Do you happen to have any of those ATX2AT adapters? If you do, you could try that ATX PSU with the adapter in your motherboard's AT power socket, if only so you get some more info for troubleshooting purposes.

I can't remember which PSU socket on the motherboard I used when I tested my AT boards which have both AT and ATX socket.

Whats the model number of your ATX PSU? And do you have any other ATX PSUs you could try? It may have to do with the PSU being too new due to some ATX specification or something?

Try with another ATX PSU.

It is possible that the motherboard does not use 3.3V from the PSU, so it is unloaded, and the PSU does not like this and some of the voltages may be out of specification (too high or too low) and the voltage monitor chip inside the PSU does not signal POWER GOOD to motherboard so motherboard stays in reset.

Jepael wrote:

Try with another ATX PSU.

It is possible that the motherboard does not use 3.3V from the PSU, so it is unloaded, and the PSU does not like this and some of the voltages may be out of specification (too high or too low) and the voltage monitor chip inside the PSU does not signal POWER GOOD to motherboard so motherboard stays in reset.

Bing bing bing! We have a winner. The HDD LED was constantly glowing, and it does that normally when I hold down the reset button. Tried another PSU and it works perfectly. Thanks 😀

Hmm... PS_ON is being switched to ground by the chipset. Everything else should be common with the AT pins, so I feel POWER GOOD is being properly handled. Sounds like the motherboard is fine.

I'd anticipate at least a small 3.3V load, but a cross loading situation could certainly exist if the supply is 12V heavy and really, really garbage. I doubt crossloading is the issue. I'm thinking Jepael is correct about the supply being the problem, but I suspect malfunction. All rails should be lightly loaded with what he's running, and normal operation on a slightly decent ATX supply shouldn't be an issue. A really dodgy PSU might suffer from crossloading, but usually those tie POWER GOOD to the 5V switched output anyway. Supervisor chips cost money, after all. Can't let safety or standards get in the way of profits! A faulty PSU, on the other hand, might not be able to stabilize it's rails in an operable condition, will fail to send a signal on the POWER GOOD pin, and will cause the CPU to not post.

I'd recommend trying a decent ATX PSU in known good condition on the board. If the PSU is fully functional, I'm confident your board will boot.

Here's the pinout if you want to manually test the PSU:
http://pinouts.ru/Power/atxpower_pinout.shtml

You can latch it on by connecting PS_ON to ground and proceed to manually verify all the voltage rails as well as the POWER GOOD signal pin are correct with a multimeter. If everything looks good and you want to continue digging, find a point on the motherboard to monitor the POWER GOOD signal with the old PSU connected and watch it while booting to see how it works in a loaded condition.