Topic: What do you do to keep your computer cool when gaming?

[Kid Carrson]

What steps do you take to keep your computer cool to prevent over heating and lock ups when doing heavy gaming? 

Thanks for any info.

KC

[FPF-SCM_TraceyG_XC]

Use an all aluminium case, the bigger the better. Aluminium cases have far better thermal properties than plastic ones (and look better too). A big case (a tower case) will also allow you to add extra chassis fans, I myself have 6 chassis fans.

You can also install better cooling fans on components. I have a Gigabyte 120mm heatpipe radiator for the CPU. Heatpipes have 2000 times the thermal conductivity of an ordinary copper pipe. They contain a fluid inside them which boils at the CPU end, the redulting gas flows along the pipe to the radiator where the heat is transferred and the gas changes phase back into a liquid. The phase change and latent heat energy (for those who remember their basic chemistry) is why the thermal conductivity is very efficient.

[E_Look]

My son has a locking up problem playing online, too.  He plays on a system built on a DFI NF2 Ultra 400 type motherboard with an AMD Athlon XP 2500+ (Barton core) @ 1.833 GHz, 1 Gb RAM (1 512 stick KingSTON and 1 512 stick KingMAX), a Sapphire OEM Radeon 9600 XT/128 Mb, and a MGE 400 W power supply.

The system just hangs on him... often.

These days, the mismatched RAM shouldn't be an issue... or is it??

Maybe the OEM Sapphire 9600XT doesn't like some online game graphics?

I don't think it's cooling in his case: three system fans in addition to the built in one on the power supply.

Is the MGE a reliable unit?  It IS a "name brand".

[KBF MalaK]

My XP2200+ gets flaky when the temp reaches redline (about 70C) but when I drop the FSB speed down to 100Mhz the temp drops about 20 degrees. I havent noticed a considerable performance difference at the lower bus speed and the CPU sure likes it so I'll keep an eye on it and let you know. The lockups are gone now.

[Bonk]

No overclocking and a chassis fan. Stays at body temperature.

[E_Look]

Bonk, I NEVER overclock.  I think I have too much heat as it is right off the bubblewrap.

I'm just wondering if any of the components I posted two posts up might have anything to do with system instability, as:

* the vid card sometimes works better with older drivers,

* sticks of RAM from different companies in the *OLD* days used to cause lockups

* maybe my "name brand" PSU's company got their name making other stuff besides power supplies.  (If it is, it'll be too bad, as, boy, it
   sure looks great since it has a built in LED lit fan.      )

(And before anyone sniffs, my own system is in a plain case, no lights or doodads; this one is for a kid.)

[Javora]

My son has a locking up problem playing online, too.  He plays on a system built on a DFI NF2 Ultra 400 type motherboard with an AMD Athlon XP 2500+ (Barton core) @ 1.833 GHz, 1 Gb RAM (1 512 stick KingSTON and 1 512 stick KingMAX), a Sapphire OEM Radeon 9600 XT/128 Mb, and a MGE 400 W power supply.

The system just hangs on him... often.

These days, the mismatched RAM shouldn't be an issue... or is it??

Maybe the OEM Sapphire 9600XT doesn't like some online game graphics?

I don't think it's cooling in his case: three system fans in addition to the built in one on the power supply.

Is the MGE a reliable unit?  It IS a "name brand".

Generally yes mismatched Ram can be a problem, I wouldn't even dream of putting mismatched Ram into my system.  But then my Asus motherboard is extremely picky.  I have never heard of MGE power supplies when I did a quick search of that power supply I found that it is usually sold with a case.  I generally don't use those power supplies unless I'm buying a Antec case, your milage may very.

What game is your son playing, if he is playing something like Battlefield 2 then I would suggest upgrading the video card to at least a ATI X800XT or nVidia 6800GT.  I this round of the video card war I would suggest nVidia over ATI right now.  But then if he is playing older games like SFC then the 9600XT should be just fine.  Also if your son is playing BF2 then I would upgrade to 2Gb of Ram (with matching sticks), I've heard a lot of people and computer magazines say that BF2 runs better with 2Gb of Ram.

[Nemesis]

Generally yes mismatched Ram can be a problem, I wouldn't even dream of putting mismatched Ram into my system.  But then my Asus motherboard is extremely picky.  I have never heard of MGE power supplies when I did a quick search of that power supply I found that it is usually sold with a case.  I generally don't use those power supplies unless I'm buying a Antec case, your milage may very.

What game is your son playing, if he is playing something like Battlefield 2 then I would suggest upgrading the video card to at least a ATI X800XT or nVidia 6800GT.  I this round of the video card war I would suggest nVidia over ATI right now.  But then if he is playing older games like SFC then the 9600XT should be just fine.  Also if your son is playing BF2 then I would upgrade to 2Gb of Ram (with matching sticks), I've heard a lot of people and computer magazines say that BF2 runs better with 2Gb of Ram.

RAM makers do seem to push the edges in different directions.  Myself I prefer to buy a full set at once. 

If it is a ram issue taking out one stick and running with a "mere" 512 MB would demonstrate it.  Of course you (E_Look) might have to do it twice if you get the wrong stick on the first try.

I don't think anyone has mentioned it but sometimes the basics do get overlooked.

1/ Are your fans all working?  Especially the CPU and video card fans?

2/ Are your heat sinks clean?  Especially the CPU and video card heat sinks?

3/ Airflow.  Are your case intake fans and vents clear and unobstucted?  No magazines stacked against the vents? 

4/ External heat sources.  Is the cooling air that is being sucked in being preheated by another unit?

5/ When did you last defrag and virus scan?

[E_Look]

Oops.

Only *I* have admin rights on his rig.  He's still little.  Maybe I ought to go and update, then do a spyware/AV/defrag session on his system.

I'll try to let you guys know how THIS works, as this is the most painless and probably on target thing to do.

[Stormbringer]

I use a Thermaltake subzero peltier effect electric cooler. i worry not about excess heat but whether it cools the hardware below the condensation point or not. It can put a ryme of ice on the CPU but the intelligent controller detects all onboard temp sensors and adjusts the power to the cooler.

I will never again buy an aluminum case sight unseen. The one i got was so flimsy it constantly popped stuff out of the expansion slots. if the aluminum is heavy enough i'll get it but reviews can be decieving and purchasing over the internet can get you a piece of crap as far as aluminum cases go.

[Javora]

I don't think anyone has mentioned it but sometimes the basics do get overlooked.

1/ Are your fans all working?  Especially the CPU and video card fans?

5/ When did you last defrag and virus scan?

Nemesis has a good point your system could be overheating, or for that matter you could have a defective part on your hand.  Go ahead and do the defrag/scan routine and let us know how things turn out.

[IAF Lyrkiller]

My system has 3 case fans w/ an option of doing a fourth. My CPU has a huge all copper heatsink w/ 4 heat pipes and  a 92mm fan.

And it does run fairly cool. PSU on the other hand runs fairly warm. 

[Tus-XC]

stormbringer, how much was that peltier and how well does it work compared to say, water cooling? been thinking about switching my system over to somthing other than aircooling, though right now under load i barely go above 50C (in a hot room), stilli rather it be a lil lower

[Stormbringer]

stormbringer, how much was that peltier and how well does it work compared to say, water cooling? been thinking about switching my system over to somthing other than aircooling, though right now under load i barely go above 50C (in a hot room), stilli rather it be a lil lower

If I recall correctly it was a little over 100 bucks but that is for a peltier plate, heat sink fan and a pci slot based power supply / controller. The draw backs are that the power supply controller is a bit bulky and takes up one pci slot. perhaps by now they have a pci express or pci II based controller but it would still take up extra room on the MB. additionally it requires another power cord in addition to your main power supply cord. if the controller goes bad you can get condensation on the MB and CPU or it could just stop working and possibly over heat.

OTOH: I have not had a problem with the controller except from case flexing loosening it's PCI connection or the power cord coming loose. the electronics have not failed yet.

[E_Look]

I don't think anyone has mentioned it but sometimes the basics do get overlooked.

1/ Are your fans all working?  Especially the CPU and video card fans?

5/ When did you last defrag and virus scan?

Nemesis has a good point your system could be overheating, or for that matter you could have a defective part on your hand.  Go ahead and do the defrag/scan routine and let us know how things turn out.

Well, I updated AdAware, Spybot and ran them.  Then I defragged.  Kid says it's worse than before (meaning it's just as bad).

You guys really think it's a cooling issue and not that of drivers?  If I save a few bucks, maybe I'll check the RAM.

[Stormbringer]

I don't think anyone has mentioned it but sometimes the basics do get overlooked.

1/ Are your fans all working?  Especially the CPU and video card fans?

5/ When did you last defrag and virus scan?

Nemesis has a good point your system could be overheating, or for that matter you could have a defective part on your hand.  Go ahead and do the defrag/scan routine and let us know how things turn out.

Well, I updated AdAware, Spybot and ran them.  Then I defragged.  Kid says it's worse than before (meaning it's just as bad).

You guys really think it's a cooling issue and not that of drivers?  If I save a few bucks, maybe I'll check the RAM.

I believe there is a a ram diagnostic in the free version of Sisoft Sandra benchmarking software. it is an awesome set of tools. you should download it anyway.

[Bonk]

Get rid of software firewalls. If Norton is installed ditch it.

[E_Look]

Oh, this machine has not yet seen any Symantec stuff on it.

And I got rid of Norton anything on my own machine.  It DOES now run a egree or even two cooler, believe it or not!

Where do I go to d/l Sandra?

[Stormbringer]

It's everywhere! It's everywhere!


http://www.sisoftware.co.uk/index.html?dir=dload&location=sware_dl_all&langx=en&a=

[Elvis]

Check the ram with memtest, can be found here: http://www.memtest.org/
             

[SkyFlyer]

Ive OCed my P3 1ghz to 1.1ghz with a copper heatsink and a big fan on my CPU. I then have a fan in the front and one in the back, not counting my PSU fan. I also have the option of having a fan on my side panel but I havent gotten around to installing it yet. After 12 hours of continuous running my system gets unstable, however temperature stays well below 50* Celcius.

I overclocked my ATI Radeon 9550 (just a heatsink) from 250mhz Core clock and 195.75mhz Memory clock to 465.75 mhz Core Clock and 233mhz Memory. I get no artifacts after testing with ATITool 0.24 for 11 hours.

[Nemesis]

You guys really think it's a cooling issue and not that of drivers?  If I save a few bucks, maybe I'll check the RAM.

It is not so much that we think it is overheating.  It is more that it could be overheating and it is a fairly quick, easy and cheap thing to check.  If it is heat any fix based on replacing hardware will fail and be a waste of time and money.  If it is not heat then the time and money to eliminate that as the source is very little in comparison. 

Also any steps taken now to clean vents and heat sinks will be needed eventually anyhow so the effort will not be wasted.

When the source of the problem is unclear always check the quick easy and cheap possiblities first.  As one example a friends computer started malfunctioning after he moved.  I was at his house shortly thereafter and he asked me about it.  I began inspecting the system and quickly noticed something.  The surge protector and a light for wiring fault detection - it was on.  He plugged it in elsewhere and the system worked stably again.

Which of course brings up another point.  Power.  Years ago a study on PCs concluded that 85% of computer crashes were power related.  Has anything else been added to the circuit that computer is on?  I have seen problems with flourescent lights on the same circuit as the monitor for example.    Do the lights in the room with the computer flicker occasionly?  Those flickers indicate power fluctuations that can cause crashes.  Low power is less damaging than a surge but both can cause a crash.

[E_Look]

Interesting, Nemesis!

I don't think there's much on the line; his brother's computer is on the same line, same basic config, same basic parts, except older models from the same manufacturers, and it runs like a charm.  I'll check the dirt inside his case first, then I'll run Memtest like Elvis suggested.

And, as always, I'll keep yousall posted.

[E_Look]

Gentlemen,

I ran MemTest 1.6+

and it found two errors:  I have no idea what they mean.  Can anyone clue me in about this?

[Elvis]

Which stick is producing the errors? And how long did you let it run? Check each stick seperately and let it run through a couple passes. If there is one stick that isn't producing errors go with that and see if the puter still crashes. As with everything "computer" you may or may not have found the culprit.

I had a matched pair of Corsair sticks and one stick went bad, ie showed errors on Memtest which was enough to get an RMA from Corsair. Lifetime warrranty is a good thing.

[Javora]

Elvis is right separate the Ram and run the program again for each stick.  Be carefull to write down the error codes, so if your Ram is still under warranty you can get a replacement.

[E_Look]

If I post the error codes, can you tell me what they mean?

[Elvis]

FAQ for Memtest:

http://forum.x86-secret.com/viewtopic.php?t=2906&sid=a7c3352df44e4e2d1492d7977b437ecd


Also this page has good information on what the tests actually do:

http://www.memtest86.com/#details

[Commander Maxillius]

Ive OCed my P3 1ghz to 1.1ghz with a copper heatsink and a big fan on my CPU. I then have a fan in the front and one in the back, not counting my PSU fan. I also have the option of having a fan on my side panel but I havent gotten around to installing it yet. After 12 hours of continuous running my system gets unstable, however temperature stays well below 50* Celcius.

I overclocked my ATI Radeon 9550 (just a heatsink) from 250mhz Core clock and 195.75mhz Memory clock to 465.75 mhz Core Clock and 233mhz Memory. I get no artifacts after testing with ATITool 0.24 for 11 hours.

To keep my gaming system cool I take it off my lap   It only hits 50C if I don't run SpeedFan and let the system decide when it's too hot.  With SpeedFan it never gets above 45C, even with all the airholes blocked.  You can configure when to have it turn each individual fan on, and it'll detect all of them.  It'll also detect any temp sensors, it the case of my laptop, it's got 4: Local Temp (CPU), Remote Temp (Processor, probably an estimate based on the CPU gauge +5C), HD0 (hard disk, though I can't see why the HD needs a temp sensor), and Temp1 (which I haven't found a purpose for yet, since it either displays what Remote Temp does or like now, says 0C).

link to SpeedFan: http://www.almico.com/speedfan.php

[SkyFlyer]

I don't even get close to 50. I get around 42* after six hours or so of 100% cpu.

Idle its around 33.

[Commander Maxillius]

I don't even get close to 50. I get around 42* after six hours or so of 100% cpu.

Idle its around 33.

My fans don't turn on until it hits 41, then it cools to 37 or 38 then switches them off.  That turns out to be on 10 mintues, off 5.  I did an experiment and turned "auto cool" off for half an hour and the processor hit 60 and the CPU was at 48, but once I turned the fans back on everything was back to normal (CPU below 50) in 3-5 minutes.  The stock fan controller doesn't turn the fans on until the CPU hits 65 or 70, because the system runs WAY cooler now.  It doesn't heat my lap at all anymore.

[SkyFlyer]

Nice. IIRC the P3 can handle up to 70 degrees so 42 isnt too bad

[Commander Maxillius]

It's actually a Celeron, and I don't notice any difference in stability between running the stock controller and SpeedFan.  I do notice, however, that the clockspeed is a true 1200MHz when it's cooler as opposed to something like 1191MHz after a night of SFC. 

I'll have to see how it does with Neverwinter Nights now.

[Commander Maxillius]

in SFC the processor sensor read a max of 57.  The processor sensor always reads higher than the fan controller, so I'll call it 53.  I think NWN will push the controller to 60.

[SkyFlyer]

I think the Fancontroller is reading the case temperature.

Just for fun try taking the case off and then play. You should see a much more drastic gap between the two.

As for 57* on the CPU thats starting to get a bit high, even for a celeron.

[Commander Maxillius]

Taking the case off would be a tad difficult since it's a laptop, and the high temp *may* be due to the fact that it was sitting on a mattress :-p

The CPU only touches 50 on a table.

[SFCOP]

i am using a very old, big(160mm, i think) fan that used to cool some industrial equipment where my dad works. It is connected to my PS by paper clips jamed into the power connector and is being run at 7v(best sound/air flow ratio) but can handle 24v. I have to have the case off all the time as the fan sits on top aiming down onto the CPU and RAM. Keeps system below 40c almost all the time. I have a Athlon XP 3200 CPU and Radeon 9800 pumping out the heat.

[Midnight Tech]

To keep my gaming system cool I take it off my lap It only hits 50C if I don't run SpeedFan and let the system decide when it's too hot. With SpeedFan it never gets above 45C, even with all the airholes blocked. You can configure when to have it turn each individual fan on, and it'll detect all of them. It'll also detect any temp sensors, it the case of my laptop, it's got 4: Local Temp (CPU), Remote Temp (Processor, probably an estimate based on the CPU gauge +5C), HD0 (hard disk, though I can't see why the HD needs a temp sensor), and Temp1 (which I haven't found a purpose for yet, since it either displays what Remote Temp does or like now, says 0C).

link to SpeedFan: http://www.almico.com/speedfan.php

Maxillius, the reason they give a HDD temp sensor is due to the high sppeds most of the new drives are running...a 10,000 RPM drive can actually run hot enough to burn you after just a short run time.

[Stormbringer]

http://www.overclockercafe.com/Reviews/cooling/Tt_SubZero4G_P4/pg_5.html

[FPF-SCM_TraceyG_XC]

There is also another issue that may need consideration. Where intermittent memory errors are causing a computer to crash may not actually indiciate a fault with the RAM, but rather the power supply. When a power supply begins to fail, or is under too much load, the voltage produced becomes unclean or can drop too low. Typically, the most sensitive computer hardware to voltage variations will be effected first, and this is usually the RAM. Testing the power supply under load may also be prudent, and even here, it needs to be tested using the same power point and place in the house where it normally runs.

A friend of mine had similar problems once, the cause was actually a faulty power point that was on a bad circuit which did not supply enough power to the computer during certain times of the day. The fault was fixed simply by moving the computer to a good power outlet.

[E_Look]

Very interesting, Tracey!

But how can one practically tell whether it's the RAM independent of any power input or "dirty" power?

[Javora]

Very interesting, Tracey!

But how can one practically tell whether it's the RAM independent of any power input or "dirty" power?

Process of elimination, start with a battery backup, then have your power supply tested and/or replaced.  If you still have problems with your system try something like Memtest and go from there.  Hope this helps.

[E_Look]

I decided not to TEST the power supply.  If you say you've never heard of MGE (local NY area stores and some on line places seem to have them popping up lately) its good enough for me... AND when I hold the bugger in my hands, it's LIGHT!  The best things about it were the shiny chrome finish and the 120 mm extra blue LED fan on it.

So, I replaced it with a Thermaltake Purepower 420 W (dual fan) unit (and cut my palm on the corner of a PCB pulling out one last recalcitrant power connector).  No fancy schmancy LED lit fans, but it's got two QUIET fans, and in reality, that counts more.

And, so far, no crashes or lock ups, except for the time I tried to move some of that rat's nest mass of wiring around to more convenient spots in the case... stupidly... with the power on and my kid playing something on it.  Outside of that, it seems stable now.  Of course, the real test is the weekend, when they go ape playing whatever online, offline, inline, outline, outright awful... on the PC.  I'll post again after then to report on the acid test results for the Thermaltake's effect on stability.

Now, mind you, my other kid's comp, an older machine, has exactly the same PSU and his system has NEVER given him any trouble.  So, if the system still hangs, I'm going to have to... sigh... investigate RAM prices further...


*EDIT* - I like the this particular Thermaltake unit because it is stable, trouble free, made by a KNOWN reputable manufacturer of such parts, and CHEAP!!  From Newegg, it's only a tad over $40 (USD) after shipping.

[Javora]

So, I replaced it with a Thermaltake Purepower 420 W (dual fan) unit (and cut my palm on the corner of a PCB pulling out one last recalcitrant power connector).  No fancy schmancy LED lit fans, but it's got two QUIET fans, and in reality, that counts more.

*EDIT* - I like the this particular Thermaltake unit because it is stable, trouble free, made by a KNOWN reputable manufacturer of such parts, and CHEAP!!  From Newegg, it's only a tad over $40 (USD) after shipping.

I use a Thermaltake power supply in my system and I have never had a problem with it.  IMHO 420 watts is on the light side, I would have went with at least 480~500 watt power supply.  But the important thing is that the power supply works and is stable, by what you said so far seems to be doing just fine.

[E_Look]

Well here's a partial update:

First, you're right about the capacity of the PSU, but I figured that for the kids, I put in what... one floppy drive, one DVD drive (CDR in the older one), either a 30 or 80 Gb HD, and a Zip drive, either a Radeon 9500 or 9600 XT, and a wireless card... that it wouldn't require more than 400 W.

Now, we gave the system a quick couple of tests, running their favorite online game, opening a whole bunch of windows in the background at the same time, and it froze once in about an hour.

Yes, that was a short time, but again, this weekend will be the real test.

But it did freeze that one time, in the middle of their online game... though there was a lot of lag due to a high number of users... I wonder if that made a difference...

... so I guess I should start checking out RAM prices?  I'll run MEMTEST again now that I've got in a PSU I know to put out a clean signal.  So, next week, I'll post again on this!

Thanks, Jav, and all you others.  It IS much appreciated.

[Javora]

Well here's a partial update:

First, you're right about the capacity of the PSU, but I figured that for the kids, I put in what... one floppy drive, one DVD drive (CDR in the older one), either a 30 or 80 Gb HD, and a Zip drive, either a Radeon 9500 or 9600 XT, and a wireless card... that it wouldn't require more than 400 W.

Now, we gave the system a quick couple of tests, running their favorite online game, opening a whole bunch of windows in the background at the same time, and it froze once in about an hour.

Yes, that was a short time, but again, this weekend will be the real test.

But it did freeze that one time, in the middle of their online game... though there was a lot of lag due to a high number of users... I wonder if that made a difference...

... so I guess I should start checking out RAM prices?  I'll run MEMTEST again now that I've got in a PSU I know to put out a clean signal.  So, next week, I'll post again on this!

Thanks, Jav, and all you others.  It IS much appreciated.

I'm not sure that is a accurate way to test the power supply.  If anything you just tested the Processor, Ram, Hard drive, the system bus, and the bandwith of your Internet connection.  You said that the system crashed after an hour, given what you had running that's not bad.  Question though if I may, what video game were they playing when it crashed?  If it was Battlefield 2 then you may want to consider 2GB of Ram as that game is a serious memory hog, for that matter I would suggest a new Video card as well.  Otherwise 1GB of Ram seems to be the sweet spot now a days.  Keep in mind though if you upgrade the Video card to say an ATI X800~850, that will put a serious load on the Power supply.  Which brings me back to your new 400 Watt Power supply, given what you have now it should be ok, buying a new Video card however may put more load than it can handle.

As for what you put into your kids system that will put a load on the power supply, what really matters is the Motherboard/Processor, Video Card, Sound Card (if any) and Ram.  The Floppy and Zip drives shouldn't draw power unless it is actually being used.  The Hard drive, Wireless network card, and DVD drive won't draw that much power compaired to the major components.  The other thing to keep an eye on is how many unpowered USB devices you have connected, if you have more than three or four than you may want to look into a powered USB hub.  As always, hope this helps.

[E_Look]

Hmmm... I see.

Anyway, the game is some Korean server-based online thing a lot of kids apparently are playing these days- MapleStory.

Yes, an up to date video card will tax the power supply, but I've got a Radeon 9600 XT in that box right now and that *SHOULD* not load up the PSU too much at all.

In the other machine that runs without any crashing, there's only a Radeon 9500 with the identical Thermaltake PSU and it runs Call of Duty fairly well; only loading is slow.

[E_Look]

Okay, Jav, and anyone else interested,

It seems that his computer freezes much, much less with the Thermaltake PSU than originally with the MGE one... but it still crashes, just much less often.

So then, I think I've only solved the problem partially.  It probably is a combination of "ripply" power signal (cheapo PSU [again, even if very pretty to look at]) and mismatched RAM.

As I've indicated much earlier, one stick is 512 Mb of Kingston and the other is 512 Mb of Kingmax.

I'll try to run MEMTEST again later tonight with only one or the other, but before I even begin, which brand is more reliable in quality?  I've heard that Kingmax might be better than KIngston.

[Stormbringer]

Okay, Jav, and anyone else interested,

It seems that his computer freezes much, much less with the Thermaltake PSU than originally with the MGE one... but it still crashes, just much less often.

So then, I think I've only solved the problem partially.  It probably is a combination of "ripply" power signal (cheapo PSU [again, even if very pretty to look at]) and mismatched RAM.

As I've indicated much earlier, one stick is 512 Mb of Kingston and the other is 512 Mb of Kingmax.

I'll try to run MEMTEST again later tonight with only one or the other, but before I even begin, which brand is more reliable in quality?  I've heard that Kingmax might be better than KIngston.

when i replaced my fancy led power supply i took the led fans out of the old one and put them in the new one. or you could use them as a case fan.

[E_Look]

Neat idea.  But I won't tell my kid yet.  I don't want to hear clamoring over it for the next umpteen years...

[E_Look]

Okay, Javora, Stormbringer, Maxillus, Tus, CayneG, and any others interested,

I ran MEMTEST with BOTH sticks of RAM still in (one 512 Mb Kingmax, one 512 Mb Kingston) with the new Thermaltake PSU...


... and there were NO ERRORS!!

Now what the heck does THIS mean??

Could it be that all the signal irregularities to the RAM chips were from "dirty" power from a cheapo power supply?

But the system still locks up, though not nearly as much.  His brother's machine, almost identical now, except the mobo is a slightly newer equivalent DFI NF2 AL 400 based board, as the earlier version was no longer available at the time I built this second machine, NEVER locks up.

I don't get it.

[Javora]

Okay, Javora, Stormbringer, Maxillus, Tus, CayneG, and any others interested,

I ran MEMTEST with BOTH sticks of RAM still in (one 512 Mb Kingmax, one 512 Mb Kingston) with the new Thermaltake PSU...


... and there were NO ERRORS!!

Now what the heck does THIS mean??

Simple, it means your Ram is doing ok...   
Unless something radically changes for the worse then I wouldn't worry about the Ram or Power supply anymore, it looks like those potential problems has been eliminated from the equation.

Could it be that all the signal irregularities to the RAM chips were from "dirty" power from a cheapo power supply?

But the system still locks up, though not nearly as much.  His brother's machine, almost identical now, except the mobo is a slightly newer equivalent DFI NF2 AL 400 based board, as the earlier version was no longer available at the time I built this second machine, NEVER locks up.

I don't get it.

That could have been a problem with dirty power feeding the Ram type of thing, but you are still experiencing lockups so something is still going on.  Is there anything else besides the motherboard that is different on these two systems?  If not then (if and only if you are confortable with it) I would suggest swapping the processors if you can.  This way if the system with the newer motherboard stops having problems but the system with older motherboard starts locking up then you could have a bad processor on your hands.  Likewise if you change processors and the system with the new motherboard still has problems then it could be a issue with the newer motherboard.

Then again before doing all of that why don't you check and see if there is a Bios update for the newer motherboard that is locking up on you.  If there is a newer Bios out, it might solve the problems you are having.  Hope this helps.

[KBF MalaK]

i am using a very old, big(160mm, i think) fan that used to cool some industrial equipment where my dad works. It is connected to my PS by paper clips jamed into the power connector and is being run at 7v(best sound/air flow ratio) but can handle 24v. I have to have the case off all the time as the fan sits on top aiming down onto the CPU and RAM. Keeps system below 40c almost all the time. I have a Athlon XP 3200 CPU and Radeon 9800 pumping out the heat.

I was gonna do something like that, but run a duct from the cover opposite the CPU and fan blows into the CPU fan. I'm running 4 case fans now with a second power supply just for the fans, and speedfan, and dropped the CPU core voltage down to .05 volts less than the CPU requires. I can keep in the low 50's with this setup running autocad, and 3dsmax.

I had the greatest temp drop when I reversed the flow of the rear case fan to blow out instead of blowing onto the CPU heatsink. I think before I was just recycling the case air instead of evacuating it like it is now. So now all the front fans pull air into the case, and all rear ones blow out.

I've got a Gigabyte GA-7dxr, dual WDC40gig on the raid, and an ATI radion 7200, cd and dvd burners on the secondary controller, and a Maxtor diamondmax 80gig on the primary.

[E_Look]

... Now what the heck does THIS mean??

Simple, it means your Ram is doing ok...   

Ha ha.  But seriously, you're right, it's a relief on that account

... But the system still locks up, though not nearly as much.  His brother's machine, almost identical now, except the mobo is a slightly newer equivalent DFI NF2 AL 400 based board, as the earlier version was no longer available at the time I built this second machine, NEVER locks up.

I don't get it.

That could have been a problem with dirty power feeding the Ram type of thing, but you are still experiencing lockups so something is still going on.  Is there anything else besides the motherboard that is different on these two systems?  If not then (if and only if you are confortable with it) I would suggest swapping the processors if you can.  This way if the system with the newer motherboard stops having problems but the system with older motherboard starts locking up then you could have a bad processor on your hands.  Likewise if you change processors and the system with the new motherboard still has problems then it could be a issue with the newer motherboard...

Ah, I don't want to do any CPU swapping... yet.  Both chips are still under warranty, as I bought the retail kits that came with the manufacturer (AMD) provided heat sinks, and I don't want to invalidate it.

... Then again before doing all of that why don't you check and see if there is a Bios update for the newer motherboard that is locking up on you.  If there is a newer Bios out, it might solve the problems you are having.  Hope this helps.

Yeah, I think I will look for the latest BIOS available for the mobo; I *MIGHT* have forgotten to check that after assembling the system...    

Thanks for your extensive and well thought out advice!

+1 for your trouble!

[Javora]

No problem Ed, I'm here for you and anyone else that asks for help.  Everyone seems to contribute here in one way or another and this seems to be the best way for me to do that.

[Nemesis]

Yeah, I think I will look for the latest BIOS available for the mobo; I *MIGHT* have forgotten to check that after assembling the system... 

One suggestion.  Before flashing a BIOS reset to defaults.  I have known of cases where flashing a BIOS that had non default settings resulting in a failed flash and non bootable computer.

[E_Look]

Hokey schmoke!  There's another old nugget of wisdom I have completely forgotten in my doting old age!

Thanks, Nem.  I haven't seriously tried to troubleshoot systems for a long time until recently and I truly appreciate, and I'm fairly sure others reading this thread appreciate, all this truly good advice.

[Javora]

Ah, I don't want to do any CPU swapping... yet.  Both chips are still under warranty, as I bought the retail kits that came with the manufacturer (AMD) provided heat sinks, and I don't want to invalidate it.

Removing the CPU from the motherboard shouldn't void the CPU warranty.  CPU companies have to allow for the possibility of the motherboard going bad.  Unless CPU companies are going to cover the motherboard as well.  One thing though when you installed your CPU did you use thermal paste or did you use the thermal pad?  If you used the thermal pad then you may not want to remove the CPU as the thermal pad turns into glue and can harm your CPU if you try to remove the heat sink.  So if you did use the thermal pad and have to remove the CPU for some reason be very careful.

[Nemesis]

Hokey schmoke!  There's another old nugget of wisdom I have completely forgotten in my doting old age!

Thanks, Nem.  I haven't seriously tried to troubleshoot systems for a long time until recently and I truly appreciate, and I'm fairly sure others reading this thread appreciate, all this truly good advice.

I have had a couple of other "anomalies" when working with machines whose BIOS settings had been tweaked for performance.  (My own machines so they were self inflicted  ).   The non booting after failed flash has not happened to me.  These days I am a little paranoid and won't even flash the BIOS unless the system is on a UPS as a failed flash due to power outage can leave your machine down until  you can get a new BIOS chip.  At least one manufacturer (gigabyte I think) puts 2 copies on your system, a master copy that you can't change and a flashable version.  That allows you to boot off the master copy even if you totally foul up the flash or use a "unauthorized" BIOS from a 3^rd party.

One had originally had a Pentium 90 (overclocked to 100MHz) which according to Norton was running like a Pentium 177MHz.  Later I swapped the CPU for a K6-2 400MHz (no overclock) ultimately I went to install Linux.  Part way through the install crashed repeatedly.  So I tried to reinstall Windows (98SE) and the same thing happened.  Eventually I reset to defaults and was able to easily run either install no problem.  I never did find out exactly what setting (or combination) caused the problem.

Later with a my 1^st Dual CPU machine I was reinstalling Win2k Pro and needed to install a driver off the floppy for the HD controller.  The floppy could be accessed but not read.  Since I had been able to write the floppy before beginning the reinstall I knew it worked.  Finally I did the reset to defaults and it worked once again.

The moral of the story of course is if your machine is acting "funky" consider resetting the BIOS to defaults.  Even if you think there is nothing odd in your configuration to cause the trouble you may be in for a surprise.  So you may wish to try that with your machine. Nothing to lose but a little time zero cash outlay and no real risk.

[E_Look]

Hey Jav & Nem & Kid Carrson & Tus & all else!

I just flashed my son's BIOS.  I'm typing from that rig right now.  I did as suggested, resetting the original BIOS to defaults... one interesting hiccup, though: it disabled the 3.5" floppy A: drive!  I had to reboot, reaccess the BIOS and reactivate the A: drive (what a pain!).  Apparently, after that however, it went off without a hitch and haven't had any lockups yet.  Funny though, when I first booted it up tonight and opened the browser to go to the mobo mfg's website for the new BIOS, guess what... IT CRASHED!

So far, it's been at least a half hour if not more since flashing the BIOS and no trouble yet.  The real acid test will be when he logs onto his usual online game in a couple of days.

But I guess the combination of a high quality power supply and an updated BIOS firmware is doing the trick!

Thanks again, all of you, including all unnamed ones!

EDIT:

Oh well.  Five minutes ago, about an hour into the new BIOS, the system froze.  It IS much less crashy than it was before the PSU change, however.

Maybe I *should* seriously consider matching the RAM (i.e., all from the same manufacturer), although MEMTEST said all was okay with the RAM.  *sigh*

But... as I said, the real test is this weekend when he gets to play for a more extended time than my light testing here.

Okay, another edit:

I just realized that maybe ATI might still have a bit of those bad old days in them, so I just d/l'd their latest drivers and fluff from their website and installed them.  Let's hope this one nails the bogeyman in the machine.

[Tus-XC]

how hot is the cpu running anyway?  idle will do nicely for this.  When it crashed, was it under load?  have you check the hd for errors (didn't read all the way through so i don't know)? 

One problem i have had in the past w/ my rig was overheating, had to replace the stock heatsink cause in game it would be at a constant 75-80*c .  I've also seen computers run great, but crash for no reason cause the HD was going bad.  Since it appears that your ram is ok, and your psu it good i think that narrows it pretty much down to the HD, heat, and/or vid card (if it crashed while in game then that could be a possibility).    if you could next time you get it up and running post the cpu temp, possibly the mb temp.  if you have anythign that tracks the temps post the underload temp as well.

[Dracho]

Topic: What do you do to keep your computer cool when gaming?

Generally, I play crappy old games that do not tax my system.   

[Darth Sidious]

I keep my gaming computer cool.... by keeping the windows open in winter; and the AC on full tilt all summer.

Then again i live in MN; and am crazy and LIKE -40 temps

[E_Look]

how hot is the cpu running anyway?  idle will do nicely for this.  When it crashed, was it under load?  have you check the hd for errors (didn't read all the way through so i don't know)? 

One problem i have had in the past w/ my rig was overheating, had to replace the stock heatsink cause in game it would be at a constant 75-80*c .  I've also seen computers run great, but crash for no reason cause the HD was going bad.  Since it appears that your ram is ok, and your psu it good i think that narrows it pretty much down to the HD, heat, and/or vid card (if it crashed while in game then that could be a possibility).    if you could next time you get it up and running post the cpu temp, possibly the mb temp.  if you have anythign that tracks the temps post the underload temp as well.

Ah!  More good questions.

It is an Athlon XP 2500+ chip with a Barton core, so it runs about 40-50 C, but my memory is foggy on this.  I never checked the difference between under load and idle conditions.  I'll have to do this, but not tonight; I'm beat tonight.

The HD is essentially new and recently defragged.  Maybe I should run scandisk on it.

It could be the vid card; I got it, a Sapphire Radeon 9600 XT from Monarch Computer for $98 USD when I saw it on Newegg, Zipzoomfly, etc. for about $120 or so.  But I haven't heard any bad scuttlebutt on this card, though.  I just d/l'd and installed the latest drivers for it.

It crashes not only when gaming, but also when surfing the 'Net, accessing drop down menus, using MS Word, etc... in other words, randomly.

[Midnight Tech]

E_Look, got another idea...download an ISO for Knoppix and burn it to a CD, then boot with it. Try running with it and see if you get any lockups.
Sometimes it is a bit tricky to get it to start properly in X-Window...it might require a couple boots...but most of the time you'll get full access to all hardware so you can troubleshoot things out.
If you don't get any lockups running Knoppix, then it's probably gonna be something that Windows doesn't like.

[Tus-XC]

if its between 40-50 idling, then my best guess (based on my own processor) its closer to 50.  depending on what kinda heat sink and fan system you have you can expect anywhere from a 10* difference to a 20* difference while under load.   With my stock fan it was 20 degree difference, went from 55* to 75* causing my computer to shut down.  I replaced the heatsink/fan with a one from cooler master (looks like an impeller ontop of a copper heatsink) and it idles between 35-40, and under load never goes over 50*. 

[E_Look]

Knoppix??  I haven't touched UNIX systems in a veeerrrrry looooonnnnng time!  But in those days of command line interfaces, it didn't seem that bad though... 

[Nemesis]

Knoppix??  I haven't touched UNIX systems in a veeerrrrry looooonnnnng time!  But in those days of command line interfaces, it didn't seem that bad though... 

Knoppix defaults to the KDE gui and is very windows like.  It also includes a number of Windows recovery tools.  I recently had an issue with a friends computer solved just by booting Linux.  It caused the motherboard to reassign irqs.  Then when windows was rebooted it reassigned them again fixing the conflict that Windows could not see until the irqs were reassigned first.

Have you tried System File Checker?  It is a utility that is built into Win98 and Win2000 (I assume it is also present in WinXP. It checks for corrupt system files and lets you restore them from the install CD.   On Win98/2000 you use the run option on the start menu and run SFC.exe.

If that fails I think that you have reached the stage where you need to pull half the memory and try running without it.   Once you know if the problem still exists or not you can try the other one.  If you have the problem with both individual memory units then it is likely not a memory issue.  If you have it with one unit and not the other then you know which unit is faulty.  If with neither then it is a conflict between two incompatible units and you should look at getting identical ones.  All this of course assumes that your motherboard does not require filling two banks at a time (some do). 

Alternately since you have more than one machine you could try a memory swap between machines (assuming they use the same memory types).  If you do and the problem persists again you have eliminated memory as an issue. If it migrates to the other machine you know that it is bad (or incompatible memory).  If it is eliminated then you just have a motherboard that can't handle that particular memory.

Right now I have my own cooling problem.  My #2 dual CPU machine has a failing CPU fan.  Of course I find that out at a moment in time where it will be difficult to get one for a day or so .  Fortunately I have two other machines to play with though my SETI output is suffering   until it is fixed.

[E_Look]

Knoppix??  I haven't touched UNIX systems in a veeerrrrry looooonnnnng time!  But in those days of command line interfaces, it didn't seem that bad though... 

Right now I have my own cooling problem.  My #2 dual CPU machine has a failing CPU fan...

At least you KNOW the problem and CAN fix it!

If that fails I think that you have reached the stage where you need to pull half the memory and try running without it.   Once you know if the problem still exists or not you can try the other one.  If you have the problem with both individual memory units then it is likely not a memory issue.  If you have it with one unit and not the other then you know which unit is faulty.  If with neither then it is a conflict between two incompatible units and you should look at getting identical ones.  All this of course assumes that your motherboard does not require filling two banks at a time (some do). 

Alternately since you have more than one machine you could try a memory swap between machines (assuming they use the same memory types).  If you do and the problem persists again you have eliminated memory as an issue. If it migrates to the other machine you know that it is bad (or incompatible memory).  If it is eliminated then you just have a motherboard that can't handle that particular memory.

When I get some time, I will try this one.

... and run SFC.exe...

Yo!  You mean go blow up some starships?!  Sure!  I never knew that fixed computers!!  Gee, thanks!  I'll go install all four Starfleet Commands on my kid's comp right now!   

[Nemesis]

... and run SFC.exe...

Yo!  You mean go blow up some starships?!  Sure!  I never knew that fixed computers!!  Gee, thanks!  I'll go install all four Starfleet Commands on my kid's comp right now!   

That is an order that I can give here and know it will be obeyed.

Go forth and get on tonights test server

[Nemesis]

At least you KNOW the problem and CAN fix it!

Fixed.  I have a couple of old heatsinks and fans for slower CPUs, though the heatsinks were not up to spec I realized that the fans were virtually identical.  I checked and the old fans fit the newer heatsinks.  Testing now.  The formerly overheating CPU now runs cooler than the other CPU.

[Midnight Tech]

That's good to hear Nemisis! I had to cannibalize an old Socket 370 heatsink for a fan to fit a (gasp) Slot 1 Pentium 2 the other day...running like a charm now!
Knoppix can be booted into command-line mode as well...but with all the software offered in GUI mode there's not much reason to.

[Nemesis]

The repair has been stable for more than 24 hours and the computer has done 13 SETI WUs since then.

[Javora]

If that fails I think that you have reached the stage where you need to pull half the memory and try running without it.   Once you know if the problem still exists or not you can try the other one.  If you have the problem with both individual memory units then it is likely not a memory issue.  If you have it with one unit and not the other then you know which unit is faulty.  If with neither then it is a conflict between two incompatible units and you should look at getting identical ones.  All this of course assumes that your motherboard does not require filling two banks at a time (some do). 

Alternately since you have more than one machine you could try a memory swap between machines (assuming they use the same memory types).  If you do and the problem persists again you have eliminated memory as an issue. If it migrates to the other machine you know that it is bad (or incompatible memory).  If it is eliminated then you just have a motherboard that can't handle that particular memory.

When I get some time, I will try this one.

Didn't you already go through all of that already?  *Looks down the thread a ways*  Yep, you already went through all of that and used memtest on top of all of that.  I think you are at the point now where using Midnight Tech's idea of booting Knoppix or swapping CPU's might be a good idea.  If you don't feel confortable about swapping CPU's, you could take it in to a repair store that you can trust and let them test the motherboard and CPU for you.  You could also try matching pairs of Ram but IMHO that would be a waste of money, but if you are using mismatched Ram make sure they both have the same power/usage ratings.

[Nemesis]

Didn't you already go through all of that already?  *Looks down the thread a ways*  Yep, you already went through all of that and used memtest on top of all of that.  I think you are at the point now where using Midnight Tech's idea of booting Knoppix or swapping CPU's might be a good idea.  If you don't feel confortable about swapping CPU's, you could take it in to a repair store that you can trust and let them test the motherboard and CPU for you.  You could also try matching pairs of Ram but IMHO that would be a waste of money, but if you are using mismatched Ram make sure they both have the same power/usage ratings.

I just looked through the thread and though this was suggested before it does not appear that E_Look has tried it yet.  He did do the memtests though.

[E_Look]

All right, folks, I am running Memtest on the second stick of RAM, the Kingston 512 Mb DDR-333 (PC-2700) stick and is so far totally free of errors.

The first stick, Kingmax 512 Mb, was also free of errors according to Memtest...

... BUT I NOTICED A VERY STRANGE THING DURING BOOT UP WITH THAT ONE STICK AND DURING THE MEMTEST RUN- IN BOTH INSTANCES IT WAS READ BY THE COMPUTER AS DDR-400 (PC-3200)!!! while the sticker and packaging (and my invoice) read DDR-333.

I'll bet that the timing and rate mismatches will lead to *a little* system instability.  This is the strangest PC building problem I've seen!

Let's see if I can get it exchanged for a stick of DDR-333 or if they'll be kind enough to offer me an extra stick of DDR-400 for all my (my son's, really) frustration!

I'll keep you guys posted.

[E_Look]

I called Newegg a couple of days ago and got a RMA number.  They gave me a credit for shipping, but it is apparently for Fed Ex and I used U.S. Mail... but that's minor.

The main thing is that I shipped it off, the mislabeled RAM chip module, and upon receipt, they promised to send me a replacement, which I will this time IMMEDIATELY test, not only for errors of the type we normally might expect, but also for "bad stickers"    .

I must say though, in general, Newegg is great to deal with!

[Nemesis]

How is the computer working with the "mere" 512mb left?  Stable?

[E_Look]

Well, tonight he gave it a whirl with the "miniscule" 512 Mb and at the very outset, it lagged a little playing his online game, but that stopped quickly and didn't crash once during a 30-45 min run.

I saw on Newegg's site one of the customers *CHEERING* that his stick of DDR-333 Kingmax 512 Mb was read by his mobo as DDR-400; I guess if you have two modules of these, then it IS cause for a little joy... but if you mix'n'match, then, well, your kid cries, kvetches, moans, and otherwise makes known his unhappiness.

Anyhow, I am SINCERELY HOPING that this replacement stick I am to receive will behave exactly as a DDR-333 module should behave.

[E_Look]

Hey, guess what... the system still crashes, say, once or twice an hour or two, with the one stick of Kingston 512 Mb in it.  I think I'm just spinning my wheels...

... 

[E_Look]

Okay folks!

I finally ended the crashes on my son's computer!  No, I didn't phaser the stupid machine...

... a friend told me that he once had an issue with a new motherboard not being able to handle its fastest rated data speeds at its FSB because of a blown cap on the board.  It appears that some of these capacitors' function is to smooth out any spikes and irregularities in this signal and without ALL the caps, it cannot smooth out the signals at the highest FSB speed.  So he tweaked it down a tad and had no probs ever since.

So I tried it.

So it works!

So now I'm only 90% happy.

This is because, while the whole system no longer hangs, it runs at less than max transmission rates.  I had to reduce the FSB speed in the BIOS from the preferred 166 MHz to the current 133 MHz.

I know, I know, the difference is miniscule and should affect nothing... but the thought of it!

[Javora]

I'm glad you found the solution Ed, although personally I would return the motherboard for a replacement.  Eventually that cap will blow leaving you with a dead motherboard.  Save yourself the aggravation and money and replace the motherboard while it is still under warranty, since the motherboard is not working as promised you have some recourse.

[E_Look]

Hmmm...

... Jav, your point is totally valid... and strong.  I only don't want to put up with the "Aww!"s, the "Dad! How'm I gonna play ____??"s, the "How long??!!!"s. 

But I see what you're saying.  I will certainly consider disassembling it and shipping it back.

Thanks, really!

[Dracho]

It may blow your ram and processor when it goes (or even hard drives).  Replace the board.

[E_Look]

Now there's another, soli... HARD reason!!

I'll try to break it to the kid as gently as I can.

[Javora]

Hmmm...

... Jav, your point is totally valid... and strong.  I only don't want to put up with the "Aww!"s, the "Dad! How'm I gonna play ____??"s, the "How long??!!!"s. 

But I see what you're saying.  I will certainly consider disassembling it and shipping it back.

Thanks, really!

Why, if you can make the company send you a new motherboard first and then send back the old motherboard after the new motherboard is installed.  I realize some companies won't let you do that but it can't hurt to ask.  Keep in mind though the company may ask for your credit card number and give you a time limit on when to get the old motherboard back.  IMHO I don't like companies that make you send the old motherboard back first.  These companies have a bad habit of returning people the same motherboard back saying it's refurbished when the company haven't even looked at it.

[E_Look]

Whoa!  I didn't think of that...

... I'm glad you mentioned it, Jav.  I'll e-mail 'em back.