more computer woes - freezing/restarting

C

cacophobia

Veteran XX
Hey all, I got this new computer built like a few months ago. It was freezing and restarting a bit at the beginning, but it's been getting worse and worse over time and now I can barely run more than a couple of harmless programs without it freezing.

Here are my specs:

2.13ghz core 2 duo
asus P5N-E SLI motherboard
geforce 7600 gt
2gb ocz dual channel pc6400 ram
one old ide 80gb hard drive
one new 220gb sata hard drive
power supply is like 415 watts or something, I can't remember exactly - I had to buy a new one over the one the case came with

just a few points on its behavior:

- it seems to do much better when computer hasn't been on for awhile
- when I'm playing tribes it usually doesn't take very long to crash. I generally can't get through a whole map without it crashing, and at times it will be crashing after like 3-4 minutes
- used to be when I loaded up tribes the graphics fan kicked up into overdrive (I think it's the graphics fan at least), and then would settle down after a bit. once it had settled down it seemed more prone to crashing, so I would just minimize and re-maximize tribes every few minutes to get the fan rolling again. this used to make a big difference. now it's getting to the point where it hardly matters if I leave the fan in normal mode or minimize/maximize to get it into hyper mode.
- I don't think it's a temperature issue, because what kind of brand-new stock stuff would have temperature problems right out of the box? Also, when my computer restarts, I will jump right into the BIOS and the cpu and mobo temps are both at a steady 40c.
- It does a lot worse when there are more programs running

oh well, if I think of anything else I'll add it. as always any help is MUCH appreciated, thanks!
 
ok, also, apparently I have a gpu temperature monitor on my computer and it says its like 46c and it just crashed, so I don't think that's the problem.

one other thing, a couple of days ago it froze and I left it there for like ~3-4 minutes while I was copying something down, and when I tried to boot it back up the motherboard screen came on, and then this little underscore was blinking in the top left and it wouldn't boot up. the underscore would just keep blinking. it stayed like that for about a day and then i turned it on and it worked again.
 
really sounds like an overheating problem from what you're describing. get a prgm to monitor cpu + gpu temps and watch to see if they spike or something.

another thing to try is running memtest86 (google it) on your memory for a few hours. RAM problems can be hard to diagnose cuz they cause all kinds of strange problems. if it runs for 3hrs on memtest86 with ZERO errors, we can rule this out. even 1 error means your RAM is bad.

try that out
 
It could still be a temp issue if your case doesn't have very good ventillation (bad airflow/shitty fans). Take off the side of the case and put a large fan next to it, see if that helps. Software sensors lie.

memtest is always a good idea in a situation like this

I'm also thinking it might be your old IDE drive crapping out on you
 
I'd say overheating too.

The sata drives run a lot hotter than the pata(ide) drives do. Can you get some extra fans?
 
does it crash/restart at ANY time or only in games?

open the case and direct a house fan into it and do your thing and see what happens
 
it's more susceptible to restarting in game, but if i'm running 4-5 regular things at once it will usually shit out then too.

vid card is not overclocked (to my knowledge)

side of my case is already off and it hasn't seemed to help much... and i don't own a house fan or really even have room for one in here. i guess maybe i can go buy one today and force it in here, but if the bios readings are fine i'm not sure i see the reason in that
 
cacophobia said:
side of my case is already off and it hasn't seemed to help much... and i don't own a house fan or really even have room for one in here. i guess maybe i can go buy one today and force it in here, but if the bios readings are fine i'm not sure i see the reason in that
Click to expand...
The side simply being off isn't much help, you need airflow to take the hot air away.


Also, my mobo reports a constant 10-11 °C case temperature in a room that's never colder than 20 °C. So much for software readings.
 
You've got blown capacitors on the motherboard and/or a blown powersupply. Visually check the motherboard.

OCZ has been known to frizz out on people, so memtest it. But I'm 99% sure that isn't your problem.

The giveaway is the blinking in the bios and the time it takes to cool off, and the fact it crashes more often during peak usage; your system has been overheating and you're slowly driving it into the ground. The real giveaway is that it's getting worse; generally nothing on a motherboard or powersupply slowly dies accept for capacitors and maybe transformers but usually when a transformer goes you've got a blown fuse and smoke.

Check your voltages in the bios and make sure they're consistent. As caps fail in a powersupply, voltages fluctuate more due the lessened ability to buffer voltage. A .3-.5 or greater fluctuation during constant load indicates the powersupply is on it's last legs.

The other thing you can do is boot it into windows, download and install hwinfo or any other hardware monitoring utility and monitor the system temperature and voltage as you run an app that uses a ton of processor and memory such as a benchmarking utility. That will give you an idea of what's dropping where. If voltage drops off slowly, the powersupply is gone. If it overheats and shuts down, then you may try installing a proper cooler kit on your memory and motherboard.

For the record, I have a Athlon XP palmino 1800+, it'll run fine up until 65C, then it'll crash. The motherboard will run find up until about 35C then crash. I have 13 fans (4 in the powersupply, 1 CPU, 1 motherboard, 5 case, 1 GPU and 1 PCI Cycloneblower under the GPU) in my case, and the thing kept crashing until I installed a cooler kit on it. Most processors are fine up until 55-65C, most other chips are fine until right around 35-40C.

Also, what brand/model of powersupply?
 
Last edited:
hey Gwok,

it's an Antec Neo480: Antec.com - NeoPower Series

as for the voltages fluctuating, will it be something immediate or should i monitor them over a couple of days? I've been using speedfan and so far it hasn't fluctuated more than about .05V

also, what should I be looking for on the motherboard?

also, thankyouthankyouthankyou (*if* this works! ;)
 
gwok, 13 fans!? what in the fuuuuuuck, that's crazy. i have 5 fans total: GPU, CPU, front intake, rear exhaust, power supply
 
cacophobia said:
hey Gwok,

it's an Antec Neo480: Antec.com - NeoPower Series

as for the voltages fluctuating, will it be something immediate or should i monitor them over a couple of days? I've been using speedfan and so far it hasn't fluctuated more than about .05V

also, what should I be looking for on the motherboard?

also, thankyouthankyouthankyou (*if* this works! ;)
Click to expand...

Blown and Leaking Motherboard Capacitors - A Serious Problem? - PCSTATS.com

Antecs are ok, although I don't like their neopower line. This is the guy I'v got: maddog 550 surepower. It's an old beater but she's lasted me 2 years.

Electrical draw increases during high-intensity activities; a GPU may only draw 10watts during normal usage but then jump to over 100 during an intense game. Same goes for cooling; my palmino can run at 50C during normal operation then jump to 62C during a game.

A drop of upto 5% voltage on a line is acceptable, some systems run at upto a 10% drop. So if the 12v line is running .6v or 1.2v under or over voltage, or if the 5v line is running .5 or 1v over or under then you should be fine; over those and you may have problems. Most hardware runs fine till right around 7-8% but remember the sensing equipment on the motherboard has a 1-2% tolerance itself and isn't always exact. Hell, your average multimeter is often off by 5%. A normal variance of upto .1v during constant low or high load is also acceptable. After that you usually start having issues. The main thing to look for in troubleshooting a power supply is when a lot of random errors happen or when the system crashes consistently under high load.

As I said before, do a stress test with benchmarking utility and watch for changes in heat and voltage. And run memtestx86 on it; google, download, and run. 3 passes, extended.
 
Last edited:
Gwok, I think he knows what a PCI slot fan is. He's just making fun of you for using one. Personally, I wouldn't use one either except in specific configurations. You seem like a knowledgable computer guy so I'm assuming you had a reason for putting one in. It is sort of weird that you have 13 fans though... I never had that many, even on my old computer.
 
iNVAR said:
Gwok, I think he knows what a PCI slot fan is. He's just making fun of you for using one. Personally, I wouldn't use one either except in specific configurations. You seem like a knowledgable computer guy so I'm assuming you had a reason for putting one in. It is sort of weird that you have 13 fans though... I never had that many, even on my old computer.
Click to expand...

why do I have 13 fans?

I just rebuilt my machine 3 weeks ago. Prior to that, I had a box running for 5 years. I had to lube 6 fans that died, replace a powersupply, and a failed soundcard in that time. For that long of a run, that's pretty damn good.

What a lot of people fail to understand is when any chip, be it a sound chip, UART, IDE controller, southbridge or north bridge especially, gets hot, it does 2 things before utter failure; it begins having processing errors, and begins throwing noise onto the system bus. This is not because the chip itself is undergoing rapid state change but because the internal expansion of the electric lines within the chips themselves allows voltage to occasionally leak or cause crosstalk. On some systems this is less of a problem and on others this is a big issue, and the only explanation I've seemed to find is that not every chip or production process is equal. Some processors can be overclocked to over 5 or 6 GHZ, others die out if you clock them normally. Same thing goes for northbridges, southbridges, IDE controllers, ect.

System busses are designed with fail-safes so if an IDE controller begins throwing off jibberish for a few seconds because it flipped a bit it won't crash or lock up the machine. Instead, what'll happen is that the thing will run slow and buggy; you'll click on my computer and find out your drives take 2 minutes to detect, for example. Or if the northbridge gets hot it's buffer might flip bits and then you've got a crashed game. Some of the newer boards even have the ability to reset interfaces on the fly.

After awhile, it does lead die warping on the chips themselves causing them to be perminantly buggy until eventually they die. But at those temperatures it's far more common for voltage regulators or especially electrolytic capacitors to fail over time. Electrolytic capacitors have stable operating temperatures up to 85C+, but just like a processor that has a maximum operating temperature of 90C, it isn't a good idea to run it anywhere near that. A north bridge running at 40C can and will crash like crazy.

Did I mention I've done 20 hour gaming spurts on my machine in games like Far Cry and never have crashes? Lots of machines shit out after 2 or 3 hours.

In the instance of the PCI blower fan, A: I have 6 PCI slots and no intention to use them all and B, what happens if my AGP card fan blows? That's right, it acts as a backup. Plus it isn't blowing air up to the processor intake or onto other cards. It isn't going to work for long but at least I get crashes instead of losing the card.

Mostly, though, because I can.
 
I've played Far Cry for several hours straight with no crashes, and I have a grand total of 5 fans in the comp (PSU, vid card, CPU, 2 case), all of them mid-low RPM. Good for you? :shrug:
 
I understand what you're talking about Gwok. Basically, heat is electronic components' enemy, but gwok, I'm sure I don't need to tell you this, but quality > quantity. Having 13 fans in your case is likely to create an airflow short circuit of some sort. I mean come on, 4 for the power supply? Honestly, that's completely unnecessary. Having 4 fans IN your power supply EACH pulling let's say... 40CFM isn't really going to make a difference than having ONE fan in there pulling 40CFM because they're all crammed into that tiny little box and therefore the bottleneck is whatever the airflow of ONE fan is. Your computer must be loud as shit too. Why don't you get new fans that don't suck? I had 80mm Panaflo fans for 6-7 years that NEVER seized and NEVER required lubing and they pretty much run as quietly as ever. But that was on my old computer (AMD Barton 2800+ w/Radeon 9700 Pro)

My current computer is this:
Core 2 Duo E6600
2GB DDR2 667
Radeon X1900XT
Seagate Cheetah X15-36LP U160 15KRPM SCSI drive
Fujitsu MAN3367MP U160 10KRPM SCSI drive
Seagate 200GB 7200RPM ATA133 PATA drive
2 CD/DVD RW drives
Fortron 500W PS

My computer uses these fans now:
Newegg.com - Scythe S-FLEX SFF21E Case Fan - Retail

1 intake, bottom front
1 exhaust upper rear (below P/S, but above PCI-E slot)
power supply has built in thermally controlled 120mm exhaust fan
side panel has a duct for the CPU HSF's intake
GPU requires 2 slots, exhausts heat to rear through 2nd slot

And I don't crash. My room's temperature is a steady 70-78F depending on outside temperatures and my mood. The GPU gets VERY hot, but that's normal from everything I've read.
 
As gwok said check your voltage rails in bios, but if they are within the norm 5% ish you are fine, I run the Neo480 in my system
Athlon X2 3800+
1GB ram
7900GT
5 hard drives

anyways it really does sound like a heat issue, monitor the heat using a program in windows as you start running more programs if it gets to the 65C area you most likely have a problem.

i agree with with Invar 13 fans will destroy good airflow, my computer has 2 intake, one rear exhaust, one top exhaust, heatsink fan, and psu fan
 
You must log in or register to reply here.
Back
Top