acceptable temperatures for motherboard and psu

hi everyone,

my machine keeps hanging, requiring a powerdown for it to start again (ctrl, alt & del doesn't work) - i'm worried it's overheating so i've checked the BIOS Health status a few times and come up with the following averages -

cpu temp         56/132
syt temp                      37/98
cpu fan speed      3260 rpm
sys fan speed       0 rpm
vcore            1.488v
      3.3v      2.976v
      +5.0v      5.030v
      +12.0v      11.680c
      -12.0v      -11.989v
      -5.0v      -5.177v
      battery      3.168v
      +5v sb      4.993v

1)  shouldn't my system fan be running at more than 0 rpm - and if so, how do i get it going?
2)  also, does the cpu fan speed seem a little low at 3260 rpm?  and again, how can i boost this?
3)  if i'm going to have to replace my RAM, should i be considering upgrading to PC2700 - is there a site that'll tell me if it's compatible with my motherboard?

I metion this last point as i first suspected it was RAM issues and have run all the tests i can find (DocMem, GoldMemory and Memtest86 amongst MANY others), repeatedly and for huge lengths of time and found nothing - i have no "known good RAM" to hand so can't test it that way, so am looking for other options before i spend money on new RAM only to find out that the problem reoccurs

and by the by, the machine is:

Windows XP Pro SP1
Intel Pentium 4A, 2400 MHz (4.5 x 533)
MSI 845G Max (MS-6580) (6 PCI, 1 AGP, 1 CNR, 3 DIMM, Audio, Video)  
Intel Brookdale-G i845G  
1 GB PC2100 DDR SDRAM
GeForce4 Ti 4400 (128mb)

Any help given would be GREATLY appreciated - this problem has been reoccuring for a while now and every time i think it's resolved it comes back again.  All drivers are up to date, and i've reinstalled windows.

HELP!!
LVL 1
tonybushellAsked:
Who is Participating?
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

SkackCommented:
1)  shouldn't my system fan be running at more than 0 rpm - and if so, how do i get it going?
A:  Do you have any additional fans?  If they are installed, check to see if they are spinning because that is the ultimate deciding factor, not what the BIOS reports.  If they are attached directly to drops of the power supply, then they will not show up in the BIOS.
2)  also, does the cpu fan speed seem a little low at 3260 rpm?  and again, how can i boost this?
A:  There is no way to boost the fan speed, but you can install a better fan.  I have a ThermalTake Volcano 7 and it is much nicer than the coolermaster fan that came with my AMD Athlon processor.  It also makes about a 8 degree drop in CPU temp which was important to me.  You really need to pick the right fan for you based on acceptible noise, but a nice aftermarket fan in the $25-$35 range will do wonders.  You may also want to add some case fans.  One or two in the front and rear of the machine will draw a lot of air across the motherboard and also remove the hot air that seems to float around the processor heatsink.
3)  if i'm going to have to replace my RAM, should i be considering upgrading to PC2700 - is there a site that'll tell me if it's compatible with my motherboard?
A:  Crucial.com is the micron memory retailer on the web.  You can go there and search for compatible memory for your motherboard.  In my experience, they have a listing of whats compatible for pretty much every OEM machine and motherboard I've ever put together or upgraded.  This is their listing for your board:  http://www.crucial.com/store/listparts.asp?model=MS%2D6580+%28845G+Max%29&DetailMB=Y&cat=RAM  It says that the fastest supported memory is PC2100.  They will gladly sell you faster memory that is still compatible (PC2700 or PC3200), but your motherboard will just lower the bus speed to PC2100 spec (meaning you will not see the performance gain).


Just a quick note, that CPU temp of 56 degrees seems high for a machine that is just sitting in the BIOS and I agree with your ideas that temperature might be playing a big role in it.  It will pick up a lot of heat when you actually try booting to the operating system.  I would try a good CPU heatsink/fan combo first to see if that resolves the issue.  Good luck!

==Skack==
0
william200mphCommented:
Just an idea:
Open the case and get a big fan (like one you would find in an office, etc) and point it into the computer.
That way you will know if it is overheating and whether you need to buy new heatsinks/fans etc.
0
jhanceCommented:
None of the temperatures you've noted at even close to MAXIMUM so I very much doubt that temperate is your problem here.  Some uninformed non-technical types will quake and shiver at a CPU temp of 56C but in fact the Intel P4 CPU is RATED by Intel to run normally up to more than 75C.  So, in fact, your CPU is actually quite cool (from a CPU's point of view!!)

The memory is my #1 suspicion here and while you've tested it, that really means nothing.  Often you'll find perfectly BAD RAM will pass software RAM testers perfectly good.  You need to SWAP OUT your RAM for some others.  Since you don't have any on hand, try BORROWING some from a friend.  Until you eliminate this as a problem, there is no point in going on to other potential issue as all of them are MORE difficult and expensive to diagnose.
0
The Ultimate Tool Kit for Technolgy Solution Provi

Broken down into practical pointers and step-by-step instructions, the IT Service Excellence Tool Kit delivers expert advice for technology solution providers. Get your free copy for valuable how-to assets including sample agreements, checklists, flowcharts, and more!

rrhunt28Commented:
1:  Yes your system fan should be running at more than 0 rpm.  However it probably is.  Look in the case and look at the big fan's at the front and our back of the case, if they are turning they are going faster than 0.  The read out is only for a fan plugged into a motherboard power source and must be read by the bios.  Yours probably do not plug in to the motherboard they probably plug into the wire's  

2.  Your fan speed seems ok, it all depends on the fan you have.  There are hundreds of cpu fan makers, and they all have differnt speeds.  Your speed is over 3000 rpm, that is probably fine for a stock fan.  And as mentioned your heat seems normal.

3.  http://www.crucial.com/ www.tomshardware.com www.pny.com www.kingston.com 

Those are all places to get memory, and information on memory.  If your P4 has a 533 bus, then you probably could get alittle speed out of the faster memory, but you have to make sure your board can take it.  Those web sites will help.  

Good luck.
0
tonybushellAuthor Commented:
hey guys,

thanks for all your posts - I did some more tinkering after posting last night and something I noticed in my BIOS was that one of my CD-Rom drives (have two, a CD-RW and a DVD) wasn't displaying properly - instead of having a make/model number, it was just showing as 40x32x16.  Also, my 2nd CD-Rom drive wasn't displaying at all.  However, within XP both were displaying fine, and I could even burn a CD rom without issue - yet when i rebooted again, the CD-Roms were still not listed correctly in the BIOS, although still displayed in windows - so, I unplugged both drives and simply had the HDD on the IDE channel and the machine hasn't bombed at all (has been up almost 14 hours now).

And as regards the fans - i took the case off last night (have an Antec Sonata, and out of interest, this is a home-made PC) and the nice big, silent 120mm fan was spinning like a demon and i have space for another internal fan, but am not sure of the value of adding one (the temps looked ok to me, but i posted here as i couldn't be 100% sure and was becoming baffled at the source of the problem), although may consider replacing the one on the proc to help keep that cooler.

I'm not closing the question out, as I'm not 100% convinced that I've resolved my issue, but will of course award points once i'm totally sure this is all done and dusted.

thanks agian for your excellent posts!!
0
rrhunt28Commented:
Adding another fan basically gives you more air flow.  The more air flow the better.  Your cpu cooler gets the heat off the cpu, but it is still in the case, along with heat from the motherboard, power supply, and hard drive amoung other things.  The more air you can move in and out of your case the cooler everything stays.  And computer parts for the most part like to be cool.  It is also a good back up, incase one case fan fails you dont have to shut down and wait till you can get a new one.  
0
DalzielCommented:
Hello,

A lot has been said but since cooling is a thing of mine I might as well put my bit's worth in. Everyone goes on about temperatures but you have to understand why it matters. Whether your CPU has a limit of 70,80,90 or even 100 isn't the point, heat (and this applies to all the components) increases the noise in the circuitry which makes it less stable and why the cooler the CPU the better it runs and why overclockers, who increase the noise by affecting the amount of data and voltage going into the chip, care so much about the temperature since they don't want any of the excess noise caused by heat. If a chip is cooled to -25/-40 you can get anywhere between 30 and 80%+ more speed from it since the noise from the heat, which it is designed to deal with normally, is not there. However your CPU outputs around 60W maximum and, depending on your heatsink and the resultant C/W, will be fine at it's rated speed up to around 60C, which is a reasonable limit. If you got it running cooler (my "quiet" 2.53Ghz setup hovers around 35/45C) it may be a bit more stable but as you are not overclocking and it is still operating within the rated specs. then it doesn't actually matter. So to sum up 56C is fine as long as you are within the rated specs.

Now on to the case temperature this is different, you don't want it going over 40C, the closer it gets to room (say 20-30C) the better as this affects your CPU temperature and how everything else performs. Now I assume you want to keep the noise levels down as you have a Sonata (good case BTW), but a temperature (considering the inaccuracies of the on-board reader) of 37C is a bit high, so I would get another 120mm fan and a simple fan controller and just adjust the speed to temperature. Also it's a good idea to check that all the cables are tied up and not in the way of airflow.

This link can let you listen to fan noise levels: http://www.sidewindercomputers.com/tecar.html
Which fan you get depends on you, since they are a shop there are many good makes of fan they don't sell, but look around.

As for the other issues (in NO particular order):

Fan speed only matters if it's failing or you want all the technical info., CFM and the pressure rating generally will tell you how much air (and how) it moves, if you are worried check it's rated speed against the manufacturers site.

You can adjust it's speed (downwards) simply by changing the voltage it gets by using something such as a fan controller.

Your fan speed will be 0 RPM if it's not plugged into the board, if it's on one of the main power connectors as far as the board is concerned it doesn't exist. You need to either connect a fan to the board connectors, which with a large fan I wouldn't advise as it can draw more power than the board can supply, or you can get a fan or an adaptor that connects to the board with a spare lead to supply just the RPM info.

It's quite possible that your CD drives, especially if one was on the same channel as the HD, were making the system unstable.
Put the drives one at a time on their own channel and see what happens. Also how they are listed in the BIOS when compared to Windows depends on the drive and the BIOS, there are often differences. Of course one of them might be faulty.

Check for conflicts in device manager and  go into the BIOS and disable anything you don't need - the Parallel and Serial ports for example, this will help free up resources. Also run Anti-Virus and adware software to check for things, here's two:
http://www.lavasoftusa.com/support/download/
http://spybot.eon.net.au/

Oh and a CPU fan (if you are concerned about noise), because it's smaller, will make more noise than a large case fan. Since your case is at 37C (ish) it's best to get the case cooler first, then only if you want to change the heatsink and fan, as this will make the biggest difference.

Your Power levels seem a tad off, though not too bad. See how the levels fluctuate when under strain, they should be around the rating e.g. 12V at 11.840 +/- or 3.3V at 3.240V +/- and fairly steady (obviously :), and don't worry about the - voltages, just the + ones. Also are you using an Antec PSU?
 
And...oh I've said enough, any typos./questions just mention/ask.
Also just to reiterate that a lot of these things have already been said, just wanted to add my bits worth.

Regards
Dalziel    
0
tonybushellAuthor Commented:
great post, Daziel!  thanks a bunch!

my problems reoccurred last night again in exactly the same manner, so am getting some new RAM tonight ('known good', on loan from a colleague) so i'll try to rule that out.

I've done a fair amount of reading since this post about temperature and cooling and now no longer feel that this is the issue (although it may be contributing), so need to start identifying hardware problems.  Starting with the RAM, and hoping i nail it there, before I have to start playign with the MBD and Proc.

on that score - does anyone know a decent set of Diags for the motherboard / proc?  like i mentioned in the first post, I've done RAM testing till i'm blue in the chips and nothing shows up, so i wanted to try and find something that will test and/or stress-test my mbd/proc to see if i can find the error.

part of the problem that i'm having with this is that windows isn't doing any error reporting - there's nothing in event manager that even indicates there was a fault - if you read the logs, you'd think everything was fine and dandy.

man, this is driving me NUTS!

thanks again for the excellent advice, gents.
0
DalzielCommented:
Well it might be the RAM so see what happens, though while I agree with jhance regarding the limitations of software RAM testers if they didn't pick up anything it's a good sign (oh is the RAM all the same make and model?), though something else occured to me: If the Northbridge is overheating it can cause a lot of problems, might be worth (it can be a lot of hassle though) checking that the heatsink is on correctly with a thin even layer of decent paste (say Arctic Silver), or even changing it for an active (more noise though) one. Also (though this is a thing of mine so it might not be worth the hassle your end -  re-sizing partitions, installing and so on) you could try setting up a small partition using a mainstream Linux distro. and running that for awhile as it's a different (and clean install) O/S then it can help isolate the problem, not to mention give you a lot of info.

Though in the end sooo many things can cause a hang/crash it can take forever to find the cause, be thankful if you can narrow it down to hardware. Take heart from the fact that whatever is happening is happening for a reason, nothing just hangs or crashes there IS either a software conflict, hardware conflict, the wrong settings either in Windows or the BIOS
(looking through the manual there are a lot of settings with your board) an obscure physics problem - something. Depending on your luck it can take forever but keep looking and you WILL find it.

As for motherboard diagnostics it's not really that simple due to the amount of things on the board and the way it all hangs together, thus whatever you use tests the board in some way. There of course thousands of tools for checking the parts of boards, AIDA : http://www.aida32.hu/aida-download.php?bit=32 (excellent tool BTW, just compare it to something commercial like SiSoft) and SiSoft : http://www.sisoftware.net/ are main ones that come immediately to mind, but you just have to find something to test what you want to test. Sorry if this doen't help much, but the way I test and check these things is of course personal and involves many different test methods I've learnt over the years.

Regards
Dalziel

PS: There are of course thousands of tools for logging sytem events that are better than the Windows ones, again you have to look around. Also you can stress test a CPU with things like the tools mentioned above or something like CPUBurn:
http://users.bigpond.net.au/cpuburn/  -- be careful you get heat (natch) as well.
0
DalzielCommented:
Oh the AIDA link was the download page, this is the home: http://www.aida32.hu/aida32.php
0
tonybushellAuthor Commented:
thanks again, Dalziel - I got some more RAM last night (a brand new and unused 512mb PNY PC2100 chip) and it worked fine for a few hours but then the machine started locking up again in exactly the same way (by which i mean there's a sudden and short electronic 'click' from the speakers, the screen locks up, and the red light underneath my USB mouse goes out), so I don't think it's the RAM anymore - either that, or I'm just SO unlucky that I got loaned a dodgy chip.   And to answer your other question, the RAM is all the same make and model (although it's Elixir, a brand I'm not convinced about yet) - I have 2x512mb chips and 1x256mb, which i've tried in various combos, always complying with the MoBo shcematics.

I took the whole thing apart last night and reseated all the components individually to ensure that there were no loose screws or bad connections or accidentally thrown jumper switches, but everything seems fine and dandy.  I've also reinstalled XP (last night) and this doesn't seem to have made any difference (it locked up several times during install, so it doesn't seem to be OS related, more hardware).  My knowledge of Linux is good enough to install and get it running, but not sure how to find the info you say it would provide me.

Again, this is leading me back to the motherboard and/or connectors on that.  Possibly even the RAM housing on the board.  I've removed everything on the MoBo except the hdd & the RAM & the graphics card and it still does it, so I don't think it's any of the component parts, and I have also tried another video card to make sure it's not related to that.  Unfortunately, I don't have another Proc to try, nor another MoBo, so I'm kind of running out of options here.  Not sure CompUSA would let me return a MoBo after a few days, if it's been in use :P

I think I need to find some of the 'low level' error reporting tools you alluded to, as I need to try and find some way of narrowing down this problem!

this is the fourth system i've built myself since '98, and is the first time i wish i'd bought a DELL!

thanks again to everyone for their excellent imput!
0
DalzielCommented:
Well as I said I have some doubts about the power supply, but that might just be me, you need to check the levels with a software monitor, like this: http://mbm.livewiredev.com/ (though if you use it check for board specific info. or you can have problems) and make sure that regardless of what you are doing the levels are always steady, that's not to say they won't change, just not too much. Also do the same for the CPU temperature, keep a window open or create a log so that you can check the temperature over long periods, as has been said 56-60C is fine, but if it gets higher 70-80C (even for a moment) you can have serious problems.

You need to find some logging software that you get on with, start the usual places:
http://www.tucows.com/
http://www.download.com
http://www.majorgeeks.com/

Failing that, a detailed mental log of what you were doing at the time, if it's a consistent problem then there is a consistent factor.

Also as annoying as it might be you need to cut your options, remove and/or disable everything you aren't using (BIOS as well, where applicable) - the soundcard/chip, modems, disable and remove USB devices (use a basic PS/2 one for now), cut your memory amount down and so on. All this stuff can cause problems.

The Linux info. is more about using a different system and seeing the effects, if you use a stable distro., Debian springs to mind, or an older one (CD disks in Library books are great - old, stable, published versions with a huge manual, just make sure it can support your hardware), install the basic system, don't touch it and let it run for 24+ hours (or however long you think), then if it doesn't crash you know it's either not using some hardware that's faulty or there is a software problem with/in Windows.

You could try starting Windows in safe mode and running that for awhile, see if it hangs.

If your board turns out to be faulty (believe me a HUGE amount are, after all just look at how much stuff is on it) then there can be difficulties returning it but stand your ground - they sell them to you, for you to install and in the end they should work. Also faults often don't show up for a few days.

Regards
Dalziel

PS: "i wish i'd bought a DELL!"

You may feel like that now but if the nice, stable and ready built one doesn't work, it will really annoy you. That's why I built my first computer, three different store bought ones were all faulty!  
0

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
DalzielCommented:
Just to add that I never write things in order so don't think that any one thing takes priority :)

Dalziel  
0
tonybushellAuthor Commented:
funny turn of events - took my machine to pieces again last night (every damn component that could be removed WAS removed, including the motherboard from the case, and then lovingly reseated).  

Now, when I took the processor off I noticed that the top layer of the proc (where the label/name/spec usually are) had peeled way and had stuck to the underside of the heat-sync.  This CAN'T be good, so I'm going to get some Arctic Silver at the weekend and apply it, as it strikes me that the proc (while not necessarily the root of the problem) is getting too hot and/or not getting enough cooling.

I re-attached all the different power leads, re-seated the proc, booted up and haven't had any problems since - this was 12 hours ago and it's been up and working ever since - it's also been running CPU Stability Test v6.0 for the last 8 hours without any issues (supposedly this is a fairly intense test and if it can do 12 hours without failing, the author deems your proc to be working fine).  

Whilst I'm not saying the problem is fixed for good (i've made the mistake of enthusiastic-optimism before), at least it's now working again (and has all the different components back in, including the Gig of RAM that I first thought might be dodgy).  

Something that does concern me, however, is that after doing all this and rebooting several times, I noticed that the CPU fan was only spinning at around 2500 rmp's, which is WAYYYY too low, so i'm going to apply the Arctic Silver and then watch for a day or two to see if I need a new heat-synch/fan for the proc.  At this stage, I don't think it can hurt, I'm just trying to be as logical and methodical with my troubleshooting as possible, to prevent me thinking that one thing fixed it when it was something else, or a combo of things.

If I can go the whole weekend without a related bomb or lock-up, I'll close this question out on monday and eagerly throw my points at you guys.

thanks again for your EXCELLENT posts and the continued help!

and PS: Daziel - you're totally right about building your own - I used to work for HP and the whole Pavilion line of PC's made me swear to myself i'd never buy another 'big name' pc, unless it was a custom job (ie, Alienware, Falcon Northwest, etc) - like I've said, this is my 4th home-made PC in 5 years, and it's the only way to go!  The other 3 are still in use at home (now a PDC, a file-server, and my daughter's games-machine) - the reason I said it was because at times like this, you wish there was a warranty and someone else had to worry about diagnosing and fixing the issue!

0
DalzielCommented:
Glad it's working, hope it stays that way. The important thing with the temperature is that while 50-60C is fine, any reading in the BIOS won't be what the CPU is at when under full load so it could easily be 10C+ higher, which would make it close to the limit. Now P4's have a raft of things designed to keep the chip safe but that's not the same thing as stable. On the label issue, P4's generally have the info. stamped or etched into the surface of the heatspreader, are you saying that it left an imprint in the heatsink or that there was some plastic type label, in other words has the P4's surface been stripped clean? You see I've never damaged the on-chip info. yet, but it often leaves an imprint in the heatsink or pad. On the Arctic Silver issue, I use it - great stuff, but make sure the chip and heatsink are clean, I get Isopropyl Alcohol from the chemists (it's better than nail polish remover) and use it with a lint free cloth to clean the chip and H/S of all remnants and that the new layer is thin and level, I'm sure know this but an even mildly thick layer can add 5-10C. On the fan speed issue, well the reason I don't go on RPM is that it depends on the fan size and shape how much air it moves, RPM is more about the speed of air (which is important though). Now 2500 isn't too bad, in my opinion, for example on this system my CPU fan is an 80mm moving 45CFM and it's rotation is only 3000 RPM, now that keeps my 2.53Ghz P4 at 35/45C idle/load which is pretty good for a low speed case fan (though add 10C -/+ because it's very hot here at the moment). I'm not sure what heatsink and fan you're using but if it's the Intel one then I think they adjust their speed to reflect the temperature, now I don't touch self adjusting fans with a bargepole (I like to be in charge of what's going on), and since a non adjusting, low speed case fan can keep the CPU under 50C most of the time, even when overclocked, then why do I need temperature sensing? Anyway it could be running too low, if you changed your heatsink for a Zalman (nice and quiet) or a brute like the Swiftech MCX4000 (haven't used one yet but have heard good things) and fitted it with a good 92mm fan, then added another 120mm case fan, then plugged them all into a fan controller you could have great temperatures and also adjust the noise and speed to suit whatever you wanted, http://www.sidewindercomputers.com/index.html is a good place to start.

Regards
Dalziel

PS: On the building your own issue, I love keeping and fixing up old machines, it's often a waste of time and money but when it all comes together - junked parts, scavenged parts, really cheap old gems - and you get it performing better than anyone thought possible, well it really makes me smile.    
0
DalzielCommented:
Just looking through things I noticed that I didn't phrase things exactly I would have liked, so I'd like to add:

I wasn't really dismissing the RPM etc. of fans, obviously a fan is built up of several parts and all of them working together produce the final result: fan shape/type for the inital way the air moves, blade size for the amount of air it can move, RPM for the speed of air and so on. What I meant to say is that it's better to use the CFM and the mm-H20 pressure rating (or their metric/imperial equiv.) when choosing a fan rather than just the RPM, as a slow fan can often perform quite well.

Regards
Dalziel
0
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Desktops

From novice to tech pro — start learning today.