Name: exus Date: April 5, 2008 at 12:50:51 Pacific Subject: random pc shut offs during cs OS: XP Pro CPU/Ram: 4800+/1GB Model/Manufacturer: me
Comment:
Hello. I'm having a strange problem with my computer. I recently retired from WoW and am back to play old CS 1.6. However at random times during play, and thus far this has only occured during CS 1.6, my computer shuts off as if I had pushed the power button or the power had gone out. I am aware there could be any number of things causing this. I have restarted my machine several times, I have unplugged all plugs, on both ends, including my power cable, surge protector, etc. to remove static electricity. No wires on my mobo are touching any screws or any large metal portion of my case, all mobo wires are plugged in tightly, including the system power. I have a very powerful computer and find it strange it only does this with a game that is nearly 10 years old now, and that I have been playing for about 5 years. I have never experienced this problem before. Any ideas or suggestions are greatly appreciated. Thank you for your time.
So the HL GoldSrc engine is causing a load on my machine? I have only experienced this problem during CS and a few times during WoW. I have played MUCH higher end games like HL2 EP2 for longer periods and have never seen this happen before. I will look into the PSU issue, though I did not before, It is almost 2 years old now, perhaps it's time has come. Nonetheless, 500W is more than enough for my machine. Ty for your suggestion and consideration, I greatly appreciate it.
Any PS can malfunction or fail. Sometimes you can't find any obvious symptoms there's something wrong with it, yet trying another PS eliminates the problem.
You could also try this.
Win XP is set by default to automatically reboot when it encounters an unrecoverable error.
(However, it does NOT make the computer shut down.)
To have XP possibly display an error message you can investigate instead of the computer rebooting:
1. Click Start, and then right-click My Computer. 2. Click Properties. 3. Click the Advanced tab, and then click Settings under Startup and Recovery. 4. Under System failure, click on the small box beside Automatically restart to remove the checkmark. 5. Click OK, and then click OK.
If you then get an error message, look at all of it's details.
Ok it no longer only shuts off during cs. Today it happened again, but right after i booted it back up, it shut down again. I still can't really tell if it is overheating or if it is the psu. It is a bit warm in my house atm, after the 2nd shut off my gpu was running at 67*C. Nvidia had a control for overheating, when to shut off when the temp reached a certain point, they have redone their control panel and i can no longer find that option. I can't get my hands on a psu atm to test anything, but i am beginning to think it is heat/temp related. I am unsure of the temp of my cpu.
Open up your case and make sure the fan on the video card is spinning when the computer is running. Take a look at that and the cpu fan and heatsink - if they have a significant amount of mung on them - dust and lint etc., clean them off. It's a good idea to remove the cpu fan from the heatsink so you can properly clean hard to see mung on the top of the heatsink. Remove the AC to the case when you do that (e.g. switch off a power bar it is connected to but leave it and the computer plugged in to keep the case grounded). Don't use a vacuum cleaner unless you can blow with it without touching the mboard rather than use attachments to suck off the dust - vacuums produce a tremendous amount of static electricity. Otherwise small artists paint brushes, preferably with natural rather than plastic bristles, work well.
"I am unsure of the temp of my cpu."
Look in the bios Setup - the current temperatures, voltages, and cpu fan rpms are displayed under the heading Hardware Monitor or similar. OR Most CDs that come with mboards have a software utility you can install in and load in Windows that allows you to see those readings at any time. e.g. a hardware monitoring utility.
"...they have redone their control panel and i can no longer find that option."
Try looking in Control Panel - Display - Setiings - Advanced.
If your computer is sometimes operating in a relatively hot environment, you may need to add a case fan - the best place is higher up at the back of the case - or one that blows from the side of the case directly at the cpu, or get a more effective cpu fan or cpu fan / heatsink combo. However, don't add any more than it takes for the temp to reach a minimum - more fans than that merely accelerates the mung buildup inside your case.
A PS fan that is no longer spinning or that is spinning too slowly, or a PS that has mung buildup inside of it, can certainly cause heat buildup inside your case.
The CPU fan/heatsink is relatively clean, and the fan is running just fine. My video card fan is also running just fine. However, you mentioned a PSU fan issue - and i may have found my problem - i never even thought about the psu fan. It is certainly running WAY to slow, it barely generates any air at all. I have been playing w/ my side panel off for quite some time, because with it on it would cause the shut offs to happen more frequently, which over the past month they have become a daily and prominent problem, i have never experienced them at this rate before. I have a small vent in the front of my case, behind which is a 120mm intake fan, and at the rear i have a 120mm exhaust fan. My side panel has a small vent, positioned so that is pretty close to being perfectly over the top of my cpu. However i think the heat contributes to my psu failure, which is becoming more and more apparent. For nearly a month, i was not able to shut off my computer AT ALL, because my mouse and keyboard drivers would somehow disapear. I have since reformatted and no longer have that problem. But 30 days with few to no proper shut downs... i think the damage has been done there. Not only that, i was a very avid computer user, i spent many hours on it gaming, torrenting, etc. So i am beginning to feel concise that it is a psu problem. I plan on buying a new one anyway, as this one is obviously bunked. I knew when i bought the parts that the cpu and gpu were "hot runners", but never experienced heat problems before. I will certainly consider adding another fan somewhere. So do you feel that the slowed psu fan and the obvious over-usage would lead to my problem? Thank you very much for your time and consideration regarding my matter. I hope to keep in touch through this thread as I continue to update you on my ridiculous problems.
"...i never even thought about the psu fan. It is certainly running WAY to slow, it barely generates any air at all." "So i am beginning to feel concise that it is a psu problem. I plan on buying a new one anyway, as this one is obviously bunked. "
How fast it spins depends on the model. The important thing is whether the PS is getting too hot - if it is it will eventually malfunction and fail. See response 4 in this and check for obvious symptoms your PS is overheating: http://www.computing.net/hardware/w...
Look up the specs for the PS model.
Better PSs have better heatsinks so the PS inherently runs cooler, and an automatic variable speed control on one or more fans - the fan(s) spin(s) faster only when the PS has more load on it and gets hotter - that's normal for that situation.
El-cheapo PSs have less efficient heatsinks and usually a single 80mm fan that is supposed to run full speed all the time. If it has a cheaper fan that does not have at least dual ball bearings, the one or two sleeve bearings eventually get worn to the point the bearing(s) produce too much friction and the fan can no longer spin at it's full speed.
If you have a failing bearing problem, it is common for you to hear rattling or screeching sounds when you first start up the system after it has cooled off to room temperature.
"... the obvious over-usage would lead to my problem? A better PS will handle any load you put on your system, no problem, as long as it has enough capacity. An el-cheapo PS may have false specs and it can't actually handle it's max ratings. Some PCI-E video card models require a lot of power. The minumum recommended PS capacity your system should have is often specified in the specs for the card on the manufacturer's web site.
Hi again. My cousin works in the IT field and is extremely knowledgeable about just about anything computer-related. Once i get an opinion here, I always go to him for a second opinion. He has not ruled out the PSU failure possibility, he also mentioned the temp-dependent fans, so it's possible mine is running slightly slower because the load upon it was not that great. However, he did bring up the idea of CPU/Mobo overheating, and after running EasyTune,an overclocking/fan/temp/volt monitoring utility by gigabyte, i found some startling results. Mind you all i can really do with this utility is monitor numbers, overclock, and change fan speeds, which i did, i changed the CPU fan 100% speed cap from 60C to 40C in hopes of it cooling the CPU earlier.
My CPU runs at about 35C resting, and it's fan at about 2900RPM. After booting up CS1.6, the temperature increased to about 45C at 3100RPM. After a little over 10 minutes of playing, my CPU temperature jumped drastically, seldom decreasing, if at all. The maximum temp I recorded was at about the 10 minute mark, running at 118C at 3150RPM. Call me crazy, but is that not a massive leap for such an old game? In general, I would call that ridiculous for any cpu to jump that high over such a short period of time. I know they run hot, but that's incredibly hot very quickly.
I play with my side panel off, and i have a 120mm intake and a 120mm exhaust. I felt little change in the PSU exhaust fan speed, with a negligibly small increase in speed. Perhaps it is a combination of the two?
I am not that savvy w/ PSU's, as you may have noticed by now, but here are some readings i am not entirely familiar with.
Under the Voltage colum of the Hardware Monitor it reads the following:
ITEM V (Voltage i'm assuming) VCoreA 1.340 +3.3V 3.290 +12V 11.980 (fluctuates up to 12.040)
Could you elaborate on what those mean and if you feel the possibility of CPU overheating is to blame, or a combination of the two? Thanks again for your time and consideration.
"The maximum temp I recorded was at about the 10 minute mark, running at 118C at 3150RPM. "
Whoa! that's way too high! That's high enough to begin damaging the cpu! Something is out of whack! That sounds like - the combo of heat sink and fan is inadequate. Did those come with the cpu in the same box or did you get them separately? - OR - the heatsink is not installed properly, and/or something's wrong with the way the proper undamaged thermal pad, or thermal grease, or thermal paste was installed. - OR - the CPU core voltage is too high.
NOTE that some mboards have a jumper that ups the cpu core voltage for overclocking use. E.g. On a Asus mboard I installed for someone, that jumper was in the position that ups the voltage, despite the fact it was brand new and the box it was in had never been opened since it was manufactured - and that fact WAS NOT shown in the bios Setup current voltage readings. The telltale sign of that was the cpu temp was 10 to 15 degrees C hotter than I excepted it to be, even at idle - yours is much higher than that! Check your mboard manual to see if the mboard has such a jumper, and if it does, make sure it is in the normal position.
To give you a comparison, I installed an AM2 64 X2 6000+ boxed set of cpu, heatsink, and fan on a friends mboard recently. The heatsink has heat pipes and is very efficient. The cpu rarely goes above the low 40's C! It idles in the low 30's.
The +5v, +3.3v, and +12v readings are fine if they're within 10% of their nominal values; the VCore voltage should be within 5% of it's nominal value. They may vary according to the load on the cpu.
AM2 cpus have the "cool and quiet" feature. If support for that is enabled in the bios Setup, and if the software support for that is loaded in Windows, the cpu is throttled back and it's power consumption goes down when there is less load on it, and the cpu fan speed is decreased, if and only if the cpu fan has a 4 wire cpu fan connector connected to a 4 pin cpu fan header on the mboard.
I knew something wasn't right. I ordered a new PSU anyway, thankfully I start my job again in May. I ordered the mobo/cpu in a bundle from Mwave.com They have an option to mount/install the cpu on your mobo before they ship it out. I know that may not be the best idea, but I had no one around for a while to help me install the cpu when I ordered the cpu, so my impatience at the time may be to blame. My cousin also thinks that the thermal paste may be causing an issue, but this is the first time this problem has arisen at this rate. The mobo itself runs a hot "north-bridge", whatever that is. I remember f---ing with the CoolNQuiet feature a while back, maybe that is to blame. I should probably note I reformatted my machine on March 15. But should I re enable CnQ if it is disabled? I heard it bottle-necked my games.
I can't think of any way this could be a Cool and Quiet problem. I haven't enabled it myself on two mboards I installed for others. In theory it should automatically adjust to more load demanded by games fine. The biggest bottleneck on more recent mboards is the hard drive max speed before anything else. In any case the cpu temp should never go above about 85 C absolute max.
Antec makes excellent quality PSs so it's certainly not a problem related to it being an el-cheapo, and it's probably normal for it's fan (it may have two fans - another one on the bottom of the PS box) to run slower when there is load on it since it has more efficient heat sinks and probably (a) temp controlled fan speed control(s).
I see the Antec has been listed all along in your signature, but I usually ignore any specs in signature text because I can't be sure the specs are for the system the thread is about.
It can't be caused by the PS - the core voltage is determined by voltage regulator circuits on the mboard, not by voltages put out by the PS, which in any case are well within specs according to your quoted readings.
Check it's mboard manual for a jumper that ups the core voltage. (Now that I know it's model, I'll take a look at your mboard manual too.) If it doesn't have that, or it does and it's in the normal position, the problem has to be either - you're using an inadequate heatsink and cpu fan combo that didn't come with the cpu - there is something seriously wrong the way the heatsink is mounted or whatever you used between the heatsink and the cpu, or there is nothing between the heatsink and the cpu.
I believe Cool and Quiet only works with 4 wire cpu fans - is yours 4 wire or 3 wire?
It does not have a jumper for upping the cpu core voltage.
Upping the cpu core voltage probably has more of an effect on how hot the cpu gets than anything else.
In your bios Setup:
Are you overclocking the mboard in any way?
MB Intelligent Tweaker
Chipset Voltage Control - Normal, or more Is that normal, or more?
CPU Voltage Control - custom Vcore setting Normal CPU Vcore - default setting Is a custom setting being used?
If you're not sure if the bios settings are being overclocked somewhere, Load Optimized Defaults. According to the manual, the bios will auto detect appropriate settings according to what components you have installed. ....
If you have flashed the bios....
Flashing the bios usually does not load bios defaults after the flash procedure - to be sure they have been loaded, you should ALWAYS load either Fail Safe or Optimized defaults after the flash - otherwise the Cmos contents and what you see in the bios setup screens DOES NOT MATCH the bios version, and that can cause all sorts of problems. The bios chip has both the Cmos data contents, which determines what you see in Setup, and any custom settings you make or the bios makes are retained by means of the mboard battery; and read only data contents not normally visible to the user, which do not need the battery to be maintained.
If you don't see anything wrong in the bios, and if loading Optimized defaults doesn't help, your problem must either - you're using an inadequate heatsink and cpu fan combo that didn't come with the cpu - there is something seriously wrong with the way the heatsink is mounted, or whatever you used between the heatsink and the cpu, or there is nothing between the heatsink and the cpu.
However... if the cpu has gotten as hot as you quote more than once, or for any significant amount of time, it's quite likely the cpu is damaged and will no longer behave normally.
See response 6 in this - you may have a serious problem if you used thermal paste and it has hardened over time: http://www.computing.net/answers/ha... You may NOT be able to release the heatsink from the cpu by simply releasing the heatsink clamp and twisting the heatsink while pressing down. That also may have changed it's thermal characteristics.
The few times I've had to pull the heatsink stuck on the cpu combo straight up it didn't hurt the ZIF socket or the cpu pins, but it's certainly risky.
If you decide you must get another cpu, I recommend the 6000+ or 6400+ boxed set with the heatsink and fan - their heatsink is much more efficient, the fan is 4 wire, and the cpu runs quite cooly, and the prices of them have come down a lot lately. You can also use a Phenom in this mboard, but if you're after max fastest overall performance, according to benckmarks I've seen, the 6400+ is the top choice. The Phenom is a moderately priced (starting at only a bit more than a 6400+) middle market quad core series - it's not meant to compete with the more expensive quad core Intel cpus - if you aren't using programs that benefit from quad core cpus, you're better off with a 6000+ or 6400+.
Note that there is a 6400+ black edition - it has a very tiny performance advantage over the 6400+ regular cpu, and is priced similarly to the regular 6400+ boxed set, but it does NOT come with a heatsink and fan.
You may need to flash the bios first though - if you do, I recommend you use the bootable Dos floppy method, even if you have to borrow a floppy drive, or even buy one, they're cheap. CPU support, according to the bios version: http://www.gigabyte.com.tw/Support/...
My board/cpu combo is still under warranty from Mwave, so i think i can still RMA it to them and get a new one. THEY mounted the cpu/heatsink on the motherboard, i have not dealt with it since i put the mobo into the case. If i did get a new CPU, say the 6400+, do i just pop it in and go? or do i need to do anything to my hdd?
"My board/cpu combo is still under warranty from Mwave,...."
If you used any non-default settings in your bios that overclock anything to do with the mboard or cpu, or higher than the Normal Vcore voltage stated in the bios, the mboard and cpu warranty does not apply.
The cpu is warrantied for three years. Mboards are warrantied for at least a year, and up - many for 3 years these days,
If the heatsink and fan are inadequate, or if the installation of any heatsink was improper, that's their fault, and they should replace whatever has become faulty because of that.
"If i did get a new CPU, say the 6400+, do i just pop it in and go?"
Yes, if it has a new enough bios version or newer; if not the bios version needs to be updated first. See the end of response 15.
FYI I got a 6400+ boxed set last Monday for $185 CDN - they may be able to get that for less - you would have to pay the diff between that and the current 4800+ price.
"...do i need to do anything to my hdd?"
NO. Even if they have to replace the mboard with a different one, in which case XP may not load, it's simple matter to fix that without losing the data on the hard drive and load the drivers for the new mboard. The ram will work fine in a different mboard - in the case of AM2 amd AM2+ mboards, the memory controller is in the cpu and that's what the ram must be compatible with, not the mboard chipset as in the case of most other cpus.
Ok, well I have some news. I went into my bios and set optimization defaults, let it do it's thing, then ran another test. This time, after 10-15 minutes of CS, my CPU temperature did not jump above 49C, with a resting temp of 33-35C. It did increase, bit by bit, so I wonder if after prolonged use, it heats up to it's unbelievable temperature again, or if after being unused for a day or 2, it's able to handle usage again. I do not want to get my hopes up on this bet for fear of them being crushed. This is the first time I've turned it on about a 36 hours, I've been using another machine in my sister's room. Do you feel that a simple change like that did the trick? I don't know why or how anything would have been OCed, I'm not a fan of OCing and have never enabled any type of setting to allow for it. I still plan on having my cousin come over and have a look at my cpu for me, and finally teach me how to open them, etc. I always dreaded dealing with them since i know they can be a pain in the ass. If there is anything I can do to help or entirely fix the situation now I will do it, I don't have the money to replace any more than the processor. If I can RMA anything back to Mwave, or send anything back or have them replace anything I will do that also, if at all possible, I'd rather have an entirely new CPU since this one has gone haywire once already. I'll continue to keep you up to date. Another thousand thanks to you for your help and advice thus far.
"Do you feel that a simple change like that did the trick?"
It could very well be. You said before the temp went up to 118 C ten minutes after starting playing cs. If it no longer does that, you may now be in the clear, IF the cpu isn't already damaged. Once the computer has fully warmed up, whatever temp the cpu gets to when you play cs or anything else that stresses the system more should be consistant and you should see the same range of temps when doing the same things, providing the ambient temp - room temp - in the room the computer is in is consistant. You shouldn't need to keep the system on for long periods of time to determine the max temp - once you get past the point where the computer has fully warmed up, e.g. 15 minutes, the cpu temp should be in the same range when doing the same things regardless of you long the system has been on.
Monitor the temp frequently for a while if you like, but it should be consistant, and there should be no need to monitor it for a long time.
On the other hand...... You said before the temp went up to 118 C ten minutes after starting playing cs. Are you SURE you had the display in degrees C and not degrees F? 118 F is about 47 degrees C.
If you DID have the display in degrees F, the cpu is NOT overheating excessively, though it could be a bit lower. In that case, it's a well known fact games tend to have more bugs in them than most other software and do not work properly on all systems they're supposed to work on. If cs is the only thing you have problems with, you could look on the game maker's web site for t-shooting suggestions or updates, or on the video maker's web site for t-shooting suggestions or special video driver versions that work better with the game (ATI sometimes has the latter - I've never heard of NVidia having the latter), but sometimes there is nothing like that, or there is but nothing you do helps, and your only option is to not play the game at all.
If you don't have such in Windows yet, most CDs that come with mboards have a software utility you can install in and load in Windows that allows you to see those readings at any time. e.g. a hardware monitoring utility. And/or the same is often available in the software downloads for your model on the mboard manufacturer's web site.
Yes I am absolutely positive my readings are in Degrees Celsius. Both are displayed, and 118C was coming out to be about 250F if I remember correctly. I will continue to monitor it's temperature, and plan on replacing/upgrading the cpu when possible.
OK 118 C is about 244 F Up until this example I had thought it was not possible for the cpu to get that hot from overclocking settings alone, unless the Vcore voltage was above the standard range or the voltage stated on the cpu, with the heatsink and fan you get in a boxed set with the cpu.
Yeah, I have no clue how anything got overclocked, never even so much as gone near any of those settings, but thus far, it runs normal again. I just hope that WAS the problem and not the PSU s--- again, though i already ordered a new one =/
And we're back to square one again. So for about 2-3 days I had no problems, CPU went no higher than 46-48C. Today I had another shut off, no idea what the temp was, so I ran another experiment and it reached about 65C within 10 minutes, and probably would have continued climbing. My room is also slightly warmer today than it has been, so any slight change in temperature seems to set it off. I am aware that the problem lies within my processor but I plan on having it looked at within the week. It really pisses me off that something that small changes it.
A small change in room temp should not make the cpu heat up that much more. A friend in Aus I correspond with has to add more than normal cooling in his cases, and/or upgrade his heatsink cpu fan combo from stock, but that's because he has no basement and the outside temps get into the 40s in the summer sometimes, and if that goes on for days on end it may get to the high 30s where he has the computer.
Since the temp was stable and reasonable, within the expected range, for a few days ,I think we can assume there's probably nothing wrong with the mboard, or the hardware monitoring circuits, or the sensor that senses the cpu temp, or the heatsink installation/thermal pad or grease or paste. That leaves - - the cpu fan could be faulty and sometimes doesn't spin or spins too slow - the PS fan(s) could be faulty and sometimes doesn't spin or spins too slow, or the controls for them could be faulty. Problems with fans are often caused by their bearings producing too much friction after they have been used for a long while. In that case, it's common to hear rumbling or ratting or screeching sounds when you first start up the computer after it has cooled down to room temp, or after it has cooled down to room temp and has been sitting for quite a while, such as overnight. Problems of variable speed fan controls not spinning fans fast enough when they should be are harder to diagnose, but usually if that is faulty, it is always faulty, so doing the same things should produce the same results if there is nothing wrong with the fans themselves.
Not likely ... - the cpu heatsink / cpu fan combo may be inadquate, if it didn't come with the cpu in a boxed set. Ask the guys who assembled the system if the combo came with the cpu or not, or look on the original bill - if the combo isn't listed separately, it probably came with the cpu. In that case the combo may not be able to cope properly with small changes in room temp or more load on the cpu, but, again, if that were the case it would be consistant, especially according to the load on the system. I have a combo that works fine with a 1ghz soket A cpu, but a 1.33ghz cpu overheats unless a fan that pushes a lot more air is installed on the heatsink.
Most CDs that come with mboards have a software utility you can install in and load in Windows that allows you to see those readings at any time. e.g. a hardware monitoring utility. Install that utility if you haven't already done so. Tweak the settings so the fan rpms trigger a warning when they drop a few hundred rpm below the rpm you see when you first turn on the computer from cold, and tweak the warning temp for the cpu temp, say, to 55C. Set the utility to load on Startup so that it is always running. (Note that you may have to turn if off when you run Defrag) That way you are warned when a problem starts happening. Some better PSs have one fan that has two wires on a three position female connector you can connect to a mboard PS or case 3 pin fan header so that it's rpm can be read.
EasyTune warns me when the RPM drops. When it's heating into the 40's-50C, it's already spinning at nearly 3100 RPM. And regardless, I would feel better with a new CPU, knowing that this one has obviously caused problems in the past.
No, you were agreeing with me earlier about how it is running too hot and has to be the cpu. But it doesn't matter anymore, I now know exactly what is causing my computer to shut the f--- off. I took my heatsink out at last, having been fed up with the bulls--- of overheating. And I have a few pictures to show you what has happened to my cpu, since it IS after all the cause of my problems, not mung buildup, not a defective Antec PSU. Please enjoy.
My cpu is G L U E D to the f---ing heatsink via cheap ass thermal grease. Mwave shall be feeling the full extent of my wrath. Thanks for your advice through this matter, it has come to an end at last. It was almost as long and as epic as a lotr movie.
I can't tell from those pictures whether the cpu is defective.
I have encountered this problem of not being able to remove the heatsink before too - it has only happened to me if thermal paste was used by someone - true thermal grease (silicon grease or silicon thermal compound with nothing else added to it) has never hardened in my experience, and the heatsink can easily be removed by merely giving the heatsink a sideways twist after releasing the heatsink clamp while pressing down. A thermal pad may stick strongly to the cpu too, but I have always been able to release the cpu the same way, with a twist of the heatsink. You can't go wrong with true thermal grease - howver most AMD boxed sets have a thermal pad on the heatsink.
If thermal paste WAS used, it's quite possible it's hardening destoyed it's thermal characteristics. But that doesn't jive logically with the fact the cpu did not overheat for several days after you loaded Optimal bios defaults.
If the cpu has been damaged by overheating, it's core is discolored, but you can't see that until you remove the heatsink. Pictures of blown capacitors, other components, power supplies, Athlon cpu's, etc.: http://www.halfdone.com/Personal/Jo...
If you are fortunate the ZIF cpu socket is still okay - you won't know for sure until you install another cpu - it has been in the few cases when I've had to do what you did, pull the whole combo straght up out of the socket because you can't release the ZIF socket lever with the heatsink in the way - but make sure there are no broken off pins in it.
I haven't yet gotten a stuck on cpu off off a heatsink when thermal paste was used and it's strongly adhered. I tried a single edge razor blade between them and that was useless, even after warming the cpu with a heat gun at a medium setting (a glorified hair dryer with higher heat capability). I gave up in all cases, and just installed a replacement fan that moves more air in some cases. I don't advise using anything narrow to pry with - the wider the better - if you want to be able to try using something else between the heatsink and cpu . A putty knife or other wider stronger blade may work.
I just looked at a AM2 mboard I have. It would be difficult to twist the heatsink because of the way the heatsink is keyed into frame around the cpu socket! That makes me realize it might be a very good idea to use thermal grease instead of a thermal pad or thermal paste - if the heatsink were stuck to the cpu you would have to unscrew the cpu socket after taking the mboard out of the case otherwise, and even then you might damage circuits on the mboard while twisting the heatsink!
Yeah, no idea yet if I f---ed up the ZIF socket - I had no choice but to pull the whole thing out. However I was very careful to pull straight up as best I could to try to avoid any pin bending. Today I called AMD and they agreed they'd send me another CPU, a 5000+ this time since the 4800+ is not in stock. They did have one condition - I had to pry the CPU off the heatsink and send it back. I thought with a little twist and pull it would come off. I could NOT remove the damn thing with my hands! I had to use a knife and pry it off very carefully, and after a few good knocks with a crab knocker (I'm from Maryland) it popped off. But anyway, problem halfway solved - not sure about the ZIF socket on the mobo. But we shall soon find out.
"I could NOT remove the damn thing with my hands!"
I'm not surprised. See response 29.
Is the cpu core discolored? See the pics in response 28. If it is not, you should have no problem with AMD accepting your RMA claim - if it is, they may not accept your RMA claim and you will have to pay for the replacement cpu.
Did the heatsink have a thermal pad, or did they use thermal paste? A thermal pad is what is on the heatsink in the AMD boxed set. If it had a thermal pad you will see the remains of a definite square or rectangle of pad material on the bottom of the heatsink.
If it has a thermal pad, there are two mboards I have recently installed I may need to get rid of that on!
Because of it being difficult to twist the heatsink because of the frame around the bottom of it on the mboard, I recommend you get some true thermal grease (silicon grease, silicon thermal compound) rather than using a thermal pad. It is usually "milky" translucent, nearly clear when in a thin layer. Some places that have lots of computer parts or repair computers have capsules of it.
If the ZIF socket was ruined, assuming it's a AM2 cpu (940 pin) a new AM2 or AM2+ mboard can be relatively cheap, and will have no problem with your cpu or ram - the memory controller is in the cpu, that's what the ram must be compatible with rather than the mboard chipset as in on most other mboards, so the ram is compatible with any AM2 or AM2+ mboard.
If the main chipset is different enough from your old mboard, Windows will not load - typically you see the first bit of Windows graphics, then a black screen with a blinking cursor top left - that is normal. That can be easily fixed without you losing the data already on your Windows (usually C:) partition - you run a Repair Setup procedure (often called a Repair Install). I can point you to directions if you need them.
Whether Windows loads okay or not, you need to load the drivers for the new mboard after that.
The information on Computing.Net is the opinions of its users. Such
opinions may not be accurate and they are to be used at your own risk.
Computing.Net cannot verify the validity of the statements made on this site. Computing.Net and Computing.Net, LLC hereby disclaim all responsibility and liability for the content of Computing.Net and its accuracy.
PLEASE READ THE FULL DISCLAIMER AND LEGAL TERMS BY CLICKING HERE
