DFI Nforce3 UT Killed 2 3400+ mobiles

[remagavon]

Well it turns out my DFI Nforce3 killed 2 of my mobile 3400+ processors. Being oem, I can't get an rma either. When raising the voltage from something around 1.65 to 1.7ish the board failed to POST. It was definitely a trivial increase and not something like 1.4v to 2.05.

Here's the story:

I read online a bit and figured out that being a mobile cpu, my heatsink may have not been making great contact (it wasn't). I assumed this processor was damaged from excessive heat, so I bought another one and it booted just fine.

A few months passed, and I eventually got a prometeia to try out and subsequently tried the same thing (raising the voltage). As luck would have it, the same thing happened as before, except this time my cpu was at around -10c. I dismantled everything, checked for contact etc; it looked great. There was also no sign of condensation.

I tried booting a few more times with a regular heatsink, and a MOSFET (I think) below the ram slots smoked and burnt up. I RMA'd the board, thinking that it could possibly be only a motherboard problem. My new board didn't work either, the system acted like it was getting shorted out (fans spun for about 1 second). Giving my processor the benefit of the doubt, I replaced the motherboard with an Abit NF-8 (which does fully support mobile processors).

The NF-8 does the same thing. When I remove the processor, the power supply stays on forever, when it's in the socket; no dice. Fantastic. Reading around, I know that DFI has been somewhat flakey, but the board has basically ruined $550 worth of cpus for me. Great.

Just wanted to share my unfortunate experience with DFI's motherboard.

 

[imported_Kensai]

 

[Hacp]

Good thing this doesn't happen with regular Ocing..

 

[remagavon]

Originally posted by: Hacp
Good thing this doesn't happen with regular Ocing..

I've never heard of a case like this before, with any amount of overclocking. The chip that was faulty probably was tied into the memory bus somehow and took out the on die controller similar to what happend with unstable voltage mods on the shuttle 754 board.

 

[Tyson95]

Are you positve the cpu's are dead?

 

[remagavon]

Originally posted by: Tyson95
Are you positve the cpu's are dead?

When I replaced the first one with a new processor, it worked fine. I'm fairly sure that means that at least one is damaged in some way. My second stopped working, and has ben tried on 3 motherboards, so I'm pretty certain that cpu is also damaged.

I have tried all combinations with a plethora of hardware from different power supplies (3 total) to different memory, even different video cards. I can't get my new NF8 to POST whatsoever.

 

[imported_g33k]

no problems here. ive kept all voltages close to stock though.

 

[uOpt]

Welllll,

you didn't only overclock, you overvolted with no temperature monitoring.

Since the mobile chips don't have heatspreaders it is very important that any build with them is verified to be temperature-stable right after powering up. Only then can you go and start fiddling.

Now, I heared DFI boards destroy CPUs and particularly RAM before. But in this case there is no reason to blame the board publically because there is no way to rule out simple heat problems for lack of contact between chip die and HSF.

 

[remagavon]

Originally posted by: MartinCracauer
Welllll,

you didn't only overclock, you overvolted with no temperature monitoring.

Since the mobile chips don't have heatspreaders it is very important that any build with them is verified to be temperature-stable right after powering up. Only then can you go and start fiddling.

Now, I heared DFI boards destroy CPUs and particularly RAM before. But in this case there is no reason to blame the board publically because there is no way to rule out simple heat problems for lack of contact between chip die and HSF.

Temperature on my first processor may have been the problem, but the second was running around 10c max with a full load.

 

[uOpt]

Originally posted by: remagavon

Originally posted by: MartinCracauer
Welllll,

you didn't only overclock, you overvolted with no temperature monitoring.

Since the mobile chips don't have heatspreaders it is very important that any build with them is verified to be temperature-stable right after powering up. Only then can you go and start fiddling.

Now, I heared DFI boards destroy CPUs and particularly RAM before. But in this case there is no reason to blame the board publically because there is no way to rule out simple heat problems for lack of contact between chip die and HSF.

Temperature on my first processor may have been the problem, but the second was running around 10c max with a full load.

At -20 room temperature or on dry ice?

Sorry this sounds like busted temperature monitoring.

 

[uOpt]

Hm, now you even have that blame in your signature.

I really think you are not quiet fair here.

 

[remagavon]

Originally posted by: MartinCracauer

Originally posted by: remagavon

Originally posted by: MartinCracauer
Welllll,

you didn't only overclock, you overvolted with no temperature monitoring.

Since the mobile chips don't have heatspreaders it is very important that any build with them is verified to be temperature-stable right after powering up. Only then can you go and start fiddling.

Now, I heared DFI boards destroy CPUs and particularly RAM before. But in this case there is no reason to blame the board publically because there is no way to rule out simple heat problems for lack of contact between chip die and HSF.

Temperature on my first processor may have been the problem, but the second was running around 10c max with a full load.

At -20 room temperature or on dry ice?

Sorry this sounds like busted temperature monitoring.

Did you read my OP? I said I had a prometeia.. this was definitely not a case of an incorrect temperature reading.