Sign in with
Sign up | Sign in
Your question

Debunking a gremlin

Last response: in Systems
Share
December 30, 2011 4:42:18 AM

To start off this lengthy post I'll give a quick overview of the problem and what follows is a walkthrough of everything I've done to try to fix it. The problem I have to do is with the graphics in my game freezing or aliasing while playing the game. Originally it would alias and sometimes crash, but after making modifications its now mostly locking up with the only remedy as pressing ctrl-alt-del and then clicking cancel(i could click task manager or anything else, but it just highlights that it seems to need an interrupt or something to unfreeze the screen) or it will occasionally crash and sometimes bsod(which varies in terms of the issue, from ntfs errors to nvidia problems). Anyway as my post goes on to explain, i've been struggling with and researching the problem for a long time now and after trying numerous tests and hardware swaps am finally ready to throw in the towel and ask for help. So without further ado here's what I've tried.

Four years ago, I built my current rig and from that point on have had some strange gremlin living in the machine. Originally I had Windows XP installed and an EVGA 8800 GTX installed and when ever I would play video games, the screen would go black or it would artifact(show random colors and triangles) and freeze about thirty to forty minutes in.

After several years of that I decided enough was enough and I upgraded the video card to an EVGA GTX 480 and had hoped that would end any and all problems I had with playing games on my computer. However that did not fully resolve the issue. After installing the new card games would either crash(this varies from game to game in terms of frequency) or the screen would lock up(not crash but freeze) and pressing ctrl-alt-del(and then clicking cancel) would unfreeze the game and things would continue on fine. Eventually I started getting frustrated with the fact that I just cannot get the gremlin out of my system and that has pushed me to search elsewhere.

I first turned to the ram and reseated it. Then I ran memtest86 for 13 passes with no errors. Then I swapped ram with another computer to test it that way. Finally I turned to the bios and made sure the mobo was running it at 1066 and not 800, I also manually set the voltage to the recommended 2.2 volts(on my dimms it gives the timings and rates it at 2.2v). With all of those tests successfully completed I've decided that it doesn't seem like it could be a ram issue.

Next I turned to my power supply which was an OCZ 700 watt power supply. I opened up SIW and open hardware monitor and notice both reported that my voltage was 8.16 on the 12 volt line! At this point I was positive it was a defunct power supply. So I went out and bought a brand new corsair GS700 power supply. I installed it and checked the voltage in SIW and open hardware monitor and to my disappointment the 12 volt line still displayed 8.16 volts! At this point I wanted to determine if the software was correctly reading the voltage of the power supply so I dragged out my trusty multimeter and connected it to a fan power cable(the one that can also connect to older pata hard drives) and read the voltage both under normal load and under gaming load. In both cases the multimeter read a little over 12 volts. I then went in to the bios to check the voltage level of the 12 volt rail and it too reads correctly at 12 volts. Finally I recently reinstalled Asus's PC Probe and it reports 12 volts on the line. So with that much evidence I decided the culprit is not my power supply(wooo 90 dollars well spent on the new one).

So after discussing with a few of my buddies we turned our attention to both the OS(btw before the upgrade to the gtx480 i had upgraded to windows 7 64 bit) and the hard drives itself. I was originally using a 1 terrabyte drive for my OS that i purchase back when i originally built the rig. My friends suggested that the drive itself could be the culprit as bugs were still being worked out of the system at the time i purchased them. So i decided to pull out an older but reliable 80GB drive and install it in my system and do a clean install of windows 7. When i installed the smaller drive I kept the bios set to treat the hard drives as IDE as opposed to RAID or AHCI drives. I also made sure to keep the driver installation to a minimum(originally just installing the driver for the graphics card and when that didn't work adding in drivers that came with my mobo) but nothing worked. I then decided to try switching over to AHCI and reinstalled windows with the appropriate intel drivers. Still to no effect.

Lastly I've considered temperatures and have kept a close eye on GPU and CPU temperatures. Both have stayed around 40C in game. Lastly I checked my hard drive temperatures and the northbridge temperatures and both were also around 40C(though the NB did occasionally rise to 50C) .

I've spent so much time scouring this forum and others trying to figure out what could possibly be causing the prolonged issue with my computer. All seem to point to bad drivers, ram, video card, and power. The drivers i've ruled out because i've reinstalled them several times both in fresh installs and going through the appropriate procedure of uninstalling then booting to safe mode and using driver sweeper to delete any remaining files. The other issue reported with the drivers is that the sound driver can screw it up(I dont use a monitor but instead use a tv, please don't judge me, i dont think it makes me cool, im just too lazy and poor to buy monitors) so when i did my most recent install i didn't install the sound driver and disabled all audio devices except for the sound card included with my mobo. Ram i've checked and rechecked so many times, that im pretty sure it cant be the ram. As for the video card, I haven't ruled out that I got two bum video cards in a row and unfortunately have no real way of checking this, except for buying a new one. And the power I believe is okay given that I've checked the voltages in hardware and they're fine(using the law for power, the voltage has to drop if it cannot put enough current in to the system). So I'm starting to look towards the motherboard and the CPU as they're the only things I haven't really tested and to be honest when I build this rig I was pretty ignorant to the damaging effects of ESD and how important a grounding strap can really be.

If anyone has any advice I would greatly appreciate the help because I feel like I've run through every test I can think of and have no idea where to turn.

PC Specs:
Intel Core2 E8500
OCZ DDR2 1066 2GB - 2 Sticks
EVGA GTX 480
Asus Rampage Formula(Includes PCIeX1 sound card)
Corsair GS700 Power Supply
Antec 900 Case
80GB 7200rpm drive
1TB 7200 drive - 2(Original)
1.5 TB 7200 drive (Original)
1 TB 7200 drive - Green(New)






More about : debunking gremlin

a c 136 B Homebuilt system
December 30, 2011 6:58:54 AM

give the RAM slightly more voltage .
December 30, 2011 7:13:59 AM

try a BIOS update
Related resources
Can't find your answer ? Ask !
December 30, 2011 7:40:05 AM

have you considered that the MOTHERBOARD could be flakey? Stuff like fried capacitors, or a malfunctioning chipset, building static on a badly isolated/grounded circuit or even just a nick on a data pathway...? Your varying voltage readings seem to show that.

My recommendation:get a new mobo.
December 30, 2011 1:24:51 PM

@Outlander_04 How much more would you recommend? 2.3?

@1tym I'm using the most current BIOS they offer, which unfortunately was last updated 2 years ago

@mitch074 I'm definitely considering that its the motherboard that is ultimately causing the issue. I agree that when I built the computer I could have easily messed up the board when I was installing it. From a technical standpoint how does the varying voltage show that?
a b B Homebuilt system
December 30, 2011 1:47:57 PM

Whoa, hold on here!

First, 2.2V is the max you should give DDR2. You can use (with the right board) 2.3 and 2.4V, but any ram that you use that on should 1) support it and 2) provide it with "active cooling" (ram fan) I would not bump the ram voltage up at all. If anything, set it for DDR2-800 operation with 5-5-5-15 or 18. Leave it at 2.2 if you want, or drop it down to 2.0. At 2.2v it could be a ram overheat issue, but with clean passes of memtest it might not be. Did you ever try overnight running of memtest? I've never used the program so I'm not sure if 13 passes is enough to stress/heat up the ram.

Quote:
I've considered temperatures and have kept a close eye on GPU and CPU temperatures. Both have stayed around 40C in game.


That is way to low for load temps. More so for a GTX480. Might be a bad sensor like your 12V reading. Try firing up a game and after 10-15min or so see if the heatsink for the GTX480 is hot. Notice in this random review the GTX480 peaked at 94c.

http://www.guru3d.com/article/geforce-gtx-470-480-revie...

Still, the odds that you had back to back bad video cards is pretty slim. It seems that you've done all the right steps. OC the CPU at all? Load temps of 40c for a CPU seem pretty low as well, so you might want to double check the CPU HSF as well. Your post seems to say that everything is ok until after awhile of playing games, so it does seem like a heat issue to me. Check your temps on everything, check the sinks. I'd also try over night passes on your ram, and check your drives for errors as well. Baring anything popping up the board would be the next thing to "check" (replace...)
December 30, 2011 2:04:16 PM

I would have to say this is a RAM issue. I looked up your board (did not have the exact model) here is a link.
http://www.asus.com/Motherboards/Intel_Socket_775/Rampa...

If you check your ram is not on the qualified vendor list. Not saying that they test all ram but it is not there.

That being said if I was in your situation seeing as you have an extra HDD, GPU, PSU and OS I would order another smaller case (something cheap), mobo, CPU, RAM and upgrade. Throw all of the old components in the cheaper case and either gift it to someone or use it as a backup or daily browsing machine. If the only thing you have problems with are games who cares, the system is a bit out of date. Just my two cents if I was as frustrated as you seem to be. Since the technology is pretty outdated anyways, its not worth trying other RAM CPU or MOBO unless you just want to know the cause of the problem or have spare parts to swap out.
December 30, 2011 2:24:33 PM

@4745454b When I was getting those low temperatures I was playing with the windows open in my room where the ambient temperature outside was in the low 30s Fahrenheit and had the GPU fan set to max. More recently(I understand I wrote this post yesterday but I wanted to reference the situation I was most frequently using my pc in) I've been playing with the windows closed and the GPU will settle to the 55C range when playing. The CPU will maintain itself at around 44C.

I've actually, accidentally, stepped the speed of the ram down to 800 and tested it at that frequency. Also 13 passes of memtest was an overnight run for me. Also originally I was running my ram at 1.8 volts until I noticed the voltage on the ram said 2.2 volts and had seen a post on this site recommending a person with a similar problem step their ram voltage up.

I have not overclocked the CPU at all, because I have no idea what I'm doing there and since it didn't run well at the normal clock I didn't want to push anything. The low 40C temps could again be a result of my windows being open and I have corsair hydro with two fans on opposite sides of the radiator to pull the air across it.

After you pointed out that I said the problem seems to arise after a while of playing the game I went back and really thought about it and would like to clarify(sorry for not doing this early, but my frustration was blurring my thoughts at the time). The freezing can happen any time, most recently right as I loaded fable 3 and actually got in to the game. It then repeats at random intervals, so it might not happen again or it might occur two seconds later.

@PartialGenious I have noticed that it wasn't on the qualified vendor list but being ignorant I figured that was a marketing ploy rather than a technical requirement. The board also touts Amd's crossfire but has no mention of Nvidia and sli anywhere.

I am planning on building a new rig very soon so this whole process might be pointless, but I wanted to get to the root of it so I can make sure that if i bring over any old hardware, like the 480, I'm not bringing the problem with it. Also the build probably wont happen for a few months(trying to time it with ivy bridge) and seeing as I recently graduated college and dont start working for a few months, I've finally got a lot of time for gaming on my hands. That being said I think all the posts have shown me that its somehow a hardware problem, so maybe I'll be upgrading sooner.
a b B Homebuilt system
December 30, 2011 2:37:55 PM

I've heard SPD settings for the ram can get "picked up" wrong. If you have the settings for DDR2-800 handy, or 1066 if you really want, go ahead and plug them in. 2.2V is the max for DDR2. My G.Skill ram would do DDR2-800 at 5-5-5-15 at 1.8v, or DDR2-800 4-4-4-12 at 2.0. Or I could do DDR2-1066 at 5-5-5-18 at 2.0. Not sure why you are running 2.2 unless you have some cheaper stuff.

Is the power in your room bad? Using a dying battery backup? All games cause this issue? Can you load the computer any other way and make it lock up/crash?
December 30, 2011 2:53:26 PM

To answer your question the Qualified Vendor List is just a reference. I have seen cases though where certain memory is not compatible with certain mobos etc. Personally I never ordered anything but corsair or crucial until recently (I just last week put 16 gb of G skill in my build because it had great reviews and I had a friend that recommended it) because they seem to not be as temperamental. I have seen Kingston sticks that wouldn't boot in one mobo but others it would. Anyways just for reference check these things on your next build then you should have no questions if you decide to call a manufacturer about a problem you are having.

Another thing in my opinion is if you just waiting on ivy bridge I would not worry about it. If your ready for a new build (have the money) go for it because the sandy bridge platform with the right mobo can provide all of the things that Ivy supports, PCIE 3.0, USB 3.0, Sata 6gbps. The processors will be about the same as well, I am sure Ivy will have a bit better processors but not much. Just saying if your always waiting on the next big thing you will never stop waiting. A solid buy now will last a long time.
a c 136 B Homebuilt system
December 30, 2011 10:05:39 PM

rexroentgen said:
@Outlander_04 How much more would you recommend? 2.3?

@1tym I'm using the most current BIOS they offer, which unfortunately was last updated 2 years ago

@mitch074 I'm definitely considering that its the motherboard that is ultimately causing the issue. I agree that when I built the computer I could have easily messed up the board when I was installing it. From a technical standpoint how does the varying voltage show that?


If you are going to add voltage , try by adding the smallest increment your mb will allow . It will be fractions of a volt .
Repeat a couple of times and if its more stabe you may have found the problem

Doing as 4745454b says and slackening the timings and/or reducing the frequency is an alternate path .
Setting it to 4-4-4-12 at 800 Mhz would be the goal , I would say

But just try reducing the frequency and leave the timings alone at first
December 31, 2011 7:41:56 PM

4745454b said:
I've heard SPD settings for the ram can get "picked up" wrong. If you have the settings for DDR2-800 handy, or 1066 if you really want, go ahead and plug them in. 2.2V is the max for DDR2. My G.Skill ram would do DDR2-800 at 5-5-5-15 at 1.8v, or DDR2-800 4-4-4-12 at 2.0. Or I could do DDR2-1066 at 5-5-5-18 at 2.0. Not sure why you are running 2.2 unless you have some cheaper stuff.

Is the power in your room bad? Using a dying battery backup? All games cause this issue? Can you load the computer any other way and make it lock up/crash?


I tried manually entering the timing in and lowering the voltage to the ram with no change. When I opened fable 3 again immediately as i got in game the system froze up and i had to ctrl-alt-del to fix it.

I've moved this computer around alot from apartment to apartment(first it was at home, then went to the dorms at college, then went to apartments near college, and now its finally home) so I don't think its bad power. There is no battery backup on my machine. Every game will give me this error and the only other thing that can cause an error consistently is flash video. If I'm on hulu sometimes the screen will go black and then come back, or worse the image will freeze but audio will still play and if given a few seconds blocks of pixels on the screen will seem to follow what is actually going on in the video, meaning some blocks are actually refreshing properly but others are still stuck on the image it locked up on.

@Outland_04 I previously had my ram set at 800MHz but never with the voltage set above 1.8 volts or with the timings manually entered.
a c 78 B Homebuilt system
December 31, 2011 8:15:22 PM

Go into safe mode and see if you can find any way to crash it.

Watching things on Hulu or whatever.
December 31, 2011 10:57:32 PM

try disabling speed step and other power saving futures. maybe even underclock it a bit as well.
!