at r599 =10.3v in both states
Voltage at junction of r599 and r607 = 10.3v in both states
Probing junction of r607 and q586 (10.3v -fan working) with anything metalic makes the fan start working. it will continue to work for a random amount of time after probe is removed.
Voltage at r603 q585 = 11.5v fan working, 12.6v fan stalled.