Shuttle SB61G2, shorted NB? && Cap Replacement Question

[Sudos]

I have a 10 year-old Shuttle SB61G2 here that I seem to have stupidly killed.
I was testing different ways of mounting LED strip lighting inside for some illumination, but ended up letting the tip droop down and touch a resistor in between the RAM slot and the capacitor on the bottom middle of the board shot below. for reference, that's the capacitor with R233 marked by it, not for it, but just to be clear about location. The area it touched is in the vicinty of the 4 resistors there. They all test out fine, so I powered it up.

Note: the motherboard model is FB61, v2.1.

Plugging it in, I get a noticeable whine from he coil to the right of the southbridge (heatsink'd) above the two Mosfets. Both of those also check out fine. Also noticing the northbridge getting unusually hot while off with just the power plugged in. you can see where I put my bare finger to the core to confirm this. the machine turns on from this point (heatsink back on), but does not POST. fan spins. the LED next to the fan connector to the right does not turn on.

My stupidity is the weakness here, letting the end of the LED strip touch the resistors in the first place and effectively either putting ground or +12v to them directly. I'm not sure what one it touched in that small area and what it could have fried, but my worst nightmare would be the 865 NB. a new one will cost me a kidney on eBay, something I'm willing to do, but I want to see if it's repairable or not. I can take more pictures if needed.

Secondly, once I get this working again, I'm considering swapping most of the capacitors in important areas like the CPU and by the PCI/AGP slot with polymers, since these are all the original OST caps, not one bad, but aged well enough I'm going to take precautions. I have a Pentium 4 Extreme Edition 3.2 I want to drop in, but can't now, because of this little issue. if I replace the caps, which polymers should I shoot for?

Other cards installed while the machine was on during the event were an HD3650 512M AGP card and an Audigy 2 ZS PCI card. the RAM is standard run of the mill Kingston KVR 2.5v DDR-400 in the total of 2GB.

Picture of Motherbaord (10MB!)

Picture of Motherbaord (smaller, 5.3MB)

[mariushm]

Maybe you just popped the resistor or some capacitor connected to that ground plane?

Right by the capacitors you have a ground and a voltage trace... the large capacitor by the R233 seems to be connected to those large traces. There's also C104 at the end on the left.

The whining may be an overloaded regulator or power supply, which could be a sign of a short. You may have put 12v on the ground while there was 5v or 3.3v on the other trace. That may have blown up some of those ceramic capacitors between the traces or some ceramic resistors, any of which might be now shorted.

Check if there's continuity between those copper traces (put the probes on each side of R233 or on the sides of CD104... if there's full continuity, see what other values those resistors have and try to determine where the short is.

/ nice 10 meg picture, though a bit grainy

[Sudos]

Originally Posted by mariushm

The whining may be an overloaded regulator or power supply, which could be a sign of a short. You may have put 12v on the ground while there was 5v or 3.3v on the other trace. That may have blown up some of those ceramic capacitors between the traces or some ceramic resistors, any of which might be now shorted.

Check if there's continuity between those copper traces (put the probes on each side of R233 or on the sides of CD104... if there's full continuity, see what other values those resistors have and try to determine where the short is.

/ nice 10 meg picture, though a bit grainy

You can thank Sony for that. stupid camera doesn't even have a manual macro setting.

R233 and CD104 as well as CD104 and a few other capacitors around all have a continuity of .005 with my cheapie Radiocrap meter. the ones marked CD56 and CD57 up by the NB and coil show a continuity of 0. so I'd think that those are blown then? there's also that little group of smaller ones and resistors as well under that, I checked a few there and was given quite a few different readouts from .075 to .901 on one of them. so I'm not even sure at this point.

Also: Here's an image of the underside of the board. I upped the contrast slightly to better the details of the traces. you can see where they go from in that area straight into the northbridge, and that scared the ever-living crap out of me when I saw it.

Motherboard Underside (10MB!) ...and it looks like my board picked up a straggler! little fuzzy stuck to the hole by ZCD23 has been removed!

[mariushm]

Yeah, well, it's obvious you have a short there. but 0.005 and 0 are too close together to be really sure.
I'd recommend touching the probes together to eliminate the internal resistance of the probes, then using the REL mode if your meter has one, and try again.
But unless you have a 6000 count or better meter, it might be too hard to determine where the resistance changes to a much bigger degree to indicate a dead short.

Another option would be to take out resistors and capacitors one at a time and test each one separately until there's no more any short... it's doable but you have to be careful not to lose the resistors and capacitors.

[Sudos]

no REL mode on this cheap meter, a real bummer.

I'll take a look later on. for now, it's work time.

[momaka]

Put heat sink back on NB, remove all add-on cards, drives, and peripherals except for keyboard. Then check voltages on all important rails on the motherboard when it's turned ON and when it's OFF but still plugged in.

Important rails:
- CPU VCORE
- Northbridge VCC (the coil above the Southbridge, should read about 1.5V for i865 chipsets)
- RAM primary (about 2.5V to 2.6V)
- RAM secondary (about 1.25V or 1.3V, I don't remember the exact spec for DDR)
- AGP bus (if the motherboard has AGP - usually 1.5V or 0.8V for newer boards)

That's about it. There are more rails, but they vary between different motherboards and generations.

NB data sheet can be found here:
http://download.intel.com/design/chi...s/25251405.pdf

By the way, how did you establish that the MOSFETs for the Northbridge VCC supply are fine? Reason I ask is because ceramic caps CD56 and CD57 filter the NB VCC in addition to the electrolytic caps, so if the lower MOSFET for the NB VCC is shorted, ceramic caps CD56 and CD57 will appear shorted too (as would the Northbridge). Also check the gate resistance of those MOSFETs to ground.

When performing any voltage readings, I would say avoid powering the motherboard for any extended period of time.

BTW, nice picture quality. Really REALLY helpful.