First post, by stephenlamppost
- Rank
- Newbie
When you hear a 'pop!' and see a spark fly from your board, does anyone have any hints for detecting what failed? In my case there's no obvious burnt-out component.
I'm aware of tantalum capacitors' tendency to fail... the ones I've seen are little blobs in yellow or red packaging. Is it obvious when they've failed? Is there anything else that might fail in a similar way, but not be obvious on inspection?
My story: I bought a CHIPS GW-286CT 286 20MHz board, seemingly in very nice clean condition. It tested fine from an initial quick power-on test to POST using borrowed parts. I also snipped off the on-board battery (which was getting a bit furry) and ordered an external one.
A while later, I'd sourced the necessary parts to build a working machine. I did a quick boot to BIOS with just board + battery + memory + VGA + keyboard, all good again.
Then I added the I/O card (one that had lain in a drawer unused for 5--10 years). The machine would not power on.
I tried again 2--3 times, flipping the switch on/off each time. (This is with an ATX PSU + adapter, so I was flipping the hard switch on the rear... the remote switch for soft-on is kept on.)
The PSU was completely silent, suspiciously like when it detects a short.
Unsure what to do, but thinking it might be a low-power-draw issue, I decided to go ahead and connect power to the the HDD. I flipped the switch again...
Pop! A yellow spark flew up from the area around the middle of the main board, what looked like roughly in between the two cards (I/O and VGA), although I didn't see exactly where -- I couldn't tell whether it came from the main board or a card.
Despite the pop, the machine had booted -- I heard the BIOS beeps at nearly the same time as the pop. I turned the power off quickly. There was a light smell of burnt electronics.
First thought: blown capacitor. Only three boards were connected: the main board, the VGA card and the I/O card, The I/O card seemed to smell much more of burnt than anything else did, and seemed to be what had prompted the problem. Still, I examined it but could not see any exploded cap. It has several tantalums but they all look very much intact.
Experimenting some more, I put the I/O card back in and tried it. The floppy controller worked, but not the HDD: the BIOS claimed "Fixed disk controller is not responding". I successfully booted from a floppy and enjoyed the Secret of Monkey Island demo version...
... sort of. I noticed that during the MI demo, from some point it was ignoring my keystrokes -- I had to reboot. More generally, the keyboard behaviour seems erratic: sometimes the BIOS complains about the keyboard test failing, and sometimes I get random keystrokes and random PC-speaker blips (tiny short blips that I associate from BIOS, e..g. when you hold down a key).
On a whim, I tried the known-good I/O card from my 386. This worked (I booted the machine from the hard disk for the first time), albeit still with the instability issues. I also used the suspect I/O card in my 386, and that worked fine too!
This put my suspicions back on the motherboard. But again, I can't see any blown capacitors.
Any suggestions about what might be wrong, or where to look?
Thanks very much if so!
Full set of photos (sorry the lighting is not great... I can try again on request): here.