Computer keeps shutting off.

[mindlogiczero]

didnt really know where to post.

so recently my computer started randomly turning off. i thought it was due to power supply which ive since replaced. seemed to have fixed it until about a week later. started happening again.

well i did recently install a newer cpu, 5600+ which everest reads temp at 35c, same reading for speedfan. while my 4000+ i had before read 40c. but core temp reads my 5600+ as 40-50c, but even if core temp was correct 40-60c wouldn't be enough temp for my cpu to tell bios "hey shut me down" right?

my system specs are as followed,

dell inspirion 531
windows xp 32bit
5600+ amd athlon 64 x2 2.9ghz
4x 1gig sticks of 667mhz ram
1 tb seagate hdd
9600 512mb galaxy gpu
pci tv tuner
Ultra Lifetime Series Professional LSP550 Power supply power supply.
g15 keyboard
3.0 mouse
front fan controller
120mm side case fan, 80mm front fan, 92mm back exhaust fan.


what it will do is power will knock off like someone just unplugged it, and it doesn't restart unless i push power button. it happens 1-3 times a day.
ive tried the memory test but not sure if i did it for long enough, i only did one scan while all 4 sticks were in, to my understanding that works just wont tell you which stick is bad? my one scan came back as 0 errors.

off topic. can i trust the everest readinf of my cpu running at 35c?
and my cache should be reading double this. http://img375.imageshack.us/img375/4154/13876463.jpg
and my ram should be reading 667mhz http://img44.imageshack.us/img44/876/22245091.jpg
pics taken of memtest.

 

[Rogue77777]

Did this happen with the old cpu too?
If not, I would make sure that your Dell MB supports that new cpu. Dell would rather have you buy a new unit instead of upgrading your old one.

As for your power supply I've used Ultra's too andsaid that it also happened with your old one, so I would rule out the psu.

And I really do not think that it memory related, because you're not getting ramdom restarts, although anytime you're using 4 sticks of ram you should manually set the ram voltage in the bios, if you're having issues.

 

[mindlogiczero]

i dont know how to set ram voltage nor think i can with dell pc?
dell supports the cpu i use. and no it didnt happen with old cpu. so would that mean cpu related or heat related?

 

[mindlogiczero]

bump

 

[Twoboxer]

- When in doubt, believe CoreTemp or RealTemp.
- Try using CPU-Z for cache stats and CPUID HardwareMonitor to check your voltages. If they are getting low while you're running, suspect the psu.
- Did that Dell come stock with 4x1GB? In any case, pull 2x1G out and see if it stops failing.

 

[mindlogiczero]

heres pictures ive taken from both of those.

http://img248.imageshack.us/img248/1112/72267552.png
http://img44.imageshack.us/img44/3096/55297629.png
http://img190.imageshack.us/img190/7206/12948328.png
http://img11.imageshack.us/img11/7010/81148981.png


the last two pictures show the two types of ram i use, slot 1 and 2 are same, and slot 3 and 4 are same. i bought computer from friend with the slot 3 and 4 type ram from a friend in like jan? i upgraded to 4gigs about feb? turn off prob started about 2 weeks ago, newest parts in computer are cpu and gpu. besides psu which is newest part but its only new cause i replaced that trying to fix this.

 

[Twoboxer]

Pull the last two sticks you added and try again. Will look at your pictures now.

 

[mindlogiczero]

it ran months on those new sticks fine just started weeks ago to do this. and problem might not happen for weeks / days, it rare.

 

[Twoboxer]

Did your cache report correctly?

 

[Twoboxer]

PLEASE just pull the memory. Its a test of a portion of the mobo having gone bad, its happened before.

 

[Twoboxer]

Your cpu temps look a little high to me. Probably because the Dell heatsink/coole/fan/shroud(?) were specifically sized. Or maybe a flawed install when you popped in the new cpu? They don't look high enough for trouble, but dunno what they do under load.

You might try a Prime95 run and watch the temps for 30 min or so.

 

[mindlogiczero]

yea cache read correctly. so you saying the fault could be the ram slots for the 1, and 2? and if i remove those 2 it will work if those 2 slots were bad? also if it cuts off again after that, what do you think could be cause?

 

[Twoboxer]

First, at this point I'm not sure *what* the problem is. It smells of psu, but yours is new and poweful enough. So, trying to leave that for last, I'm looking at problems *I've* had or heard of that we can eliminate easily.

We got temp info that can be trusted now, and cpu looks high but not deadly at idle, as I said above. Worth a Prime95 torture test to see if temps rise higher than we might expect under load. If they do, we prolly found it.

The memory thing comes from a couple of sources, including a Dell XPS-3 mobo I upgraded to 4 sticks, shortly after got failures like yours, shortly after lost the mobo. Maybe because I installed the two new sticks while the board was still hot and pressed down hard, maybe some other reason.

In any case, pulling two chips also eases the electrical load, which may or may not be an issue on your mobo. (I really don't much care which two, but would prefer pulling the last two you installed, unless you moved them all around when you installed)

Yes, I know it has run for months using 4 sticks. But its an easy test, and I have seen a problem before.

 

[mindlogiczero]

i think its a heat factor now to cpu. i ran that prime and soon as i started that temps jump to 61, within 10 secs at 64c and rising. unless thats common? that was showing on hwmonitor and everest was showing it fine. if you agree its cpu let me know and recommend a cheap heatsink i could use? right now im using stock heatsink/fan. i can only spend around $20-25. right now i got artic silver ceramique. and the method i used was bout a rice size in middle then i used plastic around my thumb to rub it in.

 

[mindlogiczero]

actually take that back. if its cpu i can spend alot more cause i can just return this new psu and use my old one if it wasnt problem.

unless you dont think my old psu is efficient?
http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=2932371&CatId=1079

 

[Twoboxer]

I am not familiar with AMD temps. On many Intel C2D, temps will rise into 50s, 60s really fast using a stock cooler. The last build I did, an E7400, wouldn't go over 52C running Prime95.

Before buying a new cooler, let's try a couple of other things:

1) What is the temperature in your room? If the room is warmer than normal, your cpu will run warmer than "normal".

2) Maybe there's a heat buildup inside your case, raising the ambient temperature more than you think. Remove your PC's side panel and try running Prime95 again. Keep running Prime95 until temperatures stabilize, but stop the test if temps exceed 73 degrees.

3) Take a read through this thread for some background:

http://www.tomshardware.com/forum/240448-28-5600-heat-issues

4) If we're still running "too hot" after #2, we'll try re-seating the heat sink.

 

[Twoboxer]

Error

 

[mindlogiczero]

heres after like 20 minutes of running.
http://img189.imageshack.us/img189/7636/test2g.png

peaked out at 73 and raised alot slower then it was doing, but like i said cut off didnt start untill like a week after i upgraded psu, when i upgraded i redid thermal paste on cpu, so maybe its just doing better right now cause i re did? i tried the lines, thenpulled it to see how well it spreaded. didnt spread very good was just glob in middl, this type of artoc silver is very sticky and hard to spread. could the cpu be damaged? and do you think this is reason for shut offs? going to tiger direct 2morrow, was going to have them do thermal paste for me and prolly pick up heatsink. if you agree this is prob could you suggest thermal paste/heatsink from their site? thanks.

 

[mindlogiczero]

wow tigerdirect said they will charge $30 to install the heatsink wtf.

 

[Twoboxer]

Sorry, gkid's baseball game, he won . . . anyhow . . . we gotta get it cooler . . . but I'm gonna suggest just run it for a while and see if the improvements you've obviously made lead to no - or - far - fewer shutdowns.

Far less heat should be generated when running anything other than Prime95.

Just keep an eye on your temps so that you "know" the temps a few minutes before the system shuts down . . . if it does.

 

[mindlogiczero]

oh so you saying you dont think its heat related since i shouldnt even get those 70+c temps in anything i may have used besides prime95?

right now idle is 40c core 0, 47c core 1.

one thing i did do when i redid paste this time was clean alot of extra paste from sides. i had tons of paste on all 4 sides, about equal to what shoulda been on the top of cpu itself. i did that the first time i installed it by misstake, dunno if that could cause it or not. could this be from what i use to clean cpu/heatsink with? i use bd alcohol swaps 70% isopropyl alcohol. and i basically do the paste on cpu as this guy does. http://www.youtube.com/watch?v=K2-k9rvNnsI .

grats on their win.

 

[Twoboxer]

LOL thanks! If you can see paste after installing the heat sink, that's too much. Alcohol swabs are fine for cleaning. Yes, too much paste can insulate, its only supposed to fill in the gaps where the heat sink doesn't touch the cpu.

I thought your slower-rising/lower temps meant we did a better job on the heat sink this time. Its safe enough to try to run under normal loads now and let you watch the temps. Since we have "lessened" critical high temp problems, and since you will be watching them, I'd like to see the system run now.

If it fails, you will *know* its temp related because you will have seen a real high temp reading. Assuming the temps stay under 70, that's certainly not hot enough to cause shutdowns. So if it fails, its NOT cpu-heat related. If the temps go up over 73 too frequently during normal running, then we need to get a cooler.

I'd rather not spend your money unless we have to and until we know what the problem(s) really are.

 

[mindlogiczero]

well im not really spending money, im just returning that psu i had bought, and tigerdirects return policy is 15% restocking fee i5 merch is opened unless you are trading for somthing in the store, so might aswell spend in store. regardless if the heat is prob or not id like to keep cpu cooler then it is. and about each time ive reset the cpu with paste its taken about a week untill it started to fail, so if it is the paste and starts after a week, it could maybe just mean my paste is bad? i could see me doing the paste wrong the first time but these 2 last times i feel like i had it on there good. one thing i did notice was when i tried your line method then pulled to see how well it spreaded, it was so tight to the cpu i had to actually give it force, but when i do my way it comes right up when i un brackets, when it was hard to come up does that mean it was suppost to be like that and if it comes off real easy when unlocked its bad?



im looking at these heatsinks,
http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=3071635&CatId=493#
http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=4244282&CatId=493
http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=4420236&CatId=493

i know they are basically overkill for my cpu but i thought it'd be cool that i could use with a i7 when i build a new computer whenever i have the money. just trying to figure if they fit my case, http://i.t.com.com/i/lumiere/2007/06/26/32486097-320-0-28101-20070626_160345-320x240.jpg
http://img224.imageshack.us/img224/9817/133400sideopenln1.jpg
http://www.redplanettrading.com/ebay/motherboards/ry206/ry206-top-300.jpg

 

[mindlogiczero]

just re did and tryed the lines and let heatstink spread, 39c core 0, 48c core 1 idle.

ran test and it got to 73c at beggining of test 2.

everest reads core 1 at 32c, core 2 at 26c lol.

 

[Twoboxer]

Still too hot lol. Get new paste with that cooler lol.

1) Of the three you listed, the CoolerMaster appears best, at least in this review where large numbers of coolers have been tested in the same way:

http://www.frostytech.com/articleview.cfm?articleid=2314&page=4

2) Will it fit? Two ways to check. This photo (and others in the review perhaps) shows the cooler height vs expansion card slots, and overall size:

http://www.legitreviews.com/article/833/3/

The other way, measure how much headroom you have between cpu and the case sides/edges and compare to spec heights here:

http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=4420236&csid=ITD&body=MAIN#detailspecs

3) Re: your questions about the paste - I've never used anything that sounds like yours does, so I really don't know. Ideally, if the heat sink and cpu were pefrectly flat, I suppose you wouldn't need thermal paste. The paste's job is to fill in the gaps, some as small as a "molecule" of paste. So the idea is to pressure the two surfaces together, paste where they do not meet. Having a thick layer of paste and *maybe* not enough pressure can insulate the two parts rather than enhance heat conductivity.