PSU temperature and voltage readings

PSUs: The source of DC power for all components in the PC & often a big noise source.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee, Devonavar

Post Reply
shunx
Posts: 341
Joined: Sun Oct 06, 2002 1:20 pm
Location: Vancouver

PSU temperature and voltage readings

Post by shunx » Sun Dec 08, 2002 4:18 am

Can one tell when the PSU gets too hot by only looking at its voltage readings (eg in MBM)? I just got a random reboot so I'm concerned, and I haven't gotten around to getting a thermometer yet. I have one 80mm NBM at 7V in a generic 300w PSU powering. The exhaust feels warm but not too terribly hot at the moment.

quokked
Posts: 333
Joined: Sun Aug 11, 2002 3:26 pm
Location: Melbourne, Australia
Contact:

Post by quokked » Sun Dec 08, 2002 6:34 am

I had random reboots as well from my PSU, I put in a panaflow L1A into it but it didn't seem to pump enough air through the PSU... I replaced it with a M1A and now no more random reboots, what'd I suggest doing is to put another higher airflow fan in or just make that fan 12V and see if the same problems keep on occuring, don't change any other variables of course ;)

shunx
Posts: 341
Joined: Sun Oct 06, 2002 1:20 pm
Location: Vancouver

Post by shunx » Sun Dec 08, 2002 8:41 pm

Yeah I'll get more airflow if it reboots again.

Does anyone know the answer to my question though?

larrymoencurly
Posts: 170
Joined: Tue Dec 03, 2002 11:29 pm

Post by larrymoencurly » Wed Dec 11, 2002 2:03 am

I thought that power supplies now used voltage references that didn't change much with temperature, so almost all variations came from changes in load.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Wed Dec 11, 2002 8:10 am

Temperature affects PSU efficiency. Too high and it drops, which may lead to lower power & voltages, and mybe random crashes. Apparently there is an ideal temp range for efficiency -- too low is not good either.

powergyoza
Patron of SPCR
Posts: 543
Joined: Mon Oct 21, 2002 1:01 am
Location: Vancouver, BC, Canada
Contact:

Post by powergyoza » Wed Dec 11, 2002 8:54 am

Apparently there is an ideal temp range for efficiency -- too low is not good either.
Hey that sounds like an lab test idea to me! Can we first safely assume that a PSU running at high efficiency is a happy PSU? If so can we test a PSU at different internal ambient temps to see how efficiency changes. The ideal internal temp would then be the point at which its efficiency is highest? Then we could start ranking PSUs by how hot they could run while staying efficient.

Could this work?

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Wed Dec 11, 2002 9:16 am

Maybe, but it is a lot of work. Not sure how much value this info would have, the PSU reviews are already chock full of test data :roll: And it doesn't correlate temp with reliability -- there is usually a relationship. So this unit stays running at 70% to 85C HS temp while that one drops to under 60% at 65C. How long will the former last running at that temp? Would it be preferable for it NOT to run at all at such a high temp? It's an exaggerated example...

Efficiency also varies with power level, which is another consideration. You have to consider that different PSUs may have been with different goals in mind.

For example, the Antec TC550 recently reviewed reached 75% efficiency at 300W and dropped to under 72% at 550W. Temperature in the case did not vary much because of the fans speeding up. So why the lower efficiency at full power? My guess is that peak efficiency can only be obtained through a certain range of temps and power levels -- a basic bell curve, with the efficiency dropping on either side. Nothing so exaggerated as a bell, but you get the picture. They chose the mid-range for max efficiency with the Antec550, which seems a reasonable decision to me, all things considered.

shunx
Posts: 341
Joined: Sun Oct 06, 2002 1:20 pm
Location: Vancouver

Post by shunx » Thu Dec 12, 2002 5:48 am

Ack, I'm not sure what just happened to my computer.

I had left my modded generic PSU running for a few days and all was quiet and well, until Windows started being unstable yesterday. I turned the power off and the PSU smelled awful and felt very hot, but there were no firework or anything. Obviously the undervolted fan was not good enough for it. I switched back to the Enermax PSU as a precaution, and tried to reinstall the OS since it's developed errors. Next thing I know, XP wouldn't install anymore. Fatal errors, blue screens, corrupt source files. I've tried several different setup discs as well as dumping images, all of which have seemed to worked before, but now they all fail during installation. I've tried at least a dozen times. Win98 installs ok though. Scandisc doesn't seem to reveal anything. I tried hooking up a different harddrive that has XP installed, but then it just gives me a blue screen. Mm... any ideas?

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Thu Dec 12, 2002 9:17 am

Ooo, sounds bad. So the new troubles are with an unmodded Enermax? Have to tried swapping the HD? If the data is no valuable, try reformat before installing the OS. I think you need to find out whether the PSU cooled the mobo or HDD or both. Hopefully nothing -- except maybe itself. What kind of ambient room temps do you have? Which fan in the old PSU and what speed?

shunx
Posts: 341
Joined: Sun Oct 06, 2002 1:20 pm
Location: Vancouver

Post by shunx » Thu Dec 12, 2002 5:58 pm

MikeC wrote:Ooo, sounds bad. So the new troubles are with an unmodded Enermax? Have to tried swapping the HD? If the data is no valuable, try reformat before installing the OS. I think you need to find out whether the PSU cooled the mobo or HDD or both. Hopefully nothing -- except maybe itself. What kind of ambient room temps do you have? Which fan in the old PSU and what speed?
Well, the new troubles are with a modded Enermax, but I did have two 80mm fans at 12V, which I thought should be ok. I think swapping the HD didn't help, but I'll try it again. Not sure about the room temp, but the case temp was always, as I recall, no more than 35C. The old PSU was a one-fan model, I used one NMB 80mm at 7V for exhaust.

What kind of collateral damages are PSUs known to cause? MB/HDD/CD-ROMs??? I don't know why it is only XP that doesn't install or run though.

ez2remember
Friend of SPCR
Posts: 809
Joined: Thu Nov 28, 2002 5:07 pm
Location: London, UK

Post by ez2remember » Thu Dec 12, 2002 7:15 pm

I am not sure this will help, but in the past I use to use quite a few laptops, I mean about 30 in total in the period of 4 years. I came accross a few that would not install windows 2000, but all other OS installed and worked fine incl. win98, winME etc. On closer inspection I took out the hard disk and two of them had a abnormal rattling sound when I shook it very slightly (something was obviously loose inside, and no not the ibm clicking sounds when shaking it, this was definitely not normal) and these were the two drives that could not install win2k, also a thrd did not have this rattle sound also did not work with win2k.

I have a feeling the underlying cause is your hard disk, the PSU that failed may have caused damage to the IDE cable or the hard disk itself, possibly from a surge when your PSU failed. I would try replacing both with a freinds hard disk first before forking out for a new one. Even better if you have a very old one try putting it in and see if it works.

Good luck.

ez2remember
Friend of SPCR
Posts: 809
Joined: Thu Nov 28, 2002 5:07 pm
Location: London, UK

Post by ez2remember » Thu Dec 12, 2002 7:19 pm

Oh, by the way when I said all other OS installed fine, I mean excluding windows XP. This was in the past before XP was launched.

shunx
Posts: 341
Joined: Sun Oct 06, 2002 1:20 pm
Location: Vancouver

Post by shunx » Fri Dec 13, 2002 5:25 am

thanks for the suggestion ez2remember. However, I have determined via harddrive swapping that it doesn't seem to be a harddrive-specific problem: I took a harddrive from another computer that has the OS in working condition, plugged it into my computer, boot up and, surprise, blue screen occurs. Put the harddrive back to the other computer, no problem again, Windows loads fine. So, some part of my computer other than the harddrive seems faulty.

If this means anything to anyone, when I tried to boot up the working harddrive on my computer, it crashes when loading the file agp440.sys in safe-mode. For another harddrive though it is mup.sys that crashses. I'm beginning to speculate that, *gulp*, my motherboard or CPU may be damaged. Or could it be the costly videocard!? *headaches* Now I'm depressed, lol. What do you all think?

From now on I'm going to treat the PSU with a lot more respect.

ez2remember
Friend of SPCR
Posts: 809
Joined: Thu Nov 28, 2002 5:07 pm
Location: London, UK

Post by ez2remember » Fri Dec 13, 2002 6:06 am

On your misery I am not going to mod my gerneric PSU that came with my computer. I was going to mod it even though I am getting a Nexus PSU soon. Just wanted to see how quiet I could make it, but I take it the safe way and just wait till I get my new PSU.

If your PSU did not do damage to your hard disk, it could possibly be all other components which is connected to the PSU (that's literally everythig :( ). Ask your friend nicely to borrow his hardware and test one by one to see what the cause is. Do a reformat and fresh install for best results. The reason I say that, is if you use a hard disk with a OS already installed it will cause crahes anyway becuase obviously the hardware on your friends computer is not the same as yours. You can't just plug in a working OS into your computer because windows needs to detect and configure all hardware on it (This is done at installation of any OS). Windows XP maybe more likely to minimise this effect but there is no garantee it will work like that. I done this with my friends hard disk with a working win98 and it simply would not boot up on my computer.

Good luck, and hope you find the cause.

shunx
Posts: 341
Joined: Sun Oct 06, 2002 1:20 pm
Location: Vancouver

Post by shunx » Fri Dec 13, 2002 6:34 am

I think Windows can detect changes in the hardware at run time and accomodate it because my old OS images work on new harddware. For example my Win98 right now is from an image a few years ago when I used a different mobo/CPU, and there was no need to reinstall.

Ralf Hutter
SPCR Reviewer
Posts: 8636
Joined: Sat Nov 23, 2002 6:33 am
Location: Sunny SoCal

Post by Ralf Hutter » Fri Dec 13, 2002 7:24 pm

shunx wrote:However, I have determined via harddrive swapping that it doesn't seem to be a harddrive-specific problem: I took a harddrive from another computer that has the OS in working condition, plugged it into my computer, boot up and, surprise, blue screen occurs. Put the harddrive back to the other computer, no problem again, Windows loads fine. ............................. What do you all think?.............................
You can't just swap out your HDD for someone else's. Windows (especially XP. ESPECIALLY if you're going from one brand of chipset {i.e. VIA, SIS} to a different one {Intel}) doesn't like this type of thing. It will blue screen on you. Don't confuse this BSOD with your own BSODs.

If I was you I'd check out your RAM. Download memtest86 and run it. It installs to a floppy and runs from it without Windows being loaded. This allows it to test all the memory. Run all eleven tests and let it loop (it will by default) for at least 6 hours.

shunx
Posts: 341
Joined: Sun Oct 06, 2002 1:20 pm
Location: Vancouver

Post by shunx » Fri Dec 13, 2002 9:28 pm

ok I've noticed that whereas Win98 does not mind me swapping harddrives with different configurations, XP does.

I will try the memory test thing, thanks.

shunx
Posts: 341
Joined: Sun Oct 06, 2002 1:20 pm
Location: Vancouver

Post by shunx » Fri Dec 13, 2002 11:44 pm

it seems Ralf has hit the nail on the head on suggesting an issue with the memory -- the test indicated numerous errors. After swapping the RAM, my computer has booted into XP for the first time in a couple days. Kudos!

I don't know if there are more damages to be found, but I hope this is all. This has been a considerable loss, but it could've been worse. Let this be a warning to others who, like me, may not have considered the danger in modifying the power supply. Next time I'll keep a closer look at the PSU's temperature.

[update: I sent in the RAM for RMA and luckily got back new ones a month later]

Post Reply