Computer Boot Problem; Stumped

[SugarCoat]

Okay i'm pretty stumped on this. I finally got around to installing watercooling into my latest PC, which i should mention never gave me one problem, and when i put everything back together for a test run the computer began to act stragely. Basically it started by failing to POST and then shutting itself down after running for 10-15seconds showing a blank screen. Basically acting like the processor was overheating from a bad OC, it wasnt OC'd. I started by clearing the CMOS and removing a stick of ram which didnt help, it continued to boot up for a few seconds and shut down. I yanked the newly installed water block off, pulled and unplugged all non-essential components and tried to boot again, same deal. I changed the motherboard CMOS battery, same deal. Eventually it started doing what its doing now, i push the power button and its flicks on and off for about a second. Sounds like a short right? I thought so too. I pulled all the guts out and tried to see if it would even stay on, same deal. Quick on and off.

Just to sum up things i tried

changing CMOS battery
resetting CMOS switch
removing motherboard and components entirely from computer
a brand new motherboard
tried to see if it would stay running without ram, processor, graphics card, sound card, and other random component configurations.
plugged in the Processor HSF to see if there was some issue with it not liking the lack of a CPU fan
moving the computer to a different wall socket in the house


Basically i'm out of ideas other then the big obvious one, Power Supply. Two things made me think otherwise, one is the fact that it was a $250 Server rated PSU about 2 years old and never given me any trouble until yesterday (plus i just installed water cooling, surely i did something else wrong where the watercooling is actually effecting components), and the other is the fact that i quickly tried plugging in my old cheapo DFI 939 board without any components and it stayed running for about 5 seconds before i shut it off, this was after it began having its quick on/off fits. I'm in the process right now of pulling the PSU out of the computer and trying it with the guts of my previous 939 system, but to be honest i'm hoping it doesnt work because if it does i have no idea what to look at next. I've never really seen a PSU thats thrown a fit quite like this, to me it definitly seems like something is shorting, perhaps internally in the PSU, but if it does infact run the 939 guts i'll be pretty damn confused. The only difference between the two setups are the 8pin 12V vs the 4pin for the AUX motherboard power.

If anyone else has any ideas i'm all eyes but i think i covered it all pretty well so i'm problably SOL. My luck the one thing i dont change and planned on keeping for the longest would die first, if it is infact the PSU, which i dont see how it cant be.

 

[zsouthboy]

Incredibly stupid but apt question:

Have you tried plugging a 3 prong fan into the CPU fan header on the motherboard(s)?

EDIT:
Dammit, I looked back at your post, and you mentioned that you have.

PS sounds like it, at this point

 

[ERP]

You'll get exactly this if there is a problem with the CPU cooler, either not contacting the CPU, or not getting enough cooling.

 

[Monty]

Sounds like you have tested everything but the psu. Try unpluging everything from the psu apart from a fan and then short the green and black wire on the 24\20 pin connector. If it only runs for 10 secs then you know for sure its the power supply. If it continues to run, doesnt entirely discount it as a problem. If it doesnt start at all the psu may need a minimum load for it to start, most new fairly decent psu's have such a feature.

Or you could just try another psu

As for cpu overheating or cpu fan not installed, most motherboards have a constant beep on default bios settings for such failures

 

[SugarCoat]

You'll get exactly this if there is a problem with the CPU cooler, either not contacting the CPU, or not getting enough cooling.

that doesnt stop the computer from doing a Power On Self Test though ERP, which is my problem, it wont even boot really. Even without proper cooling or a HSF of any kind you can usually get into the bios for 5 seconds before the processor overheats, but with how fast the computer turns on and off it doesnt have a chance to overheat. Likewise the motherboards should turn on when there is nothing in them, which the 2 boards i tried never did. I plugged a peice of junk PSU into it and it turned on and stayed running (oddly enough the *broken* PSU worked on the AMD motherboard), so it is infact the PSU. 90% sure anyway.

Going to run out and pick up a Antec TruePower Trio 550 and see how that does. Cheap but effective solution.

Sounds like you have tested everything but the psu. Try unpluging everything from the psu apart from a fan and then short the green and black wire on the 24\20 pin connector. If it only runs for 10 secs then you know for sure its the power supply. If it continues to run, doesnt entirely discount it as a problem. If it doesnt start at all the psu may need a minimum load for it to start, most new fairly decent psu's have such a feature.

Or you could just try another psu

As for cpu overheating or cpu fan not installed, most motherboards have a constant beep on default bios settings for such failures

Well like i said, it kinda has to be the PSU, i tested everything else. I doubt it got a new feature, like i said it was working fine with everything for years, this is the first problem, even with newish hardware which it was working fine with until i tore it down to install the tubing.

 

[Sobek]

*edit* duuur. Never mind, sorry.

*edit 2* I had an almost identical problem with a board I used when building a system for a friend, but it would very rarely actually boot into windows and work. In the end it took a BIOS update that had absolutely nothing in the release notes that indicated it would help, but it did. If I read correctly, you can't get anywhere near windows, right?

 

[SugarCoat]

*edit* duuur. Never mind, sorry.

*edit 2* I had an almost identical problem with a board I used when building a system for a friend, but it would very rarely actually boot into windows and work. In the end it took a BIOS update that had absolutely nothing in the release notes that indicated it would help, but it did. If I read correctly, you can't get anywhere near windows, right?

nope wouldnt even post.

So far the following is dead;

670Watt PSU
AW9D Max
X1800XT vid card

finally got it to boot by installing a new P965 Quad GT board and an old X800XT, comically enough the mouse doesnt work at all. The good news is the water cooling is working like a charm and the proc is at 30C down from 55C.

While installing the new motherboard i noticed it was shorting when i put it in the case (yes i'm using insulating washers and spacers) it looked like it was shorting from one of the brackets in the addon slots in the back. Possible cause of the trouble, would a short still boot then fry components like that? I have never seen so many things go wrong like this, dominos effect. Anyway its better but not quite great, still have to figure out what the hell is wrong with the mouse detection and then remove the old drivers and install the new ones.

 

[{Sniping}Waste]

Open up the PSU and look were most of the wires are soldered on the board. You should see many electrolytic capacitors there. See if any of them have a bulge on the top of them or if any of them looks like something has leaked out of them on the top of them.

 

[digitalwanderer]

Open up the PSU and look were most of the wires are soldered on the board. You should see many electrolytic capacitors there. See if any of them have a bulge on the top of them or if any of them looks like something has leaked out of them on the top of them.

But whatever you do, DON'T LICK THEM!!!!

Man, I wish someone had told me that one before I'd opened up my first power supply...

 

[{Sniping}Waste]

But whatever you do, DON'T LICK THEM!!!!

Man, I wish someone had told me that one before I'd opened up my first power supply...

The only thing to lick for testing is CMOS battery's and nothing else.

Most PSU fail because the filtering electrolytic capacitors go bad from over heating. When over heating the pressure inside the electrolytic capacitor goes up and starts to draw a voltage. So if the bad cap is drawing 1 volt and its filtering the 12v rail then you will only get 11v off the 12V rail. This is what is called a voltage drop across the electrolytic capacitor. The more its overheats then more the voltage drop and it gets to a point where the system will be unstable or not even power on. For a cheap repair you can see the microferit rating on the bad cap and the voltage rating and replace it with a good one. I have bought 500W and 600W PSU with this problem for $5 and fix them with a cost of $3 for the electrolytic capacitor.

 

[Mize]

(yes i'm using insulating washers and spacers)

Why? The bolt holes in motherboards are plated contacts to the ground plane in the PCB.

It would take a lot of juice to fry all that...there would have to be a scorch somewhere on the MB/case/etc. that would be a give away.

 

[SugarCoat]

Open up the PSU and look were most of the wires are soldered on the board. You should see many electrolytic capacitors there. See if any of them have a bulge on the top of them or if any of them looks like something has leaked out of them on the top of them.

its to be disposed of, and replaced, would you like me to open it to satisfy your curiosity?

Why? The bolt holes in motherboards are plated contacts to the ground plane in the PCB.

It would take a lot of juice to fry all that...there would have to be a scorch somewhere on the MB/case/etc. that would be a give away.

Because its good practice! I'd spray down the whole damn thing with conformal coating if it didnt void the warranty! You do need to use the spacers though. Some cases have rises at the mounting points and dont need them, but the ones that dont will short without spacers, i know this from experiance. None if it is necessarily "fried". The Video card does the same thing the AW9 motherboard does, basically they'll turn on but thats it, acting exactly like it would if it had a bad firmware flash.

 

[Geo]

Sounds to me like the cpu fried, dude. It wasn't getting enough contact with the new watercooler, and it was overheating quickly. Did that a few times, and now it's dead, Jim.

Hope I'm wrong tho. . .

 

[the maddman]

Why? The bolt holes in motherboards are plated contacts to the ground plane in the PCB.

It would take a lot of juice to fry all that...there would have to be a scorch somewhere on the MB/case/etc. that would be a give away.

It's actually better to have the board grounded at as many places as possible. You want to dump any ESD to ground as soon as possible. Those stupid red washers cause more damage than they prevent.

 

[Mize]

I agree on using standoffs (aka spacers), but conductive ones (brass is typically bundled) and the the nickel plated bolts. Many ground points should mean less noise, etc.

 

[SugarCoat]

Sounds to me like the cpu fried, dude. It wasn't getting enough contact with the new watercooler, and it was overheating quickly. Did that a few times, and now it's dead, Jim.

Hope I'm wrong tho. . .

CPU and the sound card, HDDs and roms are the only things that didnt die. Honestly the waterblock was foolproof, if you mounted it wrong you'd have to be an idiot so that wasnt it.

Like i said to ERP, even an overheated processor POSTs before it shuts down the computer.

 

[the maddman]

CPU and the sound card, HDDs and roms are the only things that didnt die. Honestly the waterblock was foolproof, if you mounted it wrong you'd have to be an idiot so that wasnt it.

Like i said to ERP, even an overheated processor POSTs before it shuts down the computer.

Most likely the powersupply died and shot a power surge through the board. It really sucks when they go this bad.