could PSU be the problem?

[vgbaron]

Well - my clock went to xxxx again. Settings that used to work now reboot system when OCCT starts it's check.

I just added a GT 280 to my system. My PSU is a Thermaltake W0131RU 850W PS. I have 5 HD's + a DVD writer along with the GT 280.plus 6-120mm fans and a water pump.

I seem to recall somewhere that someone thought his clock becoming unstable was due to the PSU.

Nick - your thoughts on this?

I even tried a modest clock 400x9 with everything on Auto and it failed. Of course now I'm running 1800 COrsair DDR3 at 1600.

Glad I don't have any hair left.

[ozzy72]

Have you checked the CMOS battery on the motherboard?

[vgbaron]

yes - everything else seems ok.

[NickN]

hmmmmm

Yes a PSU issue could be a culprit

You cant set everything to 'Auto' and run 400x9.. you still must set Vcore and memory voltage to run that.

What is the 12v, 3.3 and 5v readings in Windows when you run the full load OCCT test or when you run a 3DMark benchmark test?

[vgbaron]

I've tried various combos - all with the same result. Before OCCT starts it shows 3.34 and 12.04 - no other voltage readings. Even Prime95 small fft was an instant reboot.

Just to square away my thinking - I double check memtest86 and 4 or 5 passes are stable with the Corsair ram, so can I safely eliminate that as the issue? I guess I could out the Mushkin back in and retest but if memtest is ok.....

I can safely say that every test I've run with ANY settings except default has resulted in an instantaneous reboot when the test starts. Wouldn't mind if it ran for a while and failed - might seem like it was getting somewhere.

I guess I could also put the 8800GTS back in and retry but that 280 is a PITA to squeeze in!  :)

[NickN]

Let me look up the PSU vic and run the numbers

I have to leave but will check this out first

[vgbaron]

UNderstand.

I swapped the 8800GTS and, if anything, the problem is worse! Not only do I get reboots when I test with OCCT and P95 but just browsing in IE7 now has caused a reboot. I'll try again with the Mushkin ram and the last good settings and see what happens.

Thanx!

[NickN]

Vic thats a MODULAR PSU meaning cables that are not solid wired

Check and make dang sure none of the connectors are lose and take a look for hot spots at those plugs too




The fact it just instantly reboot can suggest PSU but that does not mean its the problem. Does this reboot happen with a clock or any operation even unclocked under a load test?

If you can run lower and not reboot run the OCCT load test and see if 12v drops to 11.75 or less, or 3.3 drops to 2.8 or less and if so the PSU is definitely suspect


The PSU you have is only high current on the 12v3 and 12v4 output.. the other 2 are only 18A each.. look at how its connected in your system accordingly





This is why hard wired PSUs that have a SINGLE rail are so much better than modular units.. no matter what plug in use you get the FULL 12v amp ability when needed on single rail PSU's

In comparison the PCP&C 860 will deliver 64-70 peak amps on any 12v plug as it may need the power where modular 'RAILS' are totally limited to their rail and 'drop' from other power pulls

The Thermaltake has a regulaton of +/- 3% which I find quite high... PCP&C is 1% and it comes with a hard wired 6+8 pin PCIe for the newer cards without an adapter



Im not telling you its the PSU or getting another will solve the problem. I dont know what the problem is. If the system will run clocked at 1800 and test the memory clean in MEMTEST then I suspect its either PSU or another component failing when high power is asked for on ACPI initialazation (Windows boot)


You may need to do a process of elimination check to find out... motherboard is not out of the question either


gotta run

[NickN]

Vic

do you have windows set to display a blue screen on error?



Browse to CONTROL PANEL -
SYSTEM - ADVANCED TAB

Startup and Recovery, Settings button

UNCHECK Automatically Reboot and Send Administrative Alert

Click OK to close the box


that will allow a blue screen to appear without a automatic reboot


If this is reboot without a blue screen I would suspect defective parts or PSU

If there is a blue sceen the error message would be the next thing to look at

[vgbaron]

neg - always disable reboot on error.  BTW, just installed the Mushkin ram and setup the 400x9 that I had saved - which had worked - same reboot. So I've swapped memory and vid card with no success.

So the thought is:

is the PSU bad?
did the PSU just happen to go bad?
did the GTX 280 cause the PSU to go bad?

or is it something totally different.

[NickN]

... or is the motherboard bad


I have to run

[vgbaron]

Well - with my limited experience I'm beginning to bet on the mobo. Left 8800 and mushkin ram in - reset defaults in BIOS - so I'm running at 3.01 - reboot on prime95.

[NickN]

Its either mobo or PSU if your not pushing a clock and it does this after swapping out the primary components.

Although, it could be CPU too. Rare, but it has possibilities. You are sure the waterblock is seated correctly.. I would think so. Reseat the CPU in the socket and see if that changes the situation

PSU/MB diagnosis can be a real pain without experience with meters and knowing how to tie into it with probes during use

If you have no way of testing... you could go to a bestbuy and get a very high watt PSU and test it

if the problem is solved either keep the unit or take it back when you get another from online

[vgbaron]

I was going to need one for a second system and I live three blocks from Frys. Bought a Corsair 1000W PSU. Decide to RMA the mobo anyway. WHen all that is done, if it STILL doesn't work - gotta be the CPU.  :) If it is the CPU, that I have to eat since I lapped the bugger.

There was a time, many moons ago when I knew my way around a scope & logic probe but that skill has long gone. Easier to RMA - found out it has a flaky main power connector anyway. If I just touch the main connector it will shut down the mobo. Connector is seated properly - could be bad trace.

Anyway, I'm going to forget computers for an evening and relax.

Thanx!

[NickN]

Then you found the problem...

Make sure its not a bad molex pin or wire connection however that would explain everything you have seen.

Now, do be aware that traces and their solder joints can be damaged by high current demand and a low current PSU output which would lead back to the PSU if that were true. Not saying it is however now that you have found the culprit the reason is either a faulty manufactured board or that PSU was not delivering the current over and above what was needed

This happening after the 280 going in could be a coincidence but do keep that thought in mind

Sorry about the trouble right before Xmas but at least you found the issue and can get it resolved