Sign in with
Sign up | Sign in
Your question

Help with Freezing/Restarting PC

Last response: in Systems
Share
August 7, 2009 2:12:39 PM

This will be long, I will try to post as much information as I can.

Computer Specs:

Motherboard: ASRock X38TurboTwins
PSU: Silverstone ST70F 700W
Processor: Core 2 Duo E8400 3.0GHz
GPU: Sapphire Radeon HD 4850
RAM: 8GB (G.SKILL 4GB (2 x 2GB) DDR3 1333)x2
OS: Vista Ultimate 64-bit
Case: 1 big fan, 3 smaller fans, aluminum

The Main Problem:


The problem as it currently stands is that it restarts and freezes randomly. It has done this since I built it. It didn't bother me because it wasn't really all that often that it did it, I admit it was stupid of me to not fix it though.

I had never found a pattern, or a way to force the errors to happen. It was completely random (at least to me).

Recently I started playing some video games more often, L4D, Aion. And I finally found a way to kind of force it to restart or freeze. In the game Aion, it happens everytime I play it, sometimes it takes 20min to a couple hours but I know it will eventually happen while playing this game. In other games, like L4D it has only happened once in all the times I play. It still happens every once in a while just doing other stuff in Windows randomly.

Testing so Far:

I started monitoring while playing Aion to see anything weird. Hottest my GPU got was 80C, I turned up some fans and got it down to low 70's C. The voltages were all really close except for +12.0v which was at 12.56 while not playing a game and during play it goes down to 12.4s.

I couldn't pick up anything in particular that caused the freeze or restart.

I ran Prime95, a CPU/MEM stress test for a couple hours and got no errors. The CPU got its warmest during this, getting to 65C. The mobo temp was around 50C.

I unchecked the "Automatically restart" option in Vista hoping it would give me an error message before restarting/freezing but I still didn't get any.

I checked the event logs and only got errors saying "Previous system shutdown was unexpected".

I reset my motherboard to see if I screwed anything up settings memory timings manually and such. After I did this, the game worked for 5 straight hours until I went to bed that night and didn't freeze that night at all. It's came back though. I really thought I had it fixed after this.

I have 4 sticks of RAM and tried switching these in and out. Somehow doing this I made the restarting in Aion quicker after I start playing. Now it happens withing a minute or two into the game. This game is the only thing I can do to force this error to happen, even though it still does it doing anything else, but a lot less often and more random.

I tried moving the GPU to the other PCI express slot, same problems.

Other Problems that may/may not be related:

My monitor has always randomly not found a signal from the PC. Sometimes I would have to turn it on and off a few times till monitor recognized it.

Most of the time when I have to turn the computer off by holding the power button down, when I turn it back on, it starts up for a few seconds, turns off, than a few seconds later turns back on normally.

Maybe my installation?

Since all these problems have been pretty much from the start, I will write some goofy things in my Case. Note though that this case worked fine with my old computer parts in it:

My mobo is only held on by 2 screws, and a post on my case is missing in the one corner of my mobo, making the mobo kind of flimsy there.

My CPU is mounted pretty snug, but during mounting the one corner clip/button that you push down broke off, but its pushed down I am pretty sure and is snug on the mobo.

I have a case fan that has a power cord kind of glued together with some black stuff that sometimes turns on and off because the wire is probably not fully connected in that gooey stuff.

The motherboard has some capacitors leaning slightly but I couldn't find any evidence of black or bulging capacitors, or anything out of the ordinary.

Everything else is pretty normal in the case. At least as much as I can tell, all this new stuff I got was new to me, like the power supply connectors and everything, it was a pretty big upgrade from my old stuff that used older connectors and stuff like that.

My newbie ideas

If I knew more, I probably would have an idea to fix it by now, other than what I tried here, I'm trying to avoid buying a ton of parts if its not the problem, I don't really have any backup parts to try out otherwise I would have tried them out, and I don't really know anybody that has backup stuff I can try.

The PSU: maybe Power supply is messed up so randomly it screws up.

The motherboard: seems like it could cause these types of problems. It was kind of a cheap one.

The GPU: I kept thinking this one especially once I got that one game to always cause the error. Plus with the monitor not detecting a signal sometimes on boot, but would the vid card cause these problems when I'm not doing anything game related?

Windows Vista: I haven't tried a reinstall yet because I have a bunch of videos and stuff with nowhere to back them up, once I find a solution to that I will probably try this. But the problems still say hardware to me, but I can be wrong.

I'm getting close to biting the bullet and buying new parts, I was thinking a new case first, since my old one is falling apart, and then maybe new mobo cause this was a cheap one, what you guys think? My psu I thought was a pretty good one...I hope its not that.

edit - Was thinking, should I just throw this off a cliff and start over? =)

Update 14Aug2009:

Ran memtest86+ for 4 passes, 0 errors
Ram prime95 stress test for a couple hours, no errors

Installed my computer in a new case, properly this time:
*GPU temp dropped 20-25 degrees C! It was only in the 70s before anyway under game load, but now it is around 50C during game play
*CPU dropped 10 degrees as well to 49C during gameplay

After putting everything in the new case I booted up and went to the game where I can usually force this error to happen (freeze/restart). It was freezing or restarting within 1 minute of playing, but I played for an hour this time, and then it restarted, and I went to bed. Temps weren't hot, voltages were good. At least according to my motherboard monitor that came with mobo..I'll make sure to check with another program to be sure.

Edit - Reinstalled windows, still happens.

If that is not it, I am thinking PSU(even though numbers look fine on monitors)/Mobo/GPU/CPU(even though prime95 stress test ran 2 hours no errors)

More about : freezing restarting

a b B Homebuilt system
August 7, 2009 3:12:05 PM

My first concern is the mechanical mounting of your mobo in the case. It appears you have only 2 mounting points (usually there are 5 or 6), plus a broken clip in one spot, and a known unsupported corner. This all makes damage to the mobo during repair work quite possible. For that reason it certainly would be desirable for you to get more of the proper stand-off mounting hardware, remove the board from the case and replace the mouting system so it is solid.

At the same time you should check for a possible hidden problem. I'm wondering whether there are mounting stand-offs under the board in the wrong place that may be causing intermittent shorts. When you do the job, be sure there are stand-offs in every place they are supposed to be, and NO others.

With the poor support your board has now, any work on the board (e.g., moving RAM or Video units) may flex the board enough to create a loose connection in some unpredictable spot. Worst case, it might even cause a mobo trace to crack, creating an intermittent connection on the mobo itself.

One type of problem and solution I have dealt with is poor contacts caused simply by aging and oxidation in connectors. These often can be fixed simply by disconnecting and re-connecting every cable is the case, each several times. That action sort of "scrubs" the metal contact and removes the dirt and oxidation. Obviously you want to do this carefully so you don't break anything. In your case with the poorly-supported board, I would advise extreme care. IF you are going to remove the board from the case completely and re-install its full stand-off support system as I suggested above, that's the time to do all of the connectors and cables.

IF you do the disconnect / reconnect thing, I really advise do NOT do this for the CPU. Poor pin connections of the CPU into its socket are uncommon, whereas it is possible to damage the CPU or its socket in such a process. Unless you have a good reason to remove it and know how to deal with the CPU, its cooler and thermal transfer paste, etc., you are better to leave those parts untouched.

All this takes time and careful work, but the cost (for a few stand-offs) is minimal. If it does not solve your problems, you have not wasted much $$.
August 7, 2009 3:17:03 PM

That's good info and I plan on trying that out. I probably will just go with a new case just because it's time, but that will fix the mounting issue with the mobo, and I know what your talking about with the flex, the RAM is in the corner where the post is missing, and I hope I didn't break anything in the board with it flexing while putting RAM in..looking back it was really stupid of me to put this all together knowing that post was missing...impatience on my part.
Related resources
a b B Homebuilt system
August 7, 2009 3:21:23 PM

In addition to Paperdoc's suggestions, I'd recommend doing a stress test on your RAM. I had an old computer that had similar malfunctions to yours. Problems got notably worse when I tried to play games, but didn't find a clear pattern to the random restarting. Finally downloaded and ran a free little program, memtest86. It's been around for a very long time, but works just fine, you'll probably have to run multiple instances of it to cover all your RAM, but the help files will cover all that when you are ready to run it.

It turned out that one of my DIMMs had a bad address, though the rest of the chips were fine. Whenever I ran games that would use nearly all of my system resources, I'd get a random restart, usually without a blue screen, but those happened at times too.
August 10, 2009 5:03:53 PM

wathman said:
In addition to Paperdoc's suggestions, I'd recommend doing a stress test on your RAM. I had an old computer that had similar malfunctions to yours. Problems got notably worse when I tried to play games, but didn't find a clear pattern to the random restarting. Finally downloaded and ran a free little program, memtest86. It's been around for a very long time, but works just fine, you'll probably have to run multiple instances of it to cover all your RAM, but the help files will cover all that when you are ready to run it.

It turned out that one of my DIMMs had a bad address, though the rest of the chips were fine. Whenever I ran games that would use nearly all of my system resources, I'd get a random restart, usually without a blue screen, but those happened at times too.



I am going to run a memtest tonight overnight and see how that goes.
August 14, 2009 2:28:24 PM

Update on what I have done:

Ran memtest86+ for 4 passes, 0 errors
Ram prime95 stress test for a couple hours, no errors

Installed my computer in a new case, properly this time:
*GPU temp dropped 20-25 degrees C! It was only in the 70s before anyway under game load, but now it is around 50C during game play
*CPU dropped 10 degrees as well to 49C during gameplay

After putting everything in the new case I booted up and went to the game where I can usually force this error to happen (freeze/restart). It was freezing or restarting within 1 minute of playing, but I played for an hour this time, and then it restarted, and I went to bed. Temps weren't hot, voltages were good. At least according to my motherboard monitor that came with mobo..I'll make sure to check with another program to be sure.

Next step I am going to reinstall windows just to rule that out, plus it's a free test that doesn't require me buying a part to replace.

If that is not it, I am thinking PSU(even though numbers look fine on monitors)/Mobo/GPU/CPU(even though prime95 stress test ran 2 hours no errors)

Any thoughts anyone?
a b B Homebuilt system
August 14, 2009 2:42:44 PM

Which exact G.Skill RAM kits are you using? You said you manually set the timings, but I didn't see you mention manually setting the RAM voltage. Running 4 sticks of RAM is harder on the motherboard and sometimes requires bumping up both the RAM and northbridge voltages a bit for stability. I've even seen RAM that wasn't getting enough voltage pass Memtest86+ but fail under everyday use.
August 14, 2009 2:58:31 PM

shortstuff_mt said:
Which exact G.Skill RAM kits are you using? You said you manually set the timings, but I didn't see you mention manually setting the RAM voltage. Running 4 sticks of RAM is harder on the motherboard and sometimes requires bumping up both the RAM and northbridge voltages a bit for stability. I've even seen RAM that wasn't getting enough voltage pass Memtest86+ but fail under everyday use.



http://www.newegg.com/Product/Product.aspx?Item=N82E168...

I had them manually set since I built the computer. Recently I reset my mobo with the battery, and haven't manually reset them.

I don't remember what I had them set at before.


I think I had set the RAM voltage to 1.65 before. But I don't think I ever changed the northbridge voltage, as I don't really know where that is to set haha
a b B Homebuilt system
August 14, 2009 3:01:55 PM

I would go back into the BIOS and make sure they're manually set to DDR3 1333 @ 7-7-7-18 timings and 1.6v. You may even need to bump it up to 1.7v when running 4 sticks of RAM, but I would only do that if you continue to have problems. Bumping the northbridge voltage up a notch may also help.
August 14, 2009 3:05:28 PM

shortstuff_mt said:
I would go back into the BIOS and make sure they're manually set to DDR3 1333 @ 7-7-7-18 timings and 1.6v. You may even need to bump it up to 1.7v when running 4 sticks of RAM, but I would only do that if you continue to have problems. Bumping the northbridge voltage up a notch may also help.


Ok, thanks for the info, I won't be home till sunday to get to that computer, but I will make changes and post on here the results by monday. Thanks.
August 14, 2009 3:14:27 PM

Haha, I definitely won't, thanks.
August 17, 2009 5:33:23 PM

shortstuff_mt said:
The NB voltage setting is right under the RAM voltage setting in the BIOS. It's in the advanced chipset section and labeled "NB Core Voltage".

http://www.pro-clockers.com/images/review/asrock%20x38t...

(Don't set it to "Highest" like in the screenshot) :) 



I played with the voltages, set everything manually, and made sure RAM voltage was 1.65, Everything was 7-7-7-18 in CPUID and I tried with the NB voltage normal, than I tried middle, and then I tried high. It still did it each time.

In a game I play sometimes I get this black flicker, so I think I'm going to try another GPU when I get my hands on one.
a b B Homebuilt system
August 17, 2009 5:58:49 PM

Silverstone PSUs are supposed to be pretty nice, I was pretty confident that it wasn't causing issues, but since you are still having the same problems, I'm not so sure. I took a closer look at your PSU's specs, and it has 4 12v rails, but they each are only rated at 18A. 4850's aren't particularly power hungry cards as I know, but the trend has been for newer power supplies to offer single 12v rails capable of outputting more current total.

Also, since you are running out of things to check, testing a different Gfx card might identify a problem there instead of the PSU.
August 17, 2009 7:04:54 PM

wathman said:
Silverstone PSUs are supposed to be pretty nice, I was pretty confident that it wasn't causing issues, but since you are still having the same problems, I'm not so sure. I took a closer look at your PSU's specs, and it has 4 12v rails, but they each are only rated at 18A. 4850's aren't particularly power hungry cards as I know, but the trend has been for newer power supplies to offer single 12v rails capable of outputting more current total.

Also, since you are running out of things to check, testing a different Gfx card might identify a problem there instead of the PSU.


Ya I am going to get another video card in a few days to see if that is the cause. Thanks.
August 20, 2009 2:52:14 AM

wathman said:
Silverstone PSUs are supposed to be pretty nice, I was pretty confident that it wasn't causing issues, but since you are still having the same problems, I'm not so sure. I took a closer look at your PSU's specs, and it has 4 12v rails, but they each are only rated at 18A. 4850's aren't particularly power hungry cards as I know, but the trend has been for newer power supplies to offer single 12v rails capable of outputting more current total.

Also, since you are running out of things to check, testing a different Gfx card might identify a problem there instead of the PSU.




I got a new video card, same brand as before, tried it, still restarted..just for kicks, put it in crossfire with my old one (same cards), still did it.

I also did more messing around with dram voltages and nb voltages, still does it.

Motherboard now or PSU I guess? Here are my voltages just for kicks: They all fall within +/- according to PSU manual. Here is during a large FFT prime 95 test, that's why the cpu temps are up. (I'm working on getting them lower...adding some fans)


Hardware Monitoring : Winbond W83627DHG
Voltage CPU : 1.18 V
+3.3V Voltage : 3.26 V
+5V Voltage : 5.04 V
+12V Voltage : 12.51 V
Processor Fan : 1834 rpm
Processor Temperature : 52 °C
Mainboard Temperature : 31 °C
Power/Aux Temperature : 20 °C
:
Processor : Thermal Diode
Intel Core 2 Duo (Core 1) : 75 °C
Intel Core 2 Duo (Core 2) : 71 °C
:
Hard Disk Monitoring : S.M.A.R.T
Hard Disk Hitachi HDT725032VLA360 : 35 °C
Hard Disk SAMSUNG HD502IJ : 24 °C

I'm getting a little tired of buying new parts.. I'm hoping the next part is the right one...mobo or psu or cpu maybe? Every prime 95 test I have ever ran ran without errors, but once I ran one overnight and after 2 hours it rebooted. But that was before the case change/gpu change, voltage adjustments, ram adjustments.
a b B Homebuilt system
August 20, 2009 1:09:15 PM

Looking at your max load temp readings, I think you're running too hot. Even at idle temps, you should be maintaining a cooler reading. I was reading your older posts, and saw that you mentioned one of the heatsink clips is broken... I'm beginning to think that could be the problem. You are using the stock cooler correct? You really need more than a "snug" fit, the HSF needs to be held on under tension. Before you try more expensive replacements, I'd recommend trying a reasonably priced aftermarket cooler and fresh thermal paste.
a b B Homebuilt system
August 20, 2009 1:21:09 PM

Agreed with Wathman, an Akasa Nero AK-967 with some Arctic Silver 5 will be of great use to you.

I would have said RAM is the problem, but there are no blue screens throwing up so i guess not.

Other than your CPU, i'd say its the motherboard damaged, which means replacement.
August 20, 2009 2:44:05 PM

wathman said:
Looking at your max load temp readings, I think you're running too hot. Even at idle temps, you should be maintaining a cooler reading. I was reading your older posts, and saw that you mentioned one of the heatsink clips is broken... I'm beginning to think that could be the problem. You are using the stock cooler correct? You really need more than a "snug" fit, the HSF needs to be held on under tension. Before you try more expensive replacements, I'd recommend trying a reasonably priced aftermarket cooler and fresh thermal paste.



The clips are all through the back of the motherboard, like they are supposed too. And it feels tight, but if I can lower the temps and possibly fix this problem I can buy another one. Are there any good ones that aren't stand up ones like the fellow below me suggests? I don't really want a behemoth heat sink in there..i"m not trying to overclock this thing at all.
August 20, 2009 2:45:21 PM

Griffolion said:
Agreed with Wathman, an Akasa Nero AK-967 with some Arctic Silver 5 will be of great use to you.

I would have said RAM is the problem, but there are no blue screens throwing up so i guess not.

Other than your CPU, i'd say its the motherboard damaged, which means replacement.


I don't doubt how good that heatsink is. But it's very large, 120mm fan on it.. I would also prefer one that lays down on the cpu, not stands up. But if there aren't any good lay down ones, I would buy one of the stand up ones. But still..that Akasa is big.
a b B Homebuilt system
August 20, 2009 2:54:24 PM

the large HSF combos like that Akasa really do help with CPU and case interior temps. They direct the heated air right into the exhaust fan on your case, make a nice linear flow from your front intakes to the rear. My current desktop is similar to yours and I use a Zalman cooler that has a 92mm fan on it. Don't remember what my temps are currently, but I've never had heat issues.
August 24, 2009 11:32:18 PM

wathman said:
the large HSF combos like that Akasa really do help with CPU and case interior temps. They direct the heated air right into the exhaust fan on your case, make a nice linear flow from your front intakes to the rear. My current desktop is similar to yours and I use a Zalman cooler that has a 92mm fan on it. Don't remember what my temps are currently, but I've never had heat issues.



I got the cooler master hyper tx2

http://www.newegg.com/Product/Product.aspx?Item=N82E168...

Just hooked it up...cores are running at 47/44 now and 37c cpu during prime95 large fft test.

dropped 30 degrees..haha

Oh and the instability? My motherboard and ram can't do 8gb. If I run either pair by itself, it's stable (2x2gb), but with all 4 sticks in it, it is not stable. The ram tech guy told me my tRFC is too low(60) to run stable with 4 sticks of ram. My mobo doesn't have an option to change tRFC.. i tried changing voltages..overclocking cpu/ram, timings i could change, nothing.

I am going to try slowing the timings from 7-7-7-18 to 9-9-9-24 to see if that helps.

After that, I don't particularly want to buy a new mobo right now just to run this ram, and I don't want to buy new ram either, so I'll be running 4gb because honestly I haven't needed anymore than that yet anyway. Thanks for all your help, if you have another suggestion let me know but the gskill ram tech guy was basically sending me links to boards on newegg that will run my ram he said. I still might buy a new mobo, we will see.

a b B Homebuilt system
August 25, 2009 12:42:32 PM

Glad to hear you figured out what was wrong. Thought it was odd you were getting worse temps on a more efficient Wolfdale C2D than my much older Conroe. Though now that your temps are really good now, your CPU and motherboard will probably last longer and won't conveniently die when you want to upgrade ;) 

As for your tRFC finding, I never knew about that so I just did a little research on it. Are you sure the G.Skill guy said that your tRFC was too low(60) to run stable? I just read a thread over at tweak town that covered advanced memory timings. It summarized by saying that low tRFC (46 and lower) that leads to data corruption and can be a source of corruption. If you bump the value higher, the RAM performance decreases, but becomes more stable.

It was correct to bring up this issue when dealing with 8 GB of RAM, though I'm trying to determine if what you said the G.Skill tech told you is right, or if what I read is right. One of my friends has been plagued with system instability and it just so happens he has 8 GB of RAM on his system. It might be possible that your G.Skill RAM is clocked so high that running your tRFC that low just crashes the system.

http://forums.tweaktown.com/f69/about-trfc-27672/
August 25, 2009 1:00:29 PM

wathman said:
Glad to hear you figured out what was wrong. Thought it was odd you were getting worse temps on a more efficient Wolfdale C2D than my much older Conroe. Though now that your temps are really good now, your CPU and motherboard will probably last longer and won't conveniently die when you want to upgrade ;) 

As for your tRFC finding, I never knew about that so I just did a little research on it. Are you sure the G.Skill guy said that your tRFC was too low(60) to run stable? I just read a thread over at tweak town that covered advanced memory timings. It summarized by saying that low tRFC (46 and lower) that leads to data corruption and can be a source of corruption. If you bump the value higher, the RAM performance decreases, but becomes more stable.

It was correct to bring up this issue when dealing with 8 GB of RAM, though I'm trying to determine if what you said the G.Skill tech told you is right, or if what I read is right. One of my friends has been plagued with system instability and it just so happens he has 8 GB of RAM on his system. It might be possible that your G.Skill RAM is clocked so high that running your tRFC that low just crashes the system.

http://forums.tweaktown.com/f69/about-trfc-27672/



He told me it needs to be at least 68 to run stable. I think ddr3 tRFC is higher than ddr2. I read some forums where people were settings theirs at 80-88 to run stable. I tried running the 8gb at 9-9-9-24 and the tRFC become 74 when I did this. I still had stability issues though with this so I took the ram back out.

And the core temps before? Well that one clip was broken, it could of not been seated properly, I mean dropping 30 degrees is a lot, even for a better cpu hsf.
a b B Homebuilt system
August 25, 2009 1:10:15 PM

Ok, must have something to do with how fast the G.Skill modules are. They just need to have tRFC values that much higher to get stability I guess. Some of the forum posts I read were about DDR2 memory as well so you're probably right about the tRFC being higher on DDR3.
!