I’ve done this sort of thing hundreds of times, so I ought to know what I’m doing by now…
A colleague asked me if I’d take a look at her son’s computer (I’m the IT guy for a smallish company, but with more of an emphasis on development and support than hardware, although I’m still fairly competent in this area.)
Anyway, the machine she brought in would not boot; just producing long beeps; Long story short; a chip fan had failed on the motherboard and the Northbridge chip had been cooked, knackering the thing.
So, she decided to give her own computer to her son; she brought this in and I installed the HD from the broken machine (as a slave) and the RAM (whcih happened to be of the same type in the replacement machine - everything was hunky-dory - the new drive appeared as drive E: in windows and all else was rosy. She took the machine home, later, she phoned me saying it wasn’t working; the power LED turns on as soon as the supply cable is plugged in and pressing the power button does nothing at all; the fans don’t even spin up.
I’ve checked for any obviousl physical problems such as loose cables; I’ve tried booting with all of the hard drives etc disconnected (just to see if it would go to POST); I’ve tried swapping out the power supply; none of it makes any difference.
Am I looking at another dead motherboard here? If so, what could have happened to it?
First off, when a motherboard won’t initialize the display and produces nothing but beeps, you should check the standard beep codes to figure out what’s cooked or misplaced. Maybe you did that to solve the Northbridge problem.
After she moved it back to her house, did it work and then stop working? …or did it fail to boot once she got it home?
As it is, we can rule out most hardware failures (with the exception of a spontaneous failure, which is very unlikely).
It’s possible that in swapping out power supplies you misconnected the power switch leads, but I’m going to give you more credit than that (also: that wouldn’t explain how it could work and then not work). I don’t think that bad ventilation would cause a failure so soon, either.
If they’re using an AGP card for graphics, it’s possible that in handling, the AGP slot could be a loose connection – I always re-seat the graphics card when I’ve got a mysterious boot error like this. One other thing you can try is to disassemble the whole system and reassemble it. Slip a piece of cardboard or paper between the bottom face of the motherboard and the inside of the case; I had a motherboard once that had longer solder pins than it should have, and one of these grounded to the case, shorting out the whole thing and preventing the PSU from doing anything.
The worked-and-then-didn’t-work drives me to think there’s a loose connection that you just haven’t found yet.
Or more likely one of the memory sticks isn’t quite seated firmly enough and has shifted a bit.
That stick will likely be fried but the rest should be OK.
The one that just beeps and won’t start was the one I diagnosed as having a fried Northbridge chip; I removed the HD and RAM from this and added them to another machine (her own) that she brought in; this worked fine both before and after installing the extra hard drive and RAM; it didn’t work after they got it home and now won’t even power up, regardless of what I try.
It’s a dirt cheap motherboard with onboard graphics and most other things onboard also; even with all of the drives and all of the RAM removed, it won’t get as far as powering up to spin the processor fan or even attempting to POST.
I only tried swapping out the power supply after the problem arose and the power button connector connects onto the motherboard, which connection I have not needed to touch. I have tried disconnecting the power button connector and shorting the pins with a jumper, just in case it was a faulty power button.
This is a damn fine suggestion; it may be that plugging in the RAM (plus a few jolts on the journey home) brought about this kind of problem; I only hope that whatever got shorted out didn’t do any permanent damage.
Might even be a dead BIOS battery, they are cheap to replace. I’ve also seen similar symptoms when the BIOS reset jumper was accidentally left in the “reset” position; Mangetout you mentioned using that jumper, any chance it got left in place or that a bit of conductive junk got in there?
Another guess - the Northbridge chip wasn’t the only fried thing, the Ram or HD circuitry you moved was damaged too. Moving them to the second motherboard damaged it in turn.
The fried northbridge chip was on a different motherboard in a different machine; the one I’m having problems with right now is a different machine into which I have added the hard drive and RAM from the broken machine; it was working after I did this, it stopped working sometime after.
I tried taking the motherboard out and running it with a sheet of plastic between it and the case; the symptoms are the same.
The inside of the case is relatively clean and there isn’t any obvious sign of shorts or anything else like that; I’m sure I did replace the CMOS reset jumper in the right position, but even if I didn’t, *the problem was manifest before I tried resetting the CMOS[/]
Misseated RAM sounds like a definite possibility to me. Over the past five years, I’ve also had such symptoms occur with poorly seated PCI and AGP cards, and even slot form-factor CPUs. It’s happened often enough that I’ve quit trying to “snug things down”. It’s a waste of time. I go straight to completely removing and then reseating everything that sits in a slot (prefereably moveing the PCI cards to different slots). It only takes a few minute longer than opening up the case to jiggle and snug everything down, but you’re much more likely to get a recovery.
I’ve got a lot of computers at home, and I’ve been surprised know many “dead” PCs (i.e. unresponsive to snugging down and other basic procedures) turned out to have ZERO dead parts (including motherboards) when I later cannibalized them for other computers Sure, it could be “spontaneous IC healing” (which doesn’t happen nearly as often with todays smaller/denser IC feature size as it did with chips from 10-20 years ago). I’m pretty sure that they’d have working the first time if I’d pulled and raseated all the cards completely, rather than the old “jiggle and rap”
I’ve tried reseating the RAM and the only PCI device in there, which is a modem; I’ve tried detaching everything - RAM, PCI cards, drives and I can’t get so much as a beep or a spinning fan from the MoBo; it is failing to power up.
I have case with a funky reset switch that will stick and look like it’s not pushed in. You said that you removed the power switch and tried to start it by shorting the pins. Was the reset button connected when you tried that. You should disconnect the reset button and give the power button a try.
It’s not a dead BIOS battery. A BIOS will start a computer with the default setting if the battery dies. The settings might not get you to into the OS, but the BIOS menu can be accessed.
The significant point in all of this is that the machine went from sitting on my desk, working fine to sitting on her table at home, not booting; she hasn’t opened the case and fiddled with any connectors.
I’ve tried disconnecting the power and HD LEDs while booting - no joy; I’ve also tried attaching a spare power switch to the power switch pins - also no joy.
I’m very grateful for all these suggestions as well as being somewhat heartened that the steps I’ve already tried are the ones people are suggesting.
The problem could have happened any time after it left your site, and something unrelated to any of this could have failed.
Put the system back how it was. Drive jumper on the master set correct for no slave attached. Remove the newly installed memory and put the orginal back into the original sockets. Try to power up. Let me know how things go, and if you want to email, it is enabled. The next step is a lot more time consuming.
I don’t think the MAST/SLAV jumper would keep the PSU from firing up – nor badly-seated RAM. I suspect that there is a short on the motherboard or in the PSU somewhere, and that this is fundamentally an EE (that is: wiring) problem, and not a CE (that is: chips and blips) problem.
According to the ATX pinout here, you can send an artificial #PS_ON signal (specified at 3.3.2 of the ATX standard) and see if the PSU is truly known-good. You can short pin #14 to GND (should be the green wire) to test the power supply’s compliance.
If the PSU works, then there’s a short in the mobo, and troubleshooting will be much more difficult.
It can’t be the drive jumpers (they’re both on CSEL anyway) because the motherboard won’t even do anything with all of the drives disconnected; not a beep, no movement from the processor fan.
I suppose there’s the possibility that I have two duff PSUs; I ought to check that out tomorrow, although they can’t be dead on all pins, as there’s this power LED lit all the time.
I was just giving a reminder that the drive jumper might have been moved when you added the slave in. It wasn’t anything to do with the no power insue.
I have seen that kind of response with a fried processor also. for shits and grins, what voltage is the power supply set for? It didn’t get knocked to 220V input did it? A piece of metal could have moved around in the case and shorted the board. A screw dropped in the case at some past time, now wedged between case and mother board.
220V is the correct setting for here and I did check that at some point; If some foreign body has shorted part of the motherboard, it isn’t there now; I’ve had it out of the case to check that there wasn’t a long solder peak on the back or anything. I have tried restoring the system back to the original settings; noit a sausage.
I’m going to try shorting that pin on the ATX PSU connector tomorrow. In the meantime, I have ordered another motherboard in case it comes down to replacing it; only trouble is that this is a PC which doesn’t have a proper WinXP install CD - it’s an image-based installation - I’m not sure whether Windows will cope with starting up after a new MoBo installation.