Sign in with
Sign up | Sign in
Your question

Is this considered a symptom of a weak PSU?

Last response: in Components
Share
June 12, 2008 6:09:20 AM

Hey guys, I would appreciate a point in the right direction with this situation. Looked around and can't seem to find a topic that matches this to my liking.

It's been a long time since I've played games, and now that I come back to it I think I remember why I gave up: I seem to have a potential problem with voltage/power supply... but I'm not sure.

It only happens with games, so far I notice it with more graphic-intensive ones (Crysis, BioShock, COD4, etc). This can happen at any time, right after a finished load screen or 5 minutes to 5 hours into play... I have deduced that it generally hits on 'sudden' moments that requires more/quick processing (imagine a grenade blowing up in your face in COD4 = death). The result is a hard shutdown (no BSOD, no error message) and automatic reboot... Reminiscent of trying a faulty OC choice: system powers up, falters and shuts down quickly, reboots and restores old BIOS setup. Is this overheating, a weak power supply, failing parts? Or...?

TAT and coretemp show 35-40 C idle, no more than 50-55 C under load.

Specs:

Antec 180B case
Intel E6600 (Model BX80557E6600) - All stock
Samsung HD501LJ
Kingston (2x1gb) DDR2 800 (KVR800D2K2/2GR) - Stock
Gigabyte GA-P35-DS3R F11e rev 1.1
EVGA GeForce 8800 GTS 320 MB PCI-e - Stock 513/792
Antec SmartPower 2.0 SP-500w

In the interest of the PSU unit: http://www.antec.com/specs/SP500_spe.html

I can't offer much more, such as past experience with this setup. I've had it for some time but have not used it for games so much.

Any help appreciated.
June 12, 2008 6:31:27 AM

It's possible.

Here's a few things you may want to try:

Use a driver cleaner program and reinstall latest (possibly beta) drivers. I had some locking up and rebooting issues with my 640MB 8800GTS recently, but when i upgraded to the latest beta drivers it went away.

Here are a few driver cleaner programs i found with a quick search, there may be better ones out there:

http://www.guru3d.com/index.php?page=driversweeper
http://downloads.guru3d.com/Driver-Cleaner-Professional...

You may want to update your BIOS (as long as your sure your system won't power off in the middle of an update!)

See if memtest86+ gives you errors: http://www.memtest.org/#downiso

See if Orthos gives you errors: http://sp2004.fre3.com/beta/beta2.htm

You could buy a power supply tester and see what it tells you, although I've never used one and I'm not sure how helpful it will be in this situation.

Hope something here helps you!
June 12, 2008 6:51:57 AM

If you had that windows up and running for a long time, i suggest a format and using the drivers Pshrk recommended, also update your mobo's BIOS to the latest. That should remove any software related problems, but if format a no-go, follow Pshrk's instructions.
Related resources
June 12, 2008 7:04:22 AM

Agh, it's been so long I even forgot the basics !

Was running 175.16 for the vidcard, might have been fuddled with an old driver -- ran cleaner and reinstalled.

Have memtest and Orthos already available to run, I'll run the torture test tonight and see what comes of it.

F11e is the latest revision for BIOS, no worries with random shutdowns, the PC is generally stable.

Something interesting just happened after the drivers re-install: sitting around, doing nothing, playing .avi on VLC player -- the video output to my monitor just shuts off. Everything is still running, keys are responsive, I can pause/play on the keyboard, but no video. This is the first time this has happened. Nothing out of ordinary with my components that I can see...

It's quite unfortunate, but I think I've had this XP for less than a year.

Hmm.. plot thickens.
June 12, 2008 7:25:15 AM

Hmm... never had that problem before so I'm not sure :-/ This happend after you reinstalled the drivers and rebooted the computer?

It could just be a coincidence, but if that problem is related to the rebooting problem then it makes me think it's something with either your power supply or video card, most likely video card, but like i said i'm not sure. Hopefully it was just a freak coincidence.

Here's the only other things i could think of but you've probably already checked these out:

Make sure the cables and especially the power connectors are in fully and making good contact.

Make sure the fan on your video card is working and check the video card temps. I'm not sure of the best program to do this, but i think the NvidiaMonitor program you can download from nvidias website will tell you this info. I like gaming with dual monitors so i can view temps, usage, and other info on my second monitor.

If none of that helps maybe try the 169 Forceware drivers.
June 12, 2008 7:32:20 AM

Yeah, I'm thinking it's just a coincidence as well.

All physical operation and connectivity looks OK.

NVMonitor reads 60 C idle -- I briefly looked up standard temps and this seems appropriate.

I'm going to try out some older games, like HL2 to see if I get the same results.

If that doesn't work, after some other testing, I'll mess with the 169.

Thanks for the prompt replies!
June 12, 2008 7:37:28 AM

Good luck! Let us know what happens.
a c 144 ) Power supply
June 12, 2008 8:06:47 AM

predtetm said:
It's been a long time since I've played games, and now that I come back to it I think I remember why I gave up: I seem to have a potential problem with voltage/power supply... but I'm not sure.

It only happens with games, so far I notice it with more graphic-intensive ones ...
The result is a hard shutdown (no BSOD, no error message) and automatic reboot... Is this overheating, a weak power supply, failing parts? Or...?


My guess, because temperatures seem to be OK, would be video driver or PSU problems. Try new drivers first - doesn't cost anything and you do not need to open the box. Then try running Prime95 for 24 hours to eliminate any other stability problems. I've had Prime95 fail at 16 and 18 hours before during stability testing, so that's why I recommend 24 hours.

Symptoms do match a PSU problem. If so, here's what's happening: PSU detects some kind of an internal fault. A control signal to the motherboard called "PowerOK" (main power connector, pin 8, grey wire) goes to a "Logic LOW" (0 volts). This forces a hardware CPU reset. The load on the PSU drops, it resets itself, ouput voltages come up to specifications, and PowerOK goes HIGH (5 volts). CPU sees this and the system reboots.

The PSU testers are quite handy for testing a suspected DEAD PSU or a suspected bad cable. But they only test under a minimal load condition.

I have a similar system: EP35-DS3P motherboard, Q6600 OC'd to 3.6 MHz, 2 GB RAM, 640 MB 8800GTS, three hard drives, DVD writer, Antec 650 watt TP3, Antec 900 case with fans on Low. One day I was curious to see where the power went, so I borrowed a lab calibrated clampon ammeter from work.

Here is what I measured:
Q6600 OC'd, Prime95 load: 10 amps (120 watts)
8800GTS running 3dMark: 5 amps PCIe connector (60 watts)
running desktop: 3.5 amps
MB: 1 amp @ 12 v, 3 amps @ 5 v, 3 amps @ 3.3 v for the RAM (40 watts)
Really surprised here. Thought it would be more.
The 4 drives (calculated): 5 amps @ 12 v, 5 amps at 5 v (about 90 watts)
Fans (calculated): 1 amp (12 watts)

That's about 320 - 330 watts total, 260 - 270 watts (22 amps) @ 12 volts on my system. A P35C may need a little more power for the MB and RAM, but the E6600 will certainly need less power than an OC'd Q6600.

So your PSU has ample capacity. Something inside the PSU may be intermittently shutting down causing the reset/reboot. About the only thing you can do is test the PSU by substitution. Try to borrow an adequate PSU or get a new one. A good choice would be a Corsair 450VX. They are very conservatively rated.

June 12, 2008 8:22:26 AM

Nice post jsc.

If you don't have another power supply available you could try swapping out other parts such as the video card, if you have another video card available. It may be worth a try to boot the board with 1 stick of memory as well.
June 12, 2008 8:27:51 AM

I have a Q6600 @3.0GHz, 8800GTS 512MB (which needs more power than the 320MB version), 4GB Memory, 2 HDDs, 1 DVD Writer, 1 Combo Drive, 2 Fans, and using the 175.16 Driver. all of this is running on a 480W Generic PSU. The system is running great. with my old 8600GTS card, i used to get some restarts, and BSODs. But all of that went away after doing a fresh install of windows. Usually I do a fresh install every 4-6 months because I install/uninstall alot of programs/Games.

So I think 500W is plenty of Power, unless your PSU is dieing. personally I would do a fresh windows install first before deciding to buy a new PSU.

Hope this helps.

Regards.
a c 144 ) Power supply
June 12, 2008 11:31:36 AM

pshrk said:
Nice post jsc.

Thanks.

predtetm:
Here's a suggestion if you want to try a fresh Windows install, especially if you have a lot of software installed on your boot drive. Use another drive. Nearly everyone who has been doing this for awhile has an extra drive lying around. Install Windows, firewall and AV of choice (you will need to go online for OS updates at least), the drivers you need, and only your problem apps games. Test Play. If the same thing happens, it's not your OS installation and you have not disturbed your main drive.

And a suggestion for everyone who periodically reinstalls Windows (a little off topic, but the subject did come up) ...
Install the OS, go online and get the OS updates you need, install your standard utilities and what I would call your "base" software - your standard applications that you would always install. Use a backup program to make an image of your system drive (should fit on a SL DVD). I do not trust DL DVD's for software or data backups.

When it's time for a reinstall, backup the data you want to keep, and reinstall the image. Then go from there. Saves a lot of time.


websurfer wrote:
"I have a Q6600 @3.0GHz, 8800GTS 512MB (which needs more power than the 320MB version) ..."

Assuming the cards are running at stock nVidia clockrates, that's not true. Power consumption for the G80 320 MB, G80 640 MB, and G92 512 MB 8800GTS's is around 100 - 110 watts with the G92 using the least. nVidia gained a little efficiency over the G80 when they designed the G92 core.




June 12, 2008 1:31:53 PM

jsc said:

Assuming the cards are running at stock nVidia clockrates, that's not true. Power consumption for the G80 320 MB, G80 640 MB, and G92 512 MB 8800GTS's is around 100 - 110 watts with the G92 using the least. nVidia gained a little efficiency over the G80 when they designed the G92 core.


My bad :) , I got things mixed up :pt1cable: . But my point still holds, because my system is running fine on a generic 480W PSU.

Regards.
June 12, 2008 1:32:28 PM

I'm going with PSU. (Had this problem before and I seem to recall replacing the PSU fixed it.)

My thinking is that your total wattage on the 12v rail is 480, which should be enough but at 70% efficiency, it's only 336 which i think is cutting it close for your system under strain. I believe in overkill when it comes to PSUs because more is always better than not enough. Whatever you do I would certainly go with 80% efficiency or better.

Hope this helps and good luck.
June 12, 2008 3:54:08 PM

OOps. Sorry mixing my efficiency with another calculation so disregard but I still think the problem is in the PSU from your symptom.
a c 144 ) Power supply
June 12, 2008 10:01:26 PM

AM71,
I doubt if predtetm's system is pulling more than 20 - 22 amps off the 12 volt rails. His problem is not inadequate power. It's unstable power. If it were inadequate power, his system would crash immediately every time he loaded a 3D game.

Websurfer,
I wasn't going to address the generic vs. brand name PSU issue, because so much has already been written about it. A lot of people are successfully running generic PSU's (including me - sort of).

I have Antec 550 and 650 watt PSU's (third tier) for my two new boxes. Got them on sale at really good prices. My "workhorse" PC (AMD Barton core CPU, FX5700) has a generic PSU in it. The "sort of" part comes in because it's pretty much on all the time. I need to replace the PSU every 18 to 24 months because the sleeve bearing fans wear out.

I am in Saudi Arabia where ThermalTake is a top brand, Raidmax is an intermediate brand, and generics go for an equivalent of $12 to $15. If I want a good PSU, I need to bring it with me when I return from vacation.

I will not trust a new build to less than a third tier PSU. I also try not operate a PSU at more than a 70% load. But that's just me. :) 

!