Dead XT motherboard from power surge.

[Flamin Joe]

Hi guys,

I had a friend drop off an XT motherboard which he said he killed by accidentally getting the XT power supply connector in the wrong direction (yeah I know hey). As a result it powers on OK but won't POST, no beeps nothing. He's given it to me to look at repairing if possible as I've got the tools to do so, desoldering gun etc, so replacing dead components is pretty straightforward for me.

I've looked over it carefully hoping to pinpoint any burnt components (he said he smelt a burning smell, not good!) but everything looks immaculate. I also can't smell anywhere it may have burnt to give me any clues as to where to start looking.

Basically I'm just wanting to know if anyone could point me in the right direction as to where to start looking for damaged components. Is there anything which might be more suspectable to overvoltage then other components? Or is this a lost cause? I'd really like to recover it if possible as I'd hate to see another XT bite the dust!

Thanks in advance.

 

[Stone]

I had a friend drop off an XT motherboard which he said he killed by accidentally getting the XT power supply connector in the wrong direction...

Does that mean reversing P8 and P9 or something else?

 

[Flamin Joe]

Does that mean reversing P8 and P9 or something else?

On this particular PSU the P8 and P9 aren't separated it's all one plug. So it was a complete reversal with the 12v rail going where the 5v is and vice versa.

 

[modem7]

On this particular PSU the P8 and P9 aren't separated it's all one plug. So it was a complete reversal with the 12v rail going where the 5v is and vice versa.

So, +12 volts applied to the +5V pin on about 100 chips, each chip designed for about 5.25V maximum.
If you were to unsolder all chips, you may find cracks on the underside of some.

 

[Flamin Joe]

So, +12 volts applied to the +5V pin on about 100 chips, each chip designed for about 5.25V maximum.
If you were to unsolder all chips, you may find cracks on the underside of some.

Would the damage be something which could be picked up by a multimeter? I was hoping to narrow it down to particular set of chips rather then go about desoldering the entire board.

 

[Timo W.]

Shouldn't the +5V line be protected by a fuse? I would check first if the +5V are feed to any of the chips with the PSU connected the right way. If not, look for a fuse near the power connector and replace it.

 

[g4ugm]

Would the damage be something which could be picked up by a multimeter? I was hoping to narrow it down to particular set of chips rather then go about desoldering the entire board.

I doubt it. As the PSU lines are common to all chips a multimeter is not going to be much use.

 

[sergey]

Hmm... the power connector is polarized so your friend must have invested a lot of effort into plugging it in reverse I am not sure it is practically possible at all...

So... I think he either switched P8 and P9 (but you say that these connectors are in one plug... I personally haven't see anything like this) or motherboard was plugged correctly, and it were no power supply issues at all, and there is some another issue.
Depending on the equipment available to you, you can do some diagnostics. E.g. many multimeters come with a frequency meter (usually limited to 4 MHz). You can try measuring frequency on PCLK output of 8284, or frequency on A0 line of the processor, etc. See if there are any signs of life there. If so, plug diagnostic ROMs, etc.

Now if you're 100% sure that power connector got reversed and the +5V rail got +12V, most likely ICs are well toasted, and as a challenging project you can take out all the ICs (just cut leads and unsolder pins), solder sockets instead of ICs, and put new/working ICs.

Have fun!

 

[Stone]

Hmm... the power connector is polarized so your friend must have invested a lot of effort into plugging it in reverse I am not sure it is practically possible at all...

That's the same as I was thinking.

However, I do have a friend that managed to force the molex connector from an IDE drive in upside down. Needless to say there was much damage when he turned it on.

 

[SomeGuy]

Is this an IBM XT motherboard or a clone?

I would start by testing the socketed chips, such as the CPU, ROM and RAM. That should give you at least some idea what you are up against.

Were there any I/O cards plugged in at the time?

 

[Stone]

Is this an IBM XT motherboard or a clone?

Since there's no P8/P9 it would have to be a clone, wouldn't it?

 

[Chuck(G)]

Might only be just the PSU. I've certainly seen clone power supplies with a combined P8 and P9.

In any case, I don't see an easy way out of this. Get out your 'scope and DMM and start working your way out from the CPU. You might get lucky.

One of the simpler ways of toasting a PROM is to insert it reversed in its socket. So you'll also want to start by replacing the PROMs--a diagnostic PROM, such as the Landmark one, would be a good start.

 

[Flamin Joe]

Hmm... the power connector is polarized so your friend must have invested a lot of effort into plugging it in reverse I am not sure it is practically possible at all...

So... I think he either switched P8 and P9 (but you say that these connectors are in one plug... I personally haven't see anything like this) or motherboard was plugged correctly, and it were no power supply issues at all, and there is some another issue.
Depending on the equipment available to you, you can do some diagnostics. E.g. many multimeters come with a frequency meter (usually limited to 4 MHz). You can try measuring frequency on PCLK output of 8284, or frequency on A0 line of the processor, etc. See if there are any signs of life there. If so, plug diagnostic ROMs, etc.

Now if you're 100% sure that power connector got reversed and the +5V rail got +12V, most likely ICs are well toasted, and as a challenging project you can take out all the ICs (just cut leads and unsolder pins), solder sockets instead of ICs, and put new/working ICs.

Have fun!

I'm not sure if it's by bad design or not but in this particular case it doesn't really take that much effort to plug it in reverse. So it was definitely the case as he said he checked it afterwards and realised his error. But your right, a PSU with a P8 and P9 is definitely a lot harder to plug in and would definitely tell you your doing it wrong. But this single connector does make it a little easier.

Is this an IBM XT motherboard or a clone?

I would start by testing the socketed chips, such as the CPU, ROM and RAM. That should give you at least some idea what you are up against.

Were there any I/O cards plugged in at the time?

It's an XT clone in fact you can find the same motherboard on my website below. It's from an Atari PC3 8088, it was one of a few I had which I moved on to make room. In particular see the link below for the motherboard in question. Unfortunately as you can see the only thing socketed on this particular revision is the BIOS chip.

http://www.ataripc.net/wp-content/uploads/C103272_Rev.jpg

The only two I/O cards installed at the time is a VGA adapter (which I've tested to be still working) and a RLL controller (which I haven't got around to checking yet).

Might only be just the PSU. I've certainly seen clone power supplies with a combined P8 and P9.

In any case, I don't see an easy way out of this. Get out your 'scope and DMM and start working your way out from the CPU. You might get lucky.

One of the simpler ways of toasting a PROM is to insert it reversed in its socket. So you'll also want to start by replacing the PROMs--a diagnostic PROM, such as the Landmark one, would be a good start.

PSU tests fine. I tried another known working PSU, this time one which has the separate P8 / P9 connectors and it's the same result. But yes I think you are right, there is not going to be an easy way out of this.



Thanks for the input guys definitely food for thought and has given me some ideas on what to check. Happy to hear more suggestions though.