12 volt power supply is perfect, don't go any higher.
Before you put in another MOSFET...
Are the RGB LEDs still lighting up? Are the separator LEDs blinking? They should be because the MOSFET has nothing to do with the LEDs. What I am thinking is you could try running the clock with no MOSFET for a few days and use the LEDs to figure out if the controller is stopping. If the controller is stopping you need to find out why or else you will just burn another MOSFET.
Look into it later when the dust is clearing off the crater.
There are a couple of other things that might have happened:
So, an explanation first: The MOSFET shorts the supply for the period of PWMOn. It is only the inductor that limits the current into the MOSFET, and the limiting effect decays very quickly. For the first microseconds of PWMOn, the inductance resists the current flowing, and only a small amount is passed. Every microsecond, the current can increase as the magnetic field in the inductor is created. At a certain point the inductor has created the magnetic field and it does not resist the current any more, and at this point, both the inductor and the MOSFET must pass the whole current of the power supply.
The HV Calibration tried to find the right value for the PWMOn, by locating the shortest PWMOn point at which the high voltage is produced correctly. If PWMOn is too long, the limiting effect is only there for a part of the PWMOn time. This makes the inductor and the MOSFET get hot.
To reach the high voltage, it is important that the inductor turns off cleanly, and skips the "linear region" where the MOSFET acts as a resistor and thus creates heat. When it is completely off, it creates no heat, and when it is completely on it creates only a little heat. The on resistance of the IRF840 is 0.82ohms, and so as long as it turns fully on, it will on average create only a fraction of a Watt. (The power while on is small, and the on time is small).
There are several possibilities:
The PWMOn is wrong. This can happen because either you changed the power supply and did not recalibrate, or that you manually adjusted the PWMOn thinking that it is going to affect brightness. (Not blaming you, just trying to locate the problem - either way I will send you replacement parts or a replacement kit).
The inductor is faulty: Perhaps it has a much too low value, or overheating has caused some internal damage. If it overheats too badly, it is possible that internal turns have shorted together.
The MOSFET is not turning off cleanly. This can be caused by flux residues or board damage. Burnt board can show a resistance. There is also a resistor between GND and the gate, and the idea of this is to hold the gate low when the controller is not there to turn the gate on, otherwise noise on the gate could cause the MOSFET to turn on randomly,
I think the best I can offer is to repair or replace the board. If you want to try to fix it, I can send you a new inductor and MOSFET.
Ian wrote: Now the preferred is an IRF640. The very first ones were IRF740, and for historical reasons and safety reasons we used in the recent past IRF840.
After reading this, I decided to change original IRF840 in my clock for IRF640 to lower heat. I did reset and calibration and clock was runnin fine. Mosfet was much cooler than before. But after maybe one hour I noticed clock is not working. I touched the mosfet and it was super hot. I tought new mosfet died, so I changed it for IRF740, but clock was still not working. HV voltage was only 8V. I measured all components in boost circuit and ewerything was fine except inductor L1 - it was shorted. Mosfet is also fine. I salvaged similar inductor from some old PCB and also soldered original mosfet. Now clock is running again.
Any ideas why clock works fine for two years with IRF840 and died in one hour with IRF640?
Classic rev5, FW57. New mosfet bought at TME.EU.