Oh, the humanity....
I recently upgraded my main system to an athlon Xp3000+ CPU, mounted on a Gigabyte GA-7n400 Pro 2 motherboard, 785 MB PC 3200 DDR RAM, NVIDA Geoforce FX 5600 Ultra Video Card, yadda yadda. Swell. This is a gaming system and this morning I was playing Halo, happily kicking some ass when upon finishing I noticed my CPU core had been consistently been running at 65C for some time. Not disastrously hot, but far too hot for a machine that is meant to be on all the time. I decided to replace the OMD Heatsink/fan with a Cooler Master Jet 7. Well, there had been tape - no thermal grease applied with the original heatsink (For the first time in my life I let someone else build the machine). I decided to apply some Arctic silver 5 to the CPU core, as reccomended. Well, don't ask me how this happened, but for whatever reason - perhaps I pressed a wee bit too hard - SPOINK! A HUGE blob of the infernal goo fell out, more than covering the core, in fact an alarming portion of the CPU itself.
Okay, no sweat. I cleaned the crap off as best I could with iso, Q Tips, lint free cloth, the usual. I have a Duron I picked up from a friend who slathered grease over the sides of the core like pancake batter, and after a thorough cleaning it works normally. Well, I PROPERLY reapply the grease, install the gear, fire up the system, and after perhaps 20 seconds, the 'CPU Error' beep starts, and the system shuts off.
Mildly alarmed, I assumed the problem had something to do with heat, since it had operated normally (if a little hot) before the tube of Arctic Silver vomited all over it. I turned it on again after a few minutes, and entered the BIOS setup just long enough to see the CPU core temperature reach 98 Celsius.
That's right, I could have boiled myself a cup of tea on that bad boy. I screamed like a girl and was just reaching out to yank the cord out of the wall when it mercifully shut itself down.
I have again cleaned the CPU as best I can, but you can clearly see discoloration where the compound has stained the surface of the CPU. Two subsequent attempts to operate the CPU have yielded the same result after repeated cleanings. I enter the BIOS, check the core temperature and have to shut down immediately because the core is running hotter than a supernova.
And BTW, I feel I can effectively rule out the remaining system components. I have since slapped on the aforementioned Duron 750 to the motherboard and am writing this post on that configuration. I just won't be playing Halo with it.
My poor, poor Athlon.
Anyone have any ideas? I suspect I am going to have to just pony up the C-Bills for a new CPU. Any suggestions, solutions or thoughts would be appreciated.
Okay, no sweat. I cleaned the crap off as best I could with iso, Q Tips, lint free cloth, the usual. I have a Duron I picked up from a friend who slathered grease over the sides of the core like pancake batter, and after a thorough cleaning it works normally. Well, I PROPERLY reapply the grease, install the gear, fire up the system, and after perhaps 20 seconds, the 'CPU Error' beep starts, and the system shuts off.
Mildly alarmed, I assumed the problem had something to do with heat, since it had operated normally (if a little hot) before the tube of Arctic Silver vomited all over it. I turned it on again after a few minutes, and entered the BIOS setup just long enough to see the CPU core temperature reach 98 Celsius.
That's right, I could have boiled myself a cup of tea on that bad boy. I screamed like a girl and was just reaching out to yank the cord out of the wall when it mercifully shut itself down.
I have again cleaned the CPU as best I can, but you can clearly see discoloration where the compound has stained the surface of the CPU. Two subsequent attempts to operate the CPU have yielded the same result after repeated cleanings. I enter the BIOS, check the core temperature and have to shut down immediately because the core is running hotter than a supernova.
And BTW, I feel I can effectively rule out the remaining system components. I have since slapped on the aforementioned Duron 750 to the motherboard and am writing this post on that configuration. I just won't be playing Halo with it.
My poor, poor Athlon.
Anyone have any ideas? I suspect I am going to have to just pony up the C-Bills for a new CPU. Any suggestions, solutions or thoughts would be appreciated.
0
Comments
~dodo
For some reason the heatsink isn't absorbing and dissapating the heat properly ...is the fan turning at all?
2. You haven't got something installed right. There may be a thin plastic film on the bottom of the coolermaster; they've been known to do that. Also, you may just not have it installed properly...
Oh yes, as I said, I am running the exact same system with the Duron at this moment, and the XP 3000 was working properly before. I am vexed.
Well, I appreciate your quick response, but...
1. I sort of expected the 'some/all of your gear is crap' response from someone.
Name a piece of equipment and I guarantee someone somewhere will swear it is the biggest piece of junk ever made, regardless of how many thousands of people use it without incident. I'm really looking for positive suggestions here.
2. As I mentioned, I am presently sitting in front of the exact same system running with a different processor.
I don't mean to be immodest, but it would appear I installed everything properly. I've done this quite a few times before.
And...as I mentioned...the XP3000 WAS originally running properly. I suppose I am trying to discover if anyone else has had problems with excess use of thermal grease, and if anyone is aware of a remedy.
Thanks for your response.
That's the strange thing though...I am using the exact same gear right now, just with a different (although inferior) processor. I cleaned the bottom of the coolermaster quite thoroughly, and tried that as well as the OEM fan with the Duron. I am currently writing this on the same system with the Duron and the Cooler Master.
And the Xp3000 was working fine before the incident...but now as I said, it heats up FAST. I wonder if anyone has had similar experience after using an excess of thermal paste? If so, has anyone found a good remedy, or should I just chalk it up to 'never sneeze while applying thermal paste'? Heh heh...
Have you pulled the heatsink just to be sure that it is in fact making contact?
By the way, you'll find if you read my older posts and/or hang around for a while that I only say that something is crap if it is. The Coolermaster "Aero" heatsinks are not inherently flawed, but the fans they use are. They use dinky little squirrel cage fans, which is not a good thing. Squirrel cage fans push less air, but at much higher air pressure than regular axial fans. In fact, suitably-sized blowers significantly outperform high-flow axial fans on heatsinks such as the SLK series that have dense, thin fins.
However, the fans that Coolermaster uses push very little air; in fact, the airflow is low enough that the additional air pressure provided by a blower makes little difference. So, the Aero HSFs do not qualify as serious high performance cooling systems.
I can't say I have read many previous posts from anyone here yet, since I just stumbled here thanks to a Google search, but I assure you I have bookmarked the site.
One of the last things you said is closer to what I was looking for. I would think that although that grease bonds pretty well, a thorough cleaning would remedy this for the most part, although the Arctic Silver manufacturer does warn of this possibility on their website.
As I mentioned, the Duron I am using now was the victim of a similar - though intentional - overapplication of grease and is humming along nicely at the moment.
One would think, however, that returning to the OEM setup would at least obtain different results. Although it is quite possible the CPU is now permanently damaged, I am a little perplexed at how and why it heated up so quickly.
I am literally looking at something I have never seen in years of working on computers. This is about the only type of failure I have never seen, a CPU that heats up like that with a heatsink, fan, etc functional and installed correctly.
I wondered if perhaps there was an internal mechanism for the CPU that helped regulate voltage flow, that may have been damaged...but for the few moments it was on, the voltage readings appeared normal.
You are, of course, quite right about the airflow properties of that coolermaster fan - I noticed the RPM was somewhat lower than the OEM fan. However, it seems somewhat unlikely that use of the device would consistently lend itself to an immediate and catastrophic increase in core temperature of THAT magnitude. I could understand the CPU running a little hot over time, but I can't imagine that if everyone who used one of these experienced such a radical and fatal spike in core temperature they would still be avai;able in retail stores.
For your CPU core temperature to rocket up to 200F within ten seconds of turning on the machine you would think no sink or fan at all was attached.
So, though it might be impossible to save THIS Athlon.....*sniff* (I am coming to terms with it...it was SO fast...), I suppose I am doing a bit of troubleshooting in advance of replacing it.
Something just doesn't add up here. I guess we're throwing spaghetti at the wall, but the great thing about forums is that you gain access to many other people who might have had similar experiences.
If you ask what to use on an XP around here, you will get almost universal praise for Thermalright's SLK series of heatsinks as they are fairly reasonably priced and are among the very top tier in air cooling an XP.
The other thing that has happened to me, and I'm sure others in the past, is that when putting the cpu into the motherboard, it doesn't fully 'seat' properly. Due to a very slight misalignment, some pins go all the way down and some dont. This creates a slight angling of the processor in relation to the heatsink, rather than flush parallel contact. I've had that be a problem with cpu->heatsink contact on more than one occasion. Fortunately, I always do what I stated in the first paragraph and so could catch this problem before it was an issue too.
Forgive me if I'm talking 'below your level', I do see where you've gone thru system builds on many occasions in the past, truly just trying to help.
No problem. Say, by the way... did I mention that your problem is that all of your gear is crap?
Anyhow, I killed my 1.4GHz thunderbird when I put my SLK-800 on it; I used both a shim and the 4 foam pads the SLK comes with.
I also have seen two computers from a company my grandfather's business used to buy computers from that had the heatsinks installed bass ackwards. The buttheads put the notched side of the heatsink over the CPU, instead of over the socket. If you've accidentally done something like that, that would explain the temps.
I assume you've removed <a href="http://www.gruntville.com/images/hardware/cm_jet7/heatsink_warning.jpg">this</a> thing from the base, right? If not, that would also explain the temps.
But I can't believe that the CPU is damaged from excessive thermal compound. Running @ 100*C may have killed it, but I doubt the thermal compound did anything.
Also, this is really an exceedingly stupid thing to say, but because I don't know how much you know about computers, and because I've seen instances of this before, you do know that the cpu core is the rectangular thing in the middle of the CPU, right? Since you bought AS5 and an upgraded heatsink, I assume you do, but some maximum pc reader sent a picture in a while ago of an athlon that'd had thermal compound applied to the 4 foam pads instead of the CPU core, so you can never be too careful...
Your best bet would be to pop the heatsink off (again), re-clean it and the 3000, and put it back on.
Also make sure the fan is set at maximum speed.
And I re-read your original post; For some reason I thought it was an Aero 7/Aero 7+; I hate to sound like I'm giving you a hard time because you just signed up today, because I'm not, but the Jet 7 is honestly an awful heatsink. The reviews I've seen indicate that it's got awful performance. Admittedly, I don't consider those results to be 100% accurate for the most part, but they're in the ballpark (somewhere in it, anyhow), and the Jet 7's performance is the worst I've seen from a "performance" heatsink in some time...
If removing it and re-seating it don't help, you may simply have to upgrade the heatsink. Coolermaster rates it for a 3200+ and higher, but they also say it's got a thermal resistance of ~0.5*C/w. The 3000+ puts out ~75w maximum, so even if it WAS working properly, in a normal 20*C/68*F room, you'd still be seeing maximum CPU temperatures of around 57*C at least. So, even if it was working, it still wouldn't be able to keep the 3000 at what I would consider an acceptable temperature.
MD...thanks so much for your interest in my post. I appreciate your feedback!
As I have discussed this in a few other forums tonight, the subject of the CM HS/Fan combo has come up...and naturally you are quite right about the differing power requirements of the Duron and the Athlon Barton, not to mention the slightly inferior performance of the CM setup.
But it seems peculiar that such an immediate and catastrophic accumulation of heat would occur so suddenly in the CPU core, even with this setup. I can see the CPU slowly heating up over time, but of course, any contemporary power user should be monitoring his/her voltage/temp operations pretty closely and be able to compensate.
I guess the reason I am so focused on the thermal goo is because it was not until that point the failure occured, and it happened with the OEM AND the CM setup.
Here's the strange thing...please do not hesitate to correct me if you know differently, but a CPU is either dead or not. What I mean is that if i applied the CM HS/Fan and it killed the CPU, I should not have been able to go back to the OEM setup and run the processor at all. The fact that I can - even if it is for a few moments - monitor the core temp and voltage (which was normal!) means the CPU is not dead, but perhaps damaged or somehow not getting adequate power.
If this is true, my only concern is this:
Perhaps the processor was just bad? It was originally running with normal voltage readings, just a sligtly elevated core temp. based on discussions I have had tonight, the most commonly experienced temp range for the Athlon Xp3000+ Barton core (even with OEM cooling) is between 42 and 55 C.
I realize the Duron I am using as backup now has different operating requirements, but is seems unlikely that even a less than stellar Hs/Fan like the CM Jet could allow the core temp to spike to 98C within ten seconds of applying power. That's around 200F!!!
The sort of thing you would see were one foolish enough to turn on their system with no cooling at all!
Basically, I realize I am probably beating the issue to death, but the thing is, of all the fatal errors I have seen take down a system, such a massive and sudden heat spike has never been one of them. It doesn't add up. This isn't the first Athlon, or barton I have had, and i would think that You could monitor a faulty cooling system over time. I also want to try to come as close as I can to pinpointing the failure before I replace the processor.
My trusty Duron 750 has served me well for a long time, but it is only a backup!
I am also addicted to troubleshooting! It is part of my profession!
Let me ask you this...although it has been said that OEM cooling, specifically on AMD processors are inferior, I have not had trouble with my Xp2400, my XP 1600, or the trusty Duron, despite their differing power requirements.
I considered going with a liquid cooling system before I bought the CM. Frankly, here's why I bought the CM despite the conspicuously low fan RPM...
It pulls air to the side rather than blowing it downward. Many will disagree with me but I will insist this makes more sense than blowing the largely still and hot air in the case downward onto the CPU. I set it up so that it works in tandem with my primary chassis fan in the rear, and they both carry warm air out of the chassis.
I will certainly look into the Thermalright, and thanks for the reccomendation...but I will admit i have never used liquid cooling before. Have you had experience with this?
Not at all...the fact that you responded at all is gratifying! Actually, I am getting lots of great feedback and appreciate all of it.
Here's my angle: Yes, I have been fooling with PC's as a hobby and professionally for a long time, but nobody can know everything and I will tell you...the first PC I ever tried to build from scratch (although I had taken many things apart...including my Atari 2600 that i got for Xmas as a child!) was a 386 SX16.
I blew it up. I even got arcing. It was a beautiful blue glow...fantastic.
Ever since then, there are several things I ALWAYS check for when working on rigs...motherboard standoff seating and screws...(in other words, no shorts!) , CPU/heatsink (at least since CPUs have used them!) seating, etc...and general component compatibility.
However, I find the longer you do something the more likely you may be to overlook something stupid. So no, I am very gratified to know I am not the only one who looks into these things. Thanks so much for your input!
To answer your questions, I did check the CPU seating...of course, as my PC Sinsei tild me years ago, only the CPU truly knows what's going on in there!
However, I feel pretty confident that the basics were covered.
However, since you were kind enough to respond to me, please allow me to avail myself of your experience...
The reason I am focusing on the AS5 is because before the goo incident, the 3000 was working fine, if slightly hot.
Perhaps the CPU was simply faulty...used to happen to AMD a lot back in the day, but I have not had a bad AMD CPU since...oh...1997?
I guess I am ready to assume the CPU was gimpy, since even when it was running 98C core temp the voltage was normal. However, i am wondering if I can't pinpoint as close as i can to the failure before i replace it.
I had an alternative air cooling system reccomended to me, but I am starting to wonder now if liquid is the way to go? I have never used liquid cooling, and the reason I went with the Cm fan was because it pulls air over the CPu rather than blows it down.
This may seem trivial, but when used in tandem with an intake fan in front pulling at 3200 RPM, and an outtake fan in rear at the same RPM, I am currently maintaining an ambient case temp of 37C. I realize the Duron 750 has differing power requirements than the XP3000+, but the lower fan speed of the CM is not an issue in this setup, and it seems odd that such a major and fatal increase in core temp could happen so suddenly just from that one issue...
In any event, i don't suppose you have had any experiences with liquid cooling?
Barring that, just for the hell of it, try clearing the CMOS and see what happens. You didn't happen to get any AS5 on the bottom of the CPU, did you?
Also try picking up some Goo Gone; it'll remove any AS5 residue that is still on the CPU, and going over it with rubbing alcohol will remove the residue from the Goo Gone. If the AS5 is doing anything to the CPU, that should fix your problem.
Are you sure that the heatsinks aren't hanging up on the socket ledge where the socket pivots when working the lever to remove or install the proc? I have had that happen before which gave me only partial contact with the core and my temps would skyrocket fast(but not as fast as with you though). I imagine that it is possible that the proc could be damaged somehow but I've never heard of heat going out of sight that fast due to either too much thermal grease or possible heat damage to the proc.
BTW, if you want to give that 3000 a thorough cleaning, use mineral spirits first to remove that Arctic Silver, then wipe it down with alcohol. The mineral spirits(paint thinner) will really cut that AS off the proc and the wipedown with alcohol will clean any residue from the mineral spirits off.
If the proc still acts up, I would RMA the bugger back to AMD, since it sounds like you bought the retail version with an oem heatsink. They should replace it with no problem since you weren't overclocking. Just don't mention that you used thermal grease as AMD says you need to use a thermal pad instead.
One other thing you might want to do is to set the thermal cutoff lower in bios, if your motherboard supports that kind of adjustment. I set mine at around 75 C so that I don't have to worry about cooking my procs.
With the SLK series of heatsinks that I'm using on my XP's(SLK800 and SLK900a models), my temps rarely get close to 50 C even when the proc is loaded at 100% running the Folding@Home DC project on them. I've used quite a few high end air solutions and I find that the Thermalright SLK heatsinks are by far the best, especially when considering the performace versus price. If you look around, you can find pretty good deals on them.
By the way... this is bugging me, so I've got to ask:
Your screen name have anything to do with <a href="http://www.amazon.com/exec/obidos/tg/detail/-/0345404475/qid=1073204683//ref=sr_8_xs_ap_i0_xgl14/102-8047881-8792151?v=glance&s=books&n=507846">"Do Androids Dream of Electric Sheep?"</a>
Heh heh...yeah, I removed the warning tape, and applied the AS5 to the core, rather than the foam standoffs...but I totally understand why you would ask!
That's a great point you made about the resistance of the Cm...but hey...I could live with a consistent core temp of 57C..that is a little high...but I could live with it. In fact, my aforemtioned Duron is running 55C now, and I have a number of applications running...granted, as we have discussed, it has lower power requirements, but considering the fact that I routinely have a ridiculous number of things going on at once on my Pc and my network...I can handle it.
But no, you are right it is the jet 7 I have...but I think I will try to reclean the CPU and OMD setup and try that.
I DID look at the Jet to see if the notch was on the wrong side, all ready to commit ritual suicide...but no, it was correct...*LOL*
I think I might just have a gimpy CPU....Before returning the CPU to AMD I will attempt to use it with a better HS than the OEM or the CM....
After i blasted it and used a splintered toothpick (yeah I know, sounds silly) to dig around in them to make sure it was all out it fired up like a charm and I never had a squeak of trouble out of it.
Geeky I will tell you I am as skeptical as you are about the CPU...I just haven't seen a bad AMD CPU since about 1997. Not one.
And I have NEVER had a bad Intel CPU. Ever. In fact, no techs I can recall have ever had a bad Intel CPU.
To tell you the truth, were this my forum I would probably have concluded that the tech installed the HSF wrong, but as i told you, although I will admit I am not above doing stupid things, I just turned over that CM and the AS5 paste is on the correct side from the last installation.
BUT...as you mentioned, if i can POST at all with the 3000, there's hope, so perhaps later today...when it is not 0200 I will run out and snag another HSF. Otherwise, I suppose I can return the CPU.
Basically, I think it boils down to this...
I have not encountered a bad CPU (Well, Intel or AMD...now Cyrix....) in so long that it would literally stun me to run across one once in 100 installs.
I bought a bad mother board recently, and it was an ASUS...I don't think I have ever even HEARD of a bad ASUS motherboard...though I am sure someone somewhere will insist they are junk...*LOL*
I actually had this Duron on an A7V for several years...it was literally the best MB I ever had. In fact, it goes back on that board after this Gigabyte is satiated.
Thanks so much for the input...especially the input I got on the liquid cooled systems...being the Uber Geek I am I may just take the plunge.
In fact...since that may be the immediate future of PC cooling, any ideas on that would be welcome. Tahnsk to all who
yeah......every Asus board i've ever used though has left me uninspired. I guess if a person had never seen and used an Abit board, Asus would seem a lot better.
You must let me know how I can join the Church...I HAVE Intel machines...
Hell I have an Amiga, an Apple IIc, and various other exotic and useless devices...but I have NEVER BOUGHT an Intel.
Seriously, all my Intel PC's I have inherited.
I have been an AMD man since 1996, and those first two years were highly suspect!
Anyone who thinks Micro$oft did not try to help Intel destroy AMD in the old days is on crack.
But now...to strike a blow against Sauron and his Intel processors...would be awesome...
hey...every little bit helps.
Of course I could be talking total garbage but it still sounds like the cooler is not doing its job properly and I dont know of anything else that could cause such high temps assuming that the cpu volts are correct and the 4 protective pads are all in the right place and not fouling on anything.
Regards
Jim
I'm with those who are saying re-clean the processor top and reinstall the heatsink and processor. If it's damaged, it may or may not post. But to see 98C at start up points glaringly to a bad heatsink installation, the only other possibility would be a short (which could be caused by residual AS5) or bad circuit somewhere causing an incorrect reading.
Yes, I have a water cooled system. Relative to air cooling, it's a lot of work and somewhat high maintenance. But if you enjoy tinkering and trying new things it can be rewarding, it's cooler and quieter than air cooling.
Why yes, I happen to be something of a Blade Runner fanatic....heh heh...
However, it is still operating, hovering around 60C. I'll probably go out today and try to pick up a better HSF combo than the OEM.