Computer shuts down

I suspect my power supply isn't up to the task and I'm getting somewhat random shutdowns.
This is my system http://www.tigerdirect.com/applicat...mc=Email-_-WebletMain-_-WEBLET03ORDER-_-Deals
To that I added:
http://www.msi.com/product/vga/N8400GS-MD1GD3H-LP.html
http://store.hauppauge.com/hardware2.asp?product=hvr1600
and a used PCI SCSI card to run my scanners.
I also have a trackball mouse and keyboard (just stating the obvious).

At first I thought I was having an overheating issue. I went so far as to take the side off the case and blow a fan at the CPU and GPU. The GPU gets close to 60 deg. Celcius, but the shutdowns don't seem to be related to the temps there. The CPU temp stays well lower than that according to CoreTemp.

Is a 450 watt PSU up to my needs? Is there a way to test or check on it?
Thanks in advance
Dan

 

PS, I also plan on upgrading my video card eventually to something faster, and that will likely draw more wattage. So if this PSU's specs are marginal I'll probably replace it anyways.

 

Your system will only draw about 180w under full load so your power supply should be fine if it's not faulty, I use a wattage meter to test the actual power draw through the socket for diagnostics which are only cheap but if your power supply is over loaded it will become hot to the touch as a simple test,you can't really test the stability of a power supply under load without expensive equipment and a good working knowledge of electronics.

If it is a load or heat stability problem the best way to test is to put the computer under load and see if it crashes.

Monitor your temperatures and voltages with speedfan,make sure your cpu fan spins up with temperature.

http://www.majorgeeks.com/SpeedFan_d337.html

Here's what run,intel burn test to test your cpu don't worry you can test amd cpu's as well. Set stress to maximum then hit start,let it run for an hour,check you gflop score roughly matches what how your cpu should perform can't find a gflop score for your cpu ATM just google it.

http://majorgeeks.com/IntelBurnTest_d5987.html

3d mark vantage to test DX10 stability,loop it 2-3 times and watch for video artifacts such as flickering textures,check your score matches how your card should perform.

http://www.majorgeeks.com/3DMark_Vantage_d5948.html

Your card should score about 450 and your cpu about 5000 with the benchmark in performance mode.

http://www.3dmark.com/3dmv/3679218

Run memtest to test the stability of your memory.

http://www.majorgeeks.com/MemTest_d350.html

Go to control panel,system,advanced system settings,startup and recovery settings,set for small memory dump and uncheck automatically restart this is important as you'll get a mini dump which can help diagnose the problem.

Can you describe the shutdown in more detail,does the computer lock up,does it go through the shutdown sequence,doe it totally switch off instantly,does it happen if you leave it on the desktop for a prolonged period?

Post your results and temperatures,your gpu temp is fine they can run much hotter than cpu's in general.

 

I wish I knew what this meant. The help file for SpeedFan doesn't really say what temps are good, what's bad, etc.... So I run it and this is the first thing I get. Should I be concerned with the flame symbols?

 

ItelBurnTest results (Success):

Ran for close to 3 hours at a stress level of Maximum (10637.02 seconds total). I turned off Folding At Home and other nonessentials to give the program more than 7 of my 8 gig of ram. I ran the test overnight and didn't watch it. I have been watching it under stress loads otherwise and the only flame icons that appear are the GPU (which I sometimes put a fan on if I watch TV through it etc...) and those other two, Temp3 @ 127 and HD0 @ 128. Those two numbers never change.

Will run 3DMark Vantage next.

More info on my system:
Mobo: http://www.ecs.com.tw/ECSWebSite/Pr...?DetailID=1139&CategoryID=1&MenuID=19&LanID=0
CPU: http://products.amd.com/en-us/Deskt...2=&f3=&f4=&f5=&f6=&f7=&f8=&f9=&f10=&f11=&f12=
I forgot to mention I added a Solid State Hard drive but this behavior was present before adding this (and partially why I wanted a SSD). http://www.sandisk.com/products/ssd/sata/extreme/ 120gb

 

Your temps are fine,temp 3 and hard drive temp are both anomalous readings and known bugs and don't worry about the flames.

No need to put a fan on your gpu they are designed to run that hot,hotter than cpu's up to 100C. Just make sure the heatsink is clear with canned air and the fan is spinning,watching TV or movie's put very limited stress on your gpu anyway usually as the video is decoded by the cpu.

No restarts during intel burn and regular cpu temperature,you can conclude your cpu is fine this also with 90% certainty means your ram is stable also.When you run intel burn test though watch which temp goes up in speedfan 'this is your cpu temp' then post it's maximum temp after 30 mins to an hour with your next post,max cpu temp should be under 70C.

When testing your gpu in 3dmark you can watch and save a chart in speedfan alternatively use afterburner which easily logs your gpu temperature and usage but do spend a little time looking for artefacts 'graphical glitches' which will let you know if you gpu is unstable.

http://majorgeeks.com/MSI_Afterburner_d6254.html

EDIT you haven't described the shutdown yet either?

 

Running IntelBurnTest for half an hour makes Temp1 (the first one on the list) go to 56 but no higher. It will normalize between 52 and 56.

Right, so here's what's going on. I'm usually working along just fine and the computer just shuts down. No BSOD, no dump files, it's as if you unplugged the PC, straight from running fine to not running at all. I've only noticed this when the CPU is above 90%, but it's also happened when I've been away with FAH running at 50%, so CPU usage shouldn't have been 100%. The two main programs that I can think of that contribute to the shutdown are Canon's Digital Photo Professional and Vuescan. Both graphic processing programs. Of course it may not be related to this in the least and I might be jumping to conclusions.

I've downloaded WhoCrashed and it's report is:

I've tried following the links it gives but most results don't seem to fit my situation. It should also be noted that the latest crash dump file was generated over 2 months ago, and my shutdowns are much more recent.

Also the Radeon driver is for another video card that I tried for a while that's no longer in the computer. I tried this card for a while, but it caused crashes, so I reverted to my older card which is what I have now.

 

Yeah an instant off is most often a power supply over current problem or just plain faulty.

Temperatures are low,no performance related symptoms,no mini dump or BSOD,memory stable in intel burn,cpu stable.

These all fit a power supply problem,usually if the power supply is in an over current state it will be hot so this would be another symptom to look out for.

I'd still run 3dmark and mem test since there's nothing to lose,3dmark will draw the most current using cpu and vid card so it may shutdown during 3dmark but yeah at this point I agree with your original prognosis of power supply.

One thing I've failed to mention is the possibility of power failure at the AC socket which can cause the exact same symptom,make sure you have a good quality surge protector,if you suffer lots of power outages in your area or frequent bad weather consider a ups.

The shutdowns would most likely be infrequent though only two or three per year.

 

Memtest didn't come up with any errors. I've used that before. I've also used the Microsoft Memory Diagnostics (during POST) which came up empty too.

Tried the 3D Mark Vantage. It lags as bad as I thought it would and even wouldn't finish the test. My graphics card only cost $35 so I wasn't really expecting any real 3D performance. I was averaging 2-3 fps for 3D Mark, and when I use the graphics card program Kombuster (a stress tester) I usually get 7-10 fps but that doesn't run full screen. Anyways, the 3D Mark Vantage test gets to the fourth test and comes up with the error (see screen shot).

I'm doing all of this and no shut downs. So I tried using Canon DPP again and it shut down. This time there was a dump file generated.

My power supply is not hot to the touch. I do vacuum out the dust bunnies.
I have a battery backup power supply. It works, there's also an alarm on it that sounds during power outages and I'm not hearing that happen.

I went to Control Panel> Event Viewer and there's 3500 or so entries for this error during the last crash:

Although I've been watching the temps and nothing's been spiking this suggests there's a heat issue somewhere. But the fact that it's happening with one or two programs and not others suggests software. I've reinstalled all software that I suspect might be causing the problem but it persists.

 

Well at least we have an error to attack now,add the dump files as attachments with your next post,I doubt we can glean anything off them though,have you debugged them yourself?

ACPI 'advanced configuration and power interface' bios temp and power monitor 'thrm' thermal monitor, I think this is the cpu high temperature trip?t could be the video card temperature trip also which would make more sense. Two ways to go about this.

Go into bios 'del as computer is booting' go into power features and disable acpi completely or go into cpu temperature and raise the cpu alarm temperature,cpu alarm temperature or disable them completely,check for a system temperature alarm and raise that also or disable,post what temps they are set for.

As a basic test run one of the benchmarks,you can use 3dmark 2005 also 'see how far it gets' if you wish and feel the heatsink of your video card it should be quite warm to the touch but you should be able to hold your finger on it.

Have you updated to the latest video card driver? Have you also updated all your motherboard drivers? long shot but just checking.

You mentioned blowing a fan on the card does that stop it crashing? Idealy the the way to test for this is to run 3dmark 2005 say until it crashes then blow it with a fan or take the side of the case off see if it lasts longer before crashing.

Will check back tomorrow

 

I tried turning ACPI off and the computer wouldn't boot to Windows. Said I needed to put the Win install disk in. Decided to restart & revert bios to factory defaults. now keyboard not work. Using onscreen keyboard now. will report back when I get things sorted.

 

OK, update. Restarting the computer cured my keyboard woes.

I left the computer off overnight. When I turned it on in the morning the CPU fan was making a horrible (to me) noise that I hadn't heard it do before. Long story short I think I'm going to replace the heatsink and fan, or maybe just the fan. In the mean time I took the heatsink off, took the CPU out, cleaned and reapplied thermal paste, put it all back together and it seems to be working. I can run the two programs that were giving me shutdowns without problems so far.

It took me a while to get back to you to give the computer time to crash or not. I think I've thoroughly tortured it and it's still plugging away.

Thanks