Optiplex GX100 - Cap failure? [NooB]

[Cephas]

Today I tried upgrading my Optiplex GX100's Celeron 433 to a Celeron 733 [I didn't think it'd pose a problem as they both use the same socket]...
After I plugged all the peripherals back in and pushed the power switch, it powered up but nothing showed up on my monitor. After about 15~30 seconds of noise and activity, the computer abruptly shut down.
I immediately opened the case and attempted to boot again. As I watched, there was a sudden emision of smoke from the area a red and white arrow points to it in the attached picture. The air also stank.
I immediately unplugged the system and replaced the new processor with the original. Tried to boot again, but though there's system activity (you can hear the CDrom search during boot time, HDD activity, etc.) there's no signal to the LCD monitor.
Anyone have any idea what happened & what I can do?
After coming across this forum, my first guess is that I somehow busted some caps, but [pardon my ignorance] I thought caps were the "battery-looking things" on the motherboard. The parts that look fried or with leaks (see arrows in close-up diagram--apologies for the funky quality) look like ... I don't know. Are they caps?

[gonzo0815]

Re: Optiplex GX100 - Cap failure? [NooB]

Are your shure, that the CPU is ok and there is no short betwenn the board / case? If not either the VRM can´t supply the 733 Celery or forwhatreason ever the Transistors have given up and shorted.

[kc8adu]

Re: Optiplex GX100 - Cap failure? [NooB]

looks like the vrm mosfet to the left of the blank pad nearest the red/white arrow is blown.

[Cephas]

Re: Optiplex GX100 - Cap failure? [NooB]

thanks for the quick reply. The CPUs should both be okay... they worked this morning [the 733 in another computer]...
Are those things Transistors, then, and do they look busted?
PS. Ah. It's a VRM Mosfet, eh? I'll go look that up.

[linuxguru]

Re: Optiplex GX100 - Cap failure? [NooB]

The Coppermine Celeron 733 consumes about the same amount of power as a Mendocino Celeron 433, but at a lower voltage (1.75 vs. 2.05) and proportionately higher current. The VRM may not have been able to handle the higher current, causing the transistor to blow. You may also need the appropriate microcode update in the BIOS to support a Coppermine Celeron.

[Per Hansson]

Re: Optiplex GX100 - Cap failure? [NooB]

Well, your arrows are pointing at two unpopulated solder pads

Had there been anything there it would have been MOSFET's

But it looks like the bean counters did away with two of them, there are only two MOSFET's on your board, the two black things close to your arrows... They are most likley blown...

The capacitors are the can looking things, yours looks like Sanyo and if they are it is an excellent brand...

But look on the MOSFET's more clearly for cracks, the crack when they break most oftenly (because they get so hot so the chisel expands...)

There are also very nice smoke effects just as you saw :P

[kacper69]

Re: Optiplex GX100 - Cap failure? [NooB]

This matherboard don't support "green" P3 processors, only "black".
you can see , this processors are different.

[Rainbow]

Re: Optiplex GX100 - Cap failure? [NooB]

Yes, there are boards with Socket 370 that can run only Mendocino CPUs - but it should not smoke with Coppermine.
Black = Mendocino Celeron
Green = Coppermine Celeron or P3

[linuxguru]

Re: Optiplex GX100 - Cap failure? [NooB]

S-370 boards that support only the Mendocino can smoke with a Coppermine, if the VRM controller doesn't support the VID codes for 1.65V..1.75V. Depending on the design, it might set the Vcore to a default value - maybe 2.05v, maybe 3.5v. Whatever it is, that's much higher than a Coppermine is designed for - it may work for a few minutes, drawing much higher currents than the VRM is designed for. It will sooner or later smoke either some VRM components or the CPU itself.

[Rainbow]

Re: Optiplex GX100 - Cap failure? [NooB]

There are some VRM controllers that does not support voltages below 1.8V - but they should turn off then.

[gg1978]

Re: Optiplex GX100 - Cap failure? [NooB]

Sounds like you popped a FET, like kc8adu said.. Now as to "why" the FET popped, probably it had the correct Vid for the VRM, so you weren't overvolting the CPU, but the coppermines like their current, and that's a single phase design there.. As for the empty pads, those are for additional FET's to supply higher CPU current. There's nothing wrong with the empty pads..


I've got a GX110 at home serving as a linux firewall, same form factor and everything, but it takes a slot 1 coppermine chip.. Board layout is totally different.

It has a VRM with 4 tiny phases and not one aluminum electrolytic axial cap.. Everything is SMT.. Then again, with a slot 1 chip in there, there was no space for a fullsize VRM with axial caps..