Need help troubleshooting non-starting system

[w1ntermute]

Hi,

I recently built a new box with the following parts:

GIGABYTE GA-MA770-UD3 AM2+/AM2 AMD 770 ATX AMD Motherboard
AMD Athlon 64 X2 7750 Kuma 2.7GHz 2 x 512KB L2 Cache 2MB L3 Cache Socket AM2+ 95W Dual-Core black edition Processor
NZXT TEMPEST Crafted Series CS-NT-TEM-B Black Steel / Plastic ATX Mid Tower Computer Case
Rosewill RP550V2-S-SL 550W ATX12V v2.01 SLI Ready Power Supply
G.SKILL 2GB 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Desktop Memory
Arctic Silver 5 Thermal Compound
EVGA 128-P1-N309-LX GeForce FX 5200 128MB 64-bit DDR PCI Video Card

When I first started up the machine, it seemed to be working normally. I didn't have a monitor connected, so I couldn't see exactly what was happening. I turned it off and on again once, and it worked.

After that, every time I turn it on, it turns on for a second, and the case lights and fans turn on, and then they all turn off. Then if I push the power button, nothing happens, unless I wait a few minutes first, in which case it once again turns on for a second then turns off.

I thought the PSU was defective, but I RMA'd it and just got the replacement, and I'm still having the same issue. Could someone help me troubleshoot this problem? I'm leaning towards the motherboard being defective, but I'd like to be sure before I RMA another part.

Thanks!

 

[foolycooly]

Probably a heat issue...double check that you mounted the HSF correctly and that it is tight. After leaving it off for a while try to boot to the BIOS and check the temperatures there. While your PSU is of low quality, it should be enough to power that system.

 

[shortstuff_mt]

This sticky exists just for troubleshooting problems like yours:

http://www.tomshardware.com/forum/261145-31-read-posting-boot-problems

 

[w1ntermute]

I figured out the problem (sort of). That case has a lot of fans. When I disconnected all the fans except the HSF, the PSU fan, and one case fan, it successfully booted. Is this because the PSU can't handle any more than that?

 

[Shnur]

Very weird... the PSU should be able to handle everything. Personally my computer boots for 1 second when I change voltages or when I shut it down via the PSU button, then it shuts down and starts again a couple seconds later...
Did you wait before pressing the button again?

 

[w1ntermute]

Yes, I waited for 2-3 minutes between each boot attempt.

 

[toledo_speedo]

I am having the same problem atm and I am guessing it is the PSU as well. They will overheat and shut down so having to let it cool for 2-3 minutes makes sense. I have had trouble with Rosewill in the past so maybe borrow a PSU or RMA it again and see if you have any luck.

 

[w1ntermute]

Well, I moved it to a different room, and now I'm getting the same problem again. Is this most likely because of a loose connection or hardware problems?

Edit: I discovered something else that may be relevant. Like before, PSU switch has to be turned off then on again to get the system to start up for a second. ~13 seconds after I turn the PSU off again, a small hissing noise emanates from the PSU. After that, I can turn it back on and the box will turn on for a second.

Edit 2: It's working right now, with no explanation. Is there some way I can use this state to test for loose connections, etc.?

Edit 3: I moved it a few inches across the table it's on and it stopped working again 🙁

Edit 4: I moved the case again without touching _anything_ inside it and it's working again.

Edit 5: I tried removing/replacing the CMOS battery, and it started working again, but then it stopped. I think I'm going to RMA the motherboard.

Edit 6: I figured it out. The motherboard was screwed on too tightly, so it was occasionally touching the case and shorting out.

 

[Shnur]

^ +1
That's... awesome troubleshooting man...
I was aiming for something shorting, but I never saw a board going that direction.

 

[w1ntermute]

And the weird thing is that there are even plastic standoffs on the back of the motherboard. I guess the side of the case it rests against not being perfectly flat was the source of the issue.