I think this is a power supply problem?

CammanCamman NEW! England Icrontian
edited November 2005 in Hardware
This just started happening recently. Whenever I go into games that are 3D intensive, the computer will just shut off after playing for a couple minutes. My thought is that this less than great brand power supply I have, I think its like 400W "COOLMAX". It's not that old honestly but I can't figure out what else it could be. I took the whole computer apart, cleaned the heatsink and processor core and reapplied thermal paste, cleaned out all the dust and stuff.

I can't figure out what this is and I'm assuming its the power supply. I know it's not the video card because its a new Radeon 9800 Pro and it worked fine for a while and its the whole computer shutting down just immediately. So I'm guessing that the system is just crapping out when the video card tries drawing more power for 3D intensive operations.

My power in my house might be killing my power supplies cause like the washing machine is on the same circuit and I can hear the motor click on through the speakers and stuff, so the wiring is bad. I bought a UPS because I thought this would fix the problem, but apparently I need something with voltage regulation but all the AVR's Ive seen were like $120+

So, I am looking for like a good quality 350-400w power supply (AthlonXP 2800+, Radeon 9800 Pro, two HDD, two optical drives, soundblaster live) and possible a reasonably priced AVR, any suggestions? I'm horrible with the power stuff obviously and I figured that this power supply wouldve been fine for my system but it looks like I once again just screwed myself over buying a cheaper P/S.

Comments

  • LeonardoLeonardo Wake up and smell the glaciers Eagle River, Alaska Icrontian
    edited July 2005
    Yeah, line conditioning is very important. What are the voltage rails on your PSU when it's under load? A little undervoltage isn't necessarily bad; but large fluctuations can be damaging to hardware. (but you probably already know that)

    I have a spare PSU, decent quality, that you might want. Check your private messages.
  • CammanCamman NEW! England Icrontian
    edited July 2005
    This is what this SOYO Hardware monitor thing reports under full load (F@H)

    vcore: 1.64v - pretty constant
    +3.3v- 3.29-3.31v (continues to switch between these two ranges)
    +5v - 4.94-4.97
    +12v - 12.41-12.48
    DDR - 2.47v constant
    AGP - 1.47v constant

    I'm really not sure what any of this means, as I am a PSU newb.

    I am interested in that PSU you PMed me about, do you think that sounds like the problem? I really just want something that is going to work, as I believe most of the problems I've been having recently have stemmed from this cheapo power supply I bought last year, but sometimes its hard to diagnose stuff like this.
  • LeonardoLeonardo Wake up and smell the glaciers Eagle River, Alaska Icrontian
    edited July 2005
    Camman, your voltage rails (individual lines) look very healthy. I'm a bit concerned about the +12v line. It's pretty high, but I don't know if it's above a safe tolerance. At this time, I wouldn't look for another PSU. I'll see if I can find a tolerance guide.

    In the meantime, go to the S-M downloads page, grab Motherboard Monitor, install and run it for a second opinion, please.
  • LeonardoLeonardo Wake up and smell the glaciers Eagle River, Alaska Icrontian
    edited July 2005
    Your voltage rails appear to be very good, well within tolerances. I had thought a general rule of thumb for PC-PSU voltage tolerances was +\- 5%.

    You wrote that you cleaned the heatsink. CPU heatsink, right? What about the video card's cooler? An overheating GPU can seize a computer also. Have you recently made any BIOS changes? The problem is either heat or voltage. It's just a matter now of weeding out the variables.
  • rykoryko new york
    edited July 2005
    try bumping up your agp voltage to like 1.55. i had to do that with my 9800pro when i had similar problems with games crashing to the desktop. worked like a charm for me. :thumbsup:
  • LeonardoLeonardo Wake up and smell the glaciers Eagle River, Alaska Icrontian
    edited July 2005
    Yes, good suggestion. It would seem that a video card that has been performing well at x voltage wouldn't need a boost. Come to think of it though, I've had to do that with video cards before -- out of the blue, after they get a little old, sometimes a voltage boost will stabilize them. Guess it has to do with electron migration that occurs in all semiconductor chips after a while.

    Check the GPU cooler first. Maybe it's loose, dirty, or the fan is malfunctioning.
  • muddocktormuddocktor
    edited July 2005
    Another thing to check on the vid card is to make sure the fan is actually working on the gpu heatsink, Camman. I had a gigabyte 9700 pro die because the fan quit and the gpu got so hot it damaged the core. I was having random freezes while playing games for a couple of weeks before it died. The crap fan that gigabyte had put on the vid card had froze up. :rolleyes:
  • LeonardoLeonardo Wake up and smell the glaciers Eagle River, Alaska Icrontian
    edited July 2005
    The crap fan that gigabyte had put on the vid card had froze up.
    I wouildn't ding Gigabyte too hard on that. North bridge and video card fans across the industry are known for being less than long-lived, brand irrespective.

    On both my home computers, I removed north bridge and video fans and replaced them with large, passive heatsinks. Eliminated risk of fan failure and cut down annoying whining noise.
  • CammanCamman NEW! England Icrontian
    edited July 2005
    I'll check out the fan and stuff but the thing that leads me to believe its a power issue is because it doesnt just lock or crash, the whole computer shuts off like the power is lost to the whole system and i have to start it again, so its not like just a hard lock or a crash, the whole thing just clicks off
  • CammanCamman NEW! England Icrontian
    edited July 2005
    just install MBM5, this is what it looks like through that:

    CPU Diode - 114F, Socket 100F

    "Core 0" - 1.66v
    "Core 1" - 3.31v

    +3.3v - 2.96v
    +5.00v - 5.23v
    +12.00v- - 9.94v
    -12.00v - -5.27v

    These last two numbers lead me to believe that MBM5 is reading my sensors wrong...or my P/S is totally fubared I guess. But it seems the CPU speed and temp sensors and some of the voltage sensors are working correctly. A second opinion?

    Is there any possible damage to my card by upping the AGP voltage?
  • rykoryko new york
    edited July 2005
    those numbers look out of whack, but you can't really be sure what your psu is outputing untill you grab a mulitimeter and read the voltages direct from a spare molex.

    yes, you can damage your video card by uping the voltage too much. it's like a cpu or or anything else. more voltage = more heat = more risk of burning something up. tht being said 1.55v won't hurt you at all. i have had my 9800pro at 1.55v for a long time. 9-10 months at least. as long as your gpu fan is spinning and you have adequate case cooling you should be fine. i would say 1.6v or higher is for water cooling, so keep it below that and you should be ok.
  • ThraxThrax 🐌 Austin, TX Icrontian
    edited August 2005
    That 3.3v is a little piss-poor.
  • LeonardoLeonardo Wake up and smell the glaciers Eagle River, Alaska Icrontian
    edited August 2005
    +12.00v- - 9.94v
    If that's a correct reading, it's definitely time to ditch the PSU. For a mildly tuned (off the shelf) computer, the 3.3, +12, and +5 voltages should not deviate more than 10%. 5% or less deviation for an overclocked system.
  • TheLostSwedeTheLostSwede Trondheim, Norway Icrontian
    edited August 2005
    Camman,

    If the 3.3 rail really is 2.9 volts and the memory draws power from the vdimm, your memory have a maximum of 2.55 volts (no matter what vdimm you set in bios, it's impossible to have more since the 3.3 rail is so low) which is to low. What is your current fsb and cpu speed and at what timings? What memory do you have?
  • CammanCamman NEW! England Icrontian
    edited November 2005
    hey there, sorry I never responded about this, but I just searched for this thread because the problem went away for a while, but oh boy it's back again.

    My system info is as follows

    Athlon XP 2800+ at 2.08ghz
    System bus at 333mhz

    I have a 1gig of RAM split up across a PC2700 stick of 512 and two 256mb stick of PC2100

    I just ordered a 430w power supply ,this one, so hopefully that will determine whether its the PSU or not, I've been told that ThermalTake PSUs are decent quality, I certainly hope so, I got a pretty good deal on this one through through newegg, it was on sale and has a mail in rebate. I'm also thinking about buying a CyberPower 1200VA UPS w/AVR from BestBuy, it's like $90 on sale but I figure I might as well pay it if my power is going to keep killing PSUs.
  • TheLostSwedeTheLostSwede Trondheim, Norway Icrontian
    edited November 2005
    Sounds good to me. Get a cheap digital multimeter as well. It's great to measure stuff with. $15 or so at the closest Shack.
  • QeldromaQeldroma Arid ZoneAh Member
    edited November 2005
    PSU and maybe something more. I know you said it's a new card, but it may be in case circulation. I'd do some additional checking anyhow.

    Generally when you fire up the 3d Graphics accelerator by playing a game, more power is drawn then and the video card gets comensurately hotter. And the 98 is a pretty warm board- especially the AIWs. I'd be suspicious if there was any graphics artifacts in the display just before it dies. You may want to check adapter temperatures- I'm not sure if your card supports the ATI Overdrive feature, but you can check the temps with the Catalyst Control Center if it does. If you're cruising 70C, open the case and try using a house fan to see if it goes longer when you play.

    Regardless, on a PSU replacement, I'd recommend running Memtest afterward as well to see if there might have been some memory damage. For some reason I've had to replace a memory card twice after replacing a PSU that was clipping out.

    And for the Athalon XPs, 1.65Vcore is typical.
Sign In or Register to comment.