Strange mobo behaviour - POST or not!

[trentham]

I've been looking at an American Megatrends M758LMRE mobo, equipped with a Celeron 1100MHz CPU and 256MB of SDRAM.

It was replaced because it just wouldn't behave and I've discovered that if I reset the CMOS by shorting out the appropriate pins it will go through the POST and let me into the BIOS to set it. Once it's been set (usually to the optimum defaults) it will no longer show the POST and just won't boot up until I reset the CMOS again.

Sometimes it won't even power up at all - even with a new PSU attached - and I'll go away and leave it (for a few days usually - I have other things to do!) and then it powers up again and shows the first described behavior.

Clearly there is something wrong with it but I've never come across behavior like this before. Has anyone else? Has anyone any idea what sort of fault might cause this?

 

[Nodsu]

The CMOS battery is dead and makes the CMOS memory go bananas and corrupt itself?

Did you try the failsafe defaults?

 

[AlbertLionheart]

The board is fluster-clucked so don't waste time on it. I have had this before and it had been caused by a power surge from a dying PSU. The reason we thought at the time was that EPROM was so badly damaged that it would not hold the settings.

 

[trentham]

To Nodsu: I've tried replacing the battery with a new one and used failsafe defaults but it makes no difference.

To: Albert...: I fully expect the mobo is of no use but I'm just intrigued at this behavior. It is almost certain that it did happen because of a PSU problem (that was replaced in the original machine). The idea that the EPROM is cream crackered has some merit. My only other thought was maybe a microscopic split in a circuit tract on the mobo giving occasional disconnections but I think the problem is too consistent. As for wasting time on it; I'm not. I'm not trying to fix it (it's already been replaced) but I'm just interested to see if this problem has shown itself much before.

 

[Tedster]

on such an old board it is possible the capacitors are blown or leaking.

 

[trentham]

Although I've inspected the board with a strong light and magnifying glass, it is still possible that a capacitor is faulty. I'd suspect leaking rather than blown; blown tends to be fairly binary... it either works or it doesn't and also there is often visible evidence.

Capacitors are certainly the most likely individual components to have failed, IME.

 

[AlbertLionheart]

Blown caps (bulging tops and often discoloured as well) could well explain the behaviour, especially if by leaving it for a while this changes the response.