[problem sovled]I am about to give up

[konakona]

cut to the chase, the mobo doesnt seem to be able to power itself up. I have tried connecting reset pints with a screw driver, so its not the power button. No fans turn (i only have one in the PSU but anyway), the diiagnostic LED doesnt turn on either. Strange thing is though, if I put my ears real close to the CPU, I can hear a very tiny noise coming from it, like a airplane taking off or something about to errupt. Sorry about the funny language, but I am at loss of words, cant come up with a better descrption of what I was hearing. Same thing happens with two other PSUs, so its highly unlikely the PSUs are the culprits. Went through 3 different AGP cards with no luck, tried every possible combination of putting two ram sticks in 3 different DIMM slots at no avail. Needless to say, the whole thing is stripped down to nothing other than the PSU, mobo, cpu/HS, vga. Tried resetting the jumper and resitting the battery multiple times and nothing happened. As a last resort, I reinstalled the HS and its backplate/retention as well.

Now the mobo is taken out of the case, and careful visual inspection reveals there is no sign of bulging or leaking caps anywhere on it.

Any ideas?

I can write further on what might have lead to this situation if I dont get enough feedback (long story, want to see what people can come up with first)

This system has been running fine for past 8+ months and the mobo in question is Epox 8KDA-3J (nforce3-250GB / 2 kingston 512 valueram sticks / 2800+ palermo)

 

[LokutusofBorg]

Sounds like your mobo is dead, so it'd be interesting to hear the story you mention.

If you can take everything out of the picture but your motherboard and a known good PSU (and your case speaker) and you don't get POST beeps telling you the stuff it needs isn't plugged in, then you need look no further than your mobo...

 

[konakona]

Well, first thing one of my hdd seems to have bitten the bullet - gets detected in bios just fine, but cant be diagnosed or read in the OS. The hdd keeps on stalling and lagging the winXP like crazy, so probably needs to be reformatted for any hope of rejuvenation. Not so much of grief for the dead hdd, but man all that stuff I had on that HDD

Well as if that wasnt bad enough.. when it rains, it pours.

I went ahead and switched the HS from zalman 7700Alcu to ninja scythe. The scythe unit is awesome, much quiter than the zalman when it was running for few minutes. Then the trouble came. The system ran just fine fanless, but seeing how the 120mm yate loon that came with it was sufficiently quiet, I decided to put it on the HS. The fit was very tight over the ram modules, but I managed to pull it off. The case wouldnt close all the way tho, so I had to force it. The computer shut off itself once in the process, which I should have taken as a warning sign. It rebooted into windows just fine though, so I went to sleept to work on it the next morning. When I came back though,the computer was shut off. Upon pressing the power button, it would momentarilly flash the power LED and go out again. When I took the side panel off and reinserted the power cord, power came on fine but there was no POST. So I took apart everything except for the HS/backplate, and put everything back together. Now it wont get any power at ALL, even with another PSU that is proven to work on another computer. At this point, I suspected the backplate which was made of metal and had adhesive rubber padding on it. The seller did a terrible job of removing it from his old rig, with the rubber pad torn and missing here and there. I thought I may have done a hasty hackjob at reinforcing it with insulating foam, so I redid it with insulating tape. Unfortunately, all that effort of dismantling the HS (btw, it took me forever to do it. you really need to use a screw driver to pry it off i guess) didnt do squat to help the situation. To sum it up, it could have been the shortening due to either pressure applied directly above from the DIMM modules or bare spots on the conducting backplate. I doubt its the latter though, since if it were the case, the computer woudlnt have been able to boot at all (and it did for a while before i hit the sack)

That noise I am speaking of aint the coil whine, it is barely audible with my ear just about the HS. it is coming from the general direction of the CPU, thus my guess that it is coming from the CPU.

If it is indeed the mobo, I could probably take a gamble with newegg refurbs, I had pretty good luck with those. The sad thing is, however, I am not sure if it is the mobo cuz it didnt boot with seemingly working mobo prior to this. At that time, the diiagnostic LED would go back and forth between error codes C1 and FF (Detect memory and Boot Attempt, respectively)

Is there a way to tell the difference between a dead cpu and a dead mobo? I am hoping that weird noise I am hearing could be an indication of which part is malfuctioning. If both of them are dead, I might as well upgrade to opty144 or something. Too bad I dont have another testbed to check for the DIMMs, but my understanding is the mobo should at least get the power and give off error beep in the absence of functioning ram.

 

[Fardringle]

If you had to force the case closed on top of the heat sink fan, I would suspect that you either crushed the CPU core (or something inside it) with the pressure, or you forced the motherboard to flex around the CPU socket enough to break some traces in the board.

At this point, the only real ways I can think of to be sure which component is causing the problem is to plug the CPU into another motherboard, or put a different CPU in this motherboard. Unfortunately, you'll run the risk of damaging more components by doing this. You might be able to narrow things down by taking the CPU out of the socket and then trying to turn the computer on. If you still get the "whine" from around the CPU socket without the processor, then the problem is definitely in the motherboard...

 

[konakona]

I may have given you an impression that excesive force was used to close the case which was clearly not the case. Rather, the case side panel had to be bent very slightly, and I didnt even keep it shut fully before went to bed at last. The 754 cores are pretty durable with large surface area of contact AFAIK, and when the heatsink was dismantled, the CPU didnt seem to have suffered any physical damage anywhere. There is no way the board could have been bent enough to cause trace breakdowns.
Let me reiterate, the fan was NOT on top of the HS, rather on thop of the DIMM, you will probably know what I am tlaking about if you seen the Scythe ninja in action. the fan is mountable on 4 different sides around the HS, but not on top of the HS. I doubt there is enough room between the mobo and the other side of the case to allow that much bending to occur anyway.

Thanks for the suggestion of taking the CPU out, havnt thought of that one

 

[Severian]

have you tried posting the mainboard outside the case? If you are suffering from an intermittent short, this would be the easiest way to diagnose it.

I think it's more likely the mainboard has gone to that great boneyard in the sky though....although you can't definitively rule out either the cpu or mb without proving one works with another good component, and the other does not.

 

[konakona]

Originally posted by: Severian
have you tried posting the mainboard outside the case? If you are suffering from an intermittent short, this would be the easiest way to diagnose it.

I think it's more likely the mainboard has gone to that great boneyard in the sky though....although you can't definitively rule out either the cpu or mb without proving one works with another good component, and the other does not.

That actually came to my mind, so I tried it without the screws locking the board to the case, but still on the metal pedestals where screws get screwed onto. I guess I should try it completely out of the case, would any non-conducting surface be safe to test it? carpet isnt a good example would it be? maybe few sheets of paper?

If I had to choose between the two, I would rather find the mobo dead since the cpu was doing 2.5 quite well and it costs more to replace the chip.

 

[BadThad]

Originally posted by: VanillaH

Originally posted by: Severian
have you tried posting the mainboard outside the case? If you are suffering from an intermittent short, this would be the easiest way to diagnose it.

I think it's more likely the mainboard has gone to that great boneyard in the sky though....although you can't definitively rule out either the cpu or mb without proving one works with another good component, and the other does not.

That actually came to my mind, so I tried it without the screws locking the board to the case, but still on the metal pedestals where screws get screwed onto. I guess I should try it completely out of the case, would any non-conducting surface be safe to test it? carpet isnt a good example would it be? maybe few sheets of paper?

If I had to choose between the two, I would rather find the mobo dead since the cpu was doing 2.5 quite well and it costs more to replace the chip.

Do NOT use carpet, static hazard. The best thing is a piece of cardboard actually. The only things connected: cpu, ram, power supply and monitor.

 

[Fern]

would any non-conducting surface be safe to test it?

I usually use a phonebook. The height allows space for the lip of the gfx card to hang over the edge, etc.

Fern

 

[konakona]

ugh... the weirdest experience EVER. That strange noise (sounds like some messed up radio signal) wouldnt go away until the Njnja heatsink was removed; when I plugged in the power without a HS on the cpu, the mobo powered up! I quickly unplugged the power cord and grabbed my old 7700alcu and let it sit on top of the CPU. The diagnostics LED went through a whole bunch of things and settled at "7F", which stood for

1 . Switch back to text mode if full screen logo is supported.
-If errors occur, report errors & wait for keys
-If no errors occur or F1 key is pressed to continue:
Clear EPA or customization logo.

so I hit F1 there, and the code changed to FF! Yet there was nothing on the screen, so I replaced the vga with another AGP card, and was finally greeted with a beautiful POST message and NVIDIA bootrom (since no bootable devices were attached). The last bit was probably because the card wasnt fully seated in the AGP slot.

So was it the heatsink applying too much pressure on the CPU? or could it have been the heatsink pulling in hte backplate too close to the mobo resulting in shorted traces? Whatever it is, I am relieved everything (except for the hdd, probably) is still in good shape. Thanks for advice everyone

 

[konakona]

dunno, the system booted fine with the ninja without the clips latched on, but once the heatsink was secured, that whole noise thing took over again. Maybe I should google for "Scythe ninja", am I the first one to have something so awkward happen to his computer?

 

[konakona]

wow, got it sorted out! so it WAS the metallic backplate coming in contact with the mobo, namely the DIMM slot solder points on the bottom side of the mobo. It might have halfway killed one of the slots in the process (works sometimes, fails yet other times), I still got two healthy ones I could use. So the net result from upgrading the HS is 100Mhz loss in detail with fanless operation. Now I can really hear the VGA silencer when I turn the switch on, whereas it was totally swamped by the noise coming from the zalman alcu7700's fan in the previous setup. The loudest thing in the system now is the PSU, which already sports a low rpm yate loon.

Thanks for great suggestions everyone, especially those who suggested taking the mobo out of the case. Now, if you have just a few more minutes to waste, please take a look at my other thread regarding the hdd failure issue