Help! new PC shutting down

abbaj288abbaj288 TX
edited October 2004 in Hardware
Hello,
I just put together my new PC with parts from tigerdirect.com.

the motherboard is an Intel D865GBFL Socket 478 ATX with a Pentium 4 2.8GHz processor. 1G of Mushkin Green RAM 400MHz, GEForce FX 5700 video card. I don't know what other info is needed.

Anyway, the PC is just shutting down randomly as far as I can tell. Everything shuts off just like the switch has been turned off. When it starts back up it makes a weird beeping sound and it says at the bottom of the screen that "the computer was previous shut down because of a CPU temperature fault (overheating)". But I have installed the Intel Motherboard Monitor SW and I have never seen the CPU temp go over 120F or so (40-42C I think). There are 2 other temp monitors on the motherboard and they are both low aroudn 86F.

Any idea what is going on? I have never seen the fans stop or slow. The CPU fan has thermal compound and is rated for 3.0GHz P4. The case fan is very powerfull - almost like a hair dryer. Also, I booted up left it sitting for several hours with no activity and it was fine. As soon as I started openign programs it shut down so maybe it really is related to CPU activity. Is it a fault? Shoudl I return the motherboard and or CPU and get replacements?

Thanks VERY much in advance for ht ehelp. I'm really stumped - Intel tech didn't help eaither so I'm not sure what to do.
Michael

Comments

  • LeonardoLeonardo Wake up and smell the glaciers Eagle River, Alaska Icrontian
    edited October 2004
    What are you reading the temps with - software or through BIOS? Are those idle temperatures, or temperatures under load? Under what type of usage does your computer restart?
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited October 2004
    What power supply are you using?
  • WeedoWeedo New
    edited October 2004
    What kind of heatsink are you using? Are you absolutley positive that it is correctly situated on the processor. It may look like it is, but not really be making good contact.
  • abbaj288abbaj288 TX
    edited October 2004
    I'll try to answer all of the questions above.

    How I'm reading the temps: I have checked the temps both in the BIOS and also using a program provided by Intel called Motherboard Monitor which runs in XP. I have never seen either of them show higher than 115F on the CPU temp - well below the "yellow" range on the Intel program. It doesn't change much between idle and when under load, though I'm not really loading it that much except when playing games when I can't read the temps anyway. But it shuts off even when just idling.

    The fan is made by Speeze - also from Tigerdirect. I have checked the seating and it is perfect. Actually, I probably made the seating worse by pulling it out to check it out because I lost some of the thermal compound. But anyway, yes it was seated correctly.

    The power supply came with the case, which is a cheap chinese but it seems to be constructed OK. The power supply is 500W made by Hercules.

    I'm not sure what is meant by "under what type of usage" does it restart. could you explain Leonardo?

    By the way, it's now worse than before. It's now shutting down before I can log in or immediately after logging in but before the system starts up all the way. Man this is so frustrating!!!

    Thanks again for the help. IF I don't figure it out soon I mayh have to fix it with a sledgehammer...

    Michael
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited October 2004
    ...HOW cheap was the case/psu? All power supplies are not created equal...
  • WeedoWeedo New
    edited October 2004
    Don't get the sledgehammer dude, there IS a solution. I've been in that situation a few times. Most recently it was a Swiftech MCV-462V heatsink that for all the world looked to be seated properly but was not. One of my systems was doing a VERY similar thing to what you describe. This may not be your problem but you should get some ArcticSilver 5 and reinstall the heatsink if you think there may not be enough thermal compound left to do the job.

    I found that MBM didn't give acurate temp readings for my systems. Readings were way out of the ballpark. I use thermal probes which I personally trust more.That's just my preference. Then I can see what my temps are at all times.

    Keep messing with it, you'll figure it out eventually.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited October 2004
    The P4's thermal protection implies to me that it's most likely not a cooling issue. I'm fairly sure it's a PSU issue.
  • WeedoWeedo New
    edited October 2004
    Sounds like a good excuse to buy a new silent power supply.
  • abbaj288abbaj288 TX
    edited October 2004
    Thanks for the help again.

    The case was a $60 special but that included the case, power supply, keyboard, really horrible speakers and a mouse. Here's a link to the case:case on tigerdirect

    So Geeky1, if it is a PSU issue would I still see the message "the cpu was previously shut down because of a thermal error (overheating)"? That makes it sound like a problem with the processor to me, but what do I know. So what should I try first - maybe a better heatsink, or new power supply? I've only got a couple of weeks before my 30 day warranty runs out, and I have to allow time to get the return authorization from Intel if I need to replace the MB or processor.

    Also, when I pulled the heatsink to check it earlier the processor popped right out of the ZIF socket because it was stuck to the heatsink with the thermal compound. Will this hurt either the socket or the processor to pull it out like that? I don't really see any way around it because you can't access the lever to release the chip when the heatsink is in place. Bad motherboard design or are they all like this?

    Michael
  • WeedoWeedo New
    edited October 2004
    I've never had the processor stick to the heatsink like that before. Doesn't sound kosher to me, though if everything thing looks ok maybe it is. Arctic Silver 5.

    Do you have access to another power supply to check it out? I also recently had a stock psu go out on me. It was about 1 year old. My computer just went out one day. I didn't think it was the psu at first because I saw lights on the motherboard. It was supplying some power, but not enough. Lighting a few LED's is one thing, powering a complete system, something else. I pulled one out of one of my other rigs and plugged it in and bingo. If you had another one to plug in you could rule it in or out real quick.

    Mmm... shopping for a new psu and heatsink, I'm starting to salivate already.
  • abbaj288abbaj288 TX
    edited October 2004
    Well I bought a $40 heat sink today - good for up to 3.4GHz P4, copper water filled core, adjustable fan speed. Basically major overkill for my PC (maybe if I get it to work someday I can overclock???) Anyway, also bought a tube of Arctic Silver 5 as recommended. Mounted it all up and it actually dropped the CPU temp by 10 degrees F down to about 90-95F at Idle as measured by the MBM. Maybe it's not accurate but still the temp did get lower than it was before.

    So anyway, after all of that it still shut down as before, even without putting the processor under load at all. I'm going to call Intel tomorrow to see how they interpret the error message. Besides that all I can think of is to try a new PS but I don't want to keep throwing money on it if it's jsut the MB or CPU.

    Anyone have any other ideas? Thanks again for the help guys.
    Michael
  • WeedoWeedo New
    edited October 2004
    Borrow a psu if you can. If not, save the receipt and return the psu if that's not the problem. I've done that before. Hopefully that's it. It would be the easy fix. Good luck.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited October 2004
    abbaj288 wrote:
    Anyone have any other ideas? Thanks again for the help guys.
    Michael

    I'm fairly certain (75%) that it's the PSU. Besides, you would be well advised to replace that PSU anyhow.

    All power supplies are not created equal; I have seen "550w" PSUs die at 300w before. And when a PSU dies, it can take everything in your system with it- the motherboard, the ram, the cpu, all of your drives (and the data on them), your video card, all the PCI cards in your system... It can take EVERYTHING to the grave with it.

    And when you think about it, is it really worth rising several hundred $$ worth of computer equipment to save $50 on a PSU? It just doesn't make good economic sense to risk using a cheap generic power supply.
  • abbaj288abbaj288 TX
    edited October 2004
    Hey Geeky,
    I've never bought a PC power supply before. Do I need to pull mine out and take it with me to figure out which one will work or are they all the same as far as mounting and hook-up goes? Any particular brands I should go with? I'll most likely be shopping at Fry's if that means anything.

    Oops - nevermind on the PS suggestions. I took a look at the hardware forums for PSUs and found the sticky thread for PSUs. Thanks,
    Michael
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited October 2004
    In that case (no pun intended), you'll need an ATX power supply. Just about any PSU you find will be ATX (but not all, so check the box to be sure). What you need is an ATX power supply that's P4 compatible (and any psu on the market today is p4 compatible, you'll only need to worry about it if the Fry's around you has a habit of selling 4-year-old equipment, and while the Fry's where I am are bad, they're not THAT bad).

    As for brands, knowing what Fry's carries, I'd go with an Antec TruePower 430w, 480w, or 550w, or an Antec Solution Series 450w (SL450). :)
  • abbaj288abbaj288 TX
    edited October 2004
    I'll pick one up tonight and let you know what happens. By the way, this is the heatsink I bought at Fry's: http://www.newegg.com/app/ViewProductDesc.asp?description=35-128-002&depa=0

    Fortunately for me I only paid $40 instead of the $55 listed here. A similar heatsink got a good review on the main page of this site and it should be a pretty good product according to the review.
    Michael
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited October 2004
    Yeah. The gigabyte heatsink isn't bad at all.
  • muddocktormuddocktor
    edited October 2004
    abbaj288, the only other thing that might be the problem is possibly the motherobard itself. I'm like everyone else here in thinking that your problem is most likely to be the cheap psu but the mobo is another possibility. I've seen quite a few Intel mobo's for sale in Newgg's refurb department lately too. Kind of makes me wonder if Intel is having a quality control problem from their oem for their mobo's.
  • abbaj288abbaj288 TX
    edited October 2004
    Hi muddocktor,
    Yeah I suspected either the MB or CPU from the very start. I'm trying these other options as they are cheaper (?) and will end up giving me a better system anyway. Hopefully the PS will fix it, but if not I'll be going to Intel to help me figure out if it's the MB or CPU. I've still got some time before my 30 days from Tiger Direct is up to do an exchange.

    Is there any chance it's the CPU or is the MB the only real suspect (once I've eliminated cooling and the PSU)?

    Oh by the way, the only bad thing I noticed about the heatsink I bought is that the sides of the little 3.5" control panel don't extend far enough into the case for me to access the screw holes. So right now it's just sitting in the slot. That's really a bad case design more than anything else but it would have been nice if the panel had longer sides like a floppy drive. Once I've got this all figured out I'll probably add that potentiometer to the one I added to control the case fan speed so they'll both be on the same panel. That way I can have matching knobs and an all black case again - if the knobs I have fit that pot.

    Michael
  • abbaj288abbaj288 TX
    edited October 2004
    Well I bought a 420W Enermax PSU last night and installed it. Wow this PSU is heavy duty - probably twice as heavy as the one that came with the case. It was late so I only got to run it for an hour or so but so far it hasn't shut down. I'll give it more run time tonight.

    One question - the PSU came with a motherboard connector to control or maybe monitor one of the fans. They both run with it disconnected so maybe it does something else I don't know. But right now the only place I could connect it is to a connector marked "front case fan" on the MB. Is it OK to connect it there or should I leave it unconnected - or is it supposed to go somewhere else? As I understand it the PSU is supposed to run the fans for a couple of minutes after shutdown I guess to reduce heat-dumping.

    I was thinking about it and at one point I took the side cover off to look at the fans when it was having the shutting down problem. With the cover open I touched one of the PSU cables to move it and it immediately shut down (I had already touched the case to discharge static). I thought it was just a coincidence but now I'm thinking that could have been a sign that the PSU was screwed up. So now I'd be surprised if this doesn't fix it. I'll post an update in a day or two. Thanks for all your help once again.
    Michael
  • ChrisKChrisK UK
    edited October 2004
    Just an idea, have you updated windows against the worms etc? would be funny if it is that and not your hardware :p
  • abbaj288abbaj288 TX
    edited October 2004
    Yeah I updated Mcaffee immediately after I installed XP and did all of the updates for Windows. Would a worm kill the PC before Windows even loads?
  • primesuspectprimesuspect Beepin n' Boopin Detroit, MI Icrontian
    edited October 2004
    No, it wouldn't. This is most definitely a hardware problem.
  • profdlpprofdlp The Holy City Of Westlake, Ohio
    edited October 2004
    What are the BIOS settings for auto-shutdown temp? I've seen factory defaults set unreasonably low cause this exact problem.

    Also, does your power supply actually have the wire to monitor temps? If the board can't get a reading it might be wigging out over nothing. My board has the capability to monitor PSU temp but the PSU itself doesn't have a temp sensor. I just set it to "Ignore" in the BIOS and it runs fine.

    Hope this helps. :)
  • ChrisKChrisK UK
    edited October 2004
    that is a good point,

    is the wire that can be pluged into the MB fan connectors got 3 wires on it or only the one, usualy a yellow one. If it is only the one wire then it is just to tell the MB what speed that fan is running at.
  • abbaj288abbaj288 TX
    edited October 2004
    It's got 3 wires, which is why I was confused about it. Normally 2 are for power and 1 is for the speed sensor I think - so it's weird that the fan runs without it connected. I guess it could be power for a fan control circuit instead of power for the fan itself.

    profdlp - I'll check the BIOS for those things you suggested - but I'm hoping that I'll find that the problem is already solved with the new PS.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited October 2004
    The only other things I can think of are these:

    Check your BIOS setup for an overheat shutdown option. Check and make sure the temp for that is in the 65-75 C range. Possibly, if you still have issues, you might try turnign it off for 4-5 hours and seeing what happens as far as software shutdowns go.

    Second, do you have case fans, and if so is one hooked to the motherboard??? Some Intel BIOSs can also shut down a board if the case fan rpm detected by the motherboard environmental chip and then passed to the BIOS drops below a certain speed or is not plugged into the motherboard.

    Third, try running one monitor program for temps at a time, if boht poll at same time, then you could have a conflict in reading between multiple monitor programs. The one that cannot read data, if it is designed to do this, might shut down the computer.

    Fourth, virals and malware can cause spontaneous system shutdowns, though I have not seen one that in fact gens the message about shutting down that you are getting. I have had to RMA to Intel motherboards from them that had bad environmental chips, including your series. Typically they said first to do what has been suggested, including installing a different power supply. I had one that beeped on startup after being off for three hours, about 3 months after being purchased-- for first month and a half it ran fine, then started shutting down with thermal protect messages, then started hanging at boot, then started thermalling after being off for a long time and feeding ZERO video. Flashing was not an option, and a CMOS battery changeout and sticking the BIOS in recovery mode (CMOS\BIOS jumper block totally OFF). Intel RMA'd the thing and shipped a new rev of same board that did indeed work right and has worked in a customer's house for quite a while longer than the original board survived in anything approaching usable state.

    Bad Environmental chip circuitry was diagnosis by Intel Support. I set up a cross-ship on that one, and part of my basis for that is I wanted an RMA box from them so they got me the new one and I gave them a credit card number to charge the motheboard cost aginst as "collateral" if the old one was not returned within thirty days from the ship of the new board shipped in exchange-- it was, card was never charged. The new board came second day air. I shipped old one back ground with tracking, told Intel the tracking number.

    With TigerDirect purchases, it is often faster to deal directly with the mfr for RMAs than to ship the thing back to Tiger, wait for them to diag it, then get you a new one that might be from same batch as you board is now. I've had 30-48 day defective "send bad in to get new here" cycles from TigerDirect. Second, Tiger's analysis of things thought defective often happens in thier Miami, FL location. I have not heard if that is functional after all the hurricanes down here in FL this year. So, am recommending going right to Intel on this one and bypassing Tiger's gawdawful RMA-defective process. Probable motherboard if what I've experienced proves to hold true in this case, I've had more of thier boards than CPUs fail.
  • abbaj288abbaj288 TX
    edited October 2004
    Straight_Man wrote:
    The only other things I can think of are these:

    Check your BIOS setup for an overheat shutdown option. Check and make sure the temp for that is in the 65-75 C range. Possibly, if you still have issues, you might try turnign it off for 4-5 hours and seeing what happens as far as software shutdowns go.

    I will look for this tonight. I do know that shutting of for extended periods of time didn't help because I had the PC off for 12hrs and it shutdown immediately after start-up at least once.

    Straight_Man wrote:
    Second, do you have case fans, and if so is one hooked to the motherboard??? Some Intel BIOSs can also shut down a board if the case fan rpm detected by the motherboard environmental chip and then passed to the BIOS drops below a certain speed or is not plugged into the motherboard.

    I have one case fan that is connected to the MB. I added a pot that allows me to adjust the speed because it is a very powerful and loud fan. Last night I noticed that when in Windows, if I turn the fan down so that it stops my Intel MBM program will make a noise and pop up a window telling me that a fan stopped. But obviously that didn't cause the MB to shutdown - just alerted me to the problem.

    Straight_Man wrote:
    Third, try running one monitor program for temps at a time, if boht poll at same time, then you could have a conflict in reading between multiple monitor programs. The one that cannot read data, if it is designed to do this, might shut down the computer.

    I've only been running one program (Intel MBM). I did download another generic one but haven't tried it yet.
    Straight_Man wrote:
    Fourth, virals and malware can cause spontaneous system shutdowns, though I have not seen one that in fact gens the message about shutting down that you are getting. I have had to RMA to Intel motherboards from them that had bad environmental chips, including your series. Typically they said first to do what has been suggested, including installing a different power supply. I had one that beeped on startup after being off for three hours, about 3 months after being purchased-- for first month and a half it ran fine, then started shutting down with thermal protect messages, then started hanging at boot, then started thermalling after being off for a long time and feeding ZERO video. Flashing was not an option, and a CMOS battery changeout and sticking the BIOS in recovery mode (CMOS\BIOS jumper block totally OFF). Intel RMA'd the thing and shipped a new rev of same board that did indeed work right and has worked in a customer's house for quite a while longer than the original board survived in anything approaching usable state.

    The PC has seen very little time online, and none using MS Outlook or other virus friendly programs :) That doesn't mean it's clean but most likely it hasn't been hit yet.
    Straight_Man wrote:
    Bad Environmental chip circuitry was diagnosis by Intel Support. I set up a cross-ship on that one, and part of my basis for that is I wanted an RMA box from them so they got me the new one and I gave them a credit card number to charge the motheboard cost aginst as "collateral" if the old one was not returned within thirty days from the ship of the new board shipped in exchange-- it was, card was never charged. The new board came second day air. I shipped old one back ground with tracking, told Intel the tracking number.

    With TigerDirect purchases, it is often faster to deal directly with the mfr for RMAs than to ship the thing back to Tiger, wait for them to diag it, then get you a new one that might be from same batch as you board is now. I've had 30-48 day defective "send bad in to get new here" cycles from TigerDirect. Second, Tiger's analysis of things thought defective often happens in thier Miami, FL location. I have not heard if that is functional after all the hurricanes down here in FL this year. So, am recommending going right to Intel on this one and bypassing Tiger's gawdawful RMA-defective process. Probable motherboard if what I've experienced proves to hold true in this case, I've had more of thier boards than CPUs fail.

    Like I said, I'm hoping it won't come to this. By tomorrow I should know if the PSU fixed. Thanks for the suggestions.
    Michael
  • abbaj288abbaj288 TX
    edited October 2004
    4hrs run time last night and no problems. I think it's fixed. Thanks everyone for your help. I wouldn't have guessed in a million years that it was the PSU. I figured all PSUs are created equal. Well at least I'm learning something! :D
    Michael
Sign In or Register to comment.