Sign in with
Sign up | Sign in
Your question

Thunderbird running too hot

Last response: in CPUs
Share
January 6, 2004 3:36:35 PM

HELP everything is gettin too hot and crashing. Let me explain.

I have a thunderbird athlon 900mhz running on an Asus A7v, this has been running fine for about three years now with no probs. However recently is started to crash, it would freeze and just go dead. This was seldom at the start and I just assumed that is was to do with the game I was playing as it was the only one I played at the time. The problem got worse, ie it happed more frequently, and within different games too. I thought then that its not the software causing this, it must be hardware. Recently I have been playing starcraft with my brother over the LAN and the last time we attempted it, it crashed four time with me and would only play with the case off for additional cooling.

Now I bought a ZALMAN 6000alu copper monstrosity as I thought, I NEED MORE COOLING, plus I use the PC to play music and dont really like the fan noise as it intrudes. On the second note, its fantastic the noise is now acceptable. But the cooling isnt any better maybe a few centigrade less but not enough.

There in lies the problem. I have seen many reviews on the Zalman sink one of which was for a 1.4 thunderbird, his temperature was 31 degrees at idle and 51 at max usage. I got an idle of 44 with the case open (his case was too for the test) with three case fans threatening to levitate the PC all, on 12v.

Now I have set the core voltage to 1.5 volts to lower the temperature, with everything set to 5 volts (it made a 2 degree difference from 46-44 going from 12 - 5 with the case open at a core of 1.7v) I now have a temperature of 51 degrees and all im doing is writing this!

Is my chip coming to the end of its life and is half bust, so is producing LOTS more heat then its supposed to. A 1.4 gives off more heat then the 900 too, thus adding to my conclusion that the chip is having a transistor breakdown ( nervous breakdown for electronics??)

What do all you people think????
Any advice is welcome.

Cheers Matt
January 6, 2004 6:43:55 PM

Come on sum1 here must have some idea wats wrong does my post smell or something?
January 6, 2004 7:31:30 PM

51 degrees isn't to hot for a thunderbird CPU and as long as your temperatures aren't fluctuating too wildly like going up over 60 degrees you cooling is probably fine. The temperature sensors on motherboards are vague at best anyway it could be running a lot cooler than its saying! I'd be inclined to suspect the Memory you are using might not be up to scratch substandard quality RAM is often the cause of lockups in games. Also consider upgrading your graphics card as your system is aging therefore I assume your graphics card would be pretty obsolete and unable to cope with the high resolutions of more modern games.
Related resources
January 6, 2004 8:20:36 PM

Try giving everthing a good cleaning. Dust can cause overheating.
January 6, 2004 8:46:22 PM

Cheers for your responses ageing? who said it was ageing!
The graphics card is a voodoo3 2k and runs the games I wanna play perfectly, Diablo2 expansion in 800*600 runs as smoothly as a babies bum, now I dont play all these high tech all singin and dancin computer games and thats not the point of the PC anyway I use it to play music, uncompressed wav files on the hard disk. Either way I only put it together (upgraded main PC to XP2.6 so I got the mobo and ram and proc and added the rest).

This is also the 1st time in three years that the problem has occured, so I doubt the ram is at fault.

After I set the core to 1.5 volts the problem has not persisted but its running at about 5 degrees less then before. I know the problem IS heat and is clearly heat in the processor as reducing the voltage, and therefore the temperature by a couple of degrees has stopped the crashing. I played diablo2 for a good two hours without any problems. MB monitor says that in this boot up the min temp is 46 and the max 53. This is fine and works fine without any problems.
The voodoo isnt to blame, it has never caused any problems the five years we have had it and used it, the ram is also fine as is the rest of the system.

Its just the processor is supposed to work at 1.75, set to 1.75 Motherboard mon says its at 1.81 or something like that I doubt an increase of .06 will cause crashing as I have read overcloking at 1.85 is stable. So my concern is that I cant run the processor as its supposed to be.

I have no problem running at 1.5 volts but would like to know why I cant run at 1.75. It crashed at about 55-58 with voltage at 1.75.
January 6, 2004 8:51:01 PM

Try cleaning as suggested above, and if no luck I would point the finger at first a Failing Power supply [quite likely] followed by mobo/Ram.

Some steps to try:

<b>Test RAM:</b>
Download Memtest86, follow the instructions (it runs of a boot floppy) and leave running for a couple of hours. if you get no memory errors it's probably not that.

<b>Temps/Power supply</b>
Download motherboard monitor 5, set it up so all the Voltages/temperature readings look about right and then set it to log to a file every 5-10 seconds, for 50 entries or so, and then leave running while you play a game or whatever. This will show you either how the temps got immediately prior to crashing, and more importantly what the voltages on the 3 voltage lines were just before crashing (12V, 5V, 3.3V) if any of those voltages drop by more than ~5% then a replacement power supply is in order.

Any questions (or if you find a solution!) do post back.


---
<font color=red>Those of you who think you know everything are annoying to those of us who do.</font color=red> :wink:
January 6, 2004 9:01:04 PM

Beat me to it there... :smile:

That's a bit more info... If it <i>is</i> temperature then that suggests it either needs a dust removal, or the airflow thru your case is not good. you mentioned 3 case fans - where are they and what direction are they blowing? for 3 fans I would suggest one sucking air into the case (at the front normally) and the other two blowing it out(rear) - otherwise airflow won't be too great.

suprised it works happily @ 1.5V - that suggests to me that the mobo's ability to regulate the voltage supplied to the CPU is failing, and the extra voltage is making it overheat, but when you undervolt it, it's adding just enough on to make it stable again.. :eek: 

I had an Asus board which was fine for ages and then lost it's ability to control all the voltages properly and caused similar problems.

It might still be worth checking the 3.3V, 5V & 12V lines though...

---
<font color=red>Those of you who think you know everything are annoying to those of us who do.</font color=red> :wink:
January 7, 2004 12:46:22 AM

Hmm the power supply is relatevely new :(  I bought it with the case when we upgraded the comp to xp and i got the mobo processor ram, its literally a max of 4 months old. I will try the ram thingy altho I doubt its the ram, the voltages on MM which I had used to see what things were at anyway, well the core for sum reason is on 1.81.... should be about 1.5 hmmm (frown lol dont know command) I checked in bios and it said that it was on for 1.5? confusing, when I had it on 1.5 it kept the voltage at 1.51 with a max of 1.57 that seems a little high it does sound as if there is a prob with the mobo then, I dont know how tight the voltage tollerances are on core settings so it could be defaulting to the 1.75 setting, is 1.81 too high for a settin like this? I dont really want to have to purchase a new mobo :( 
The mobo does have some dust on it and if the area which regulates core V is getting too hot it mite kill its ability to regulate efficiently. OK well as its running at 1.81 it should crash! So ill pop on the recording facility in MM and see what happens.

Cheers Matt

I hope I get to the bottom of this :) 
January 7, 2004 12:52:00 AM

Ok since the PC has been on and I last posted about five minute have gone by but I opened up MM and looked at the highest and lowest core voltages 1.76-1.82 low high, and I watched it for a few seconds and it jumped about within those bands quite a lot. For the max min the PC has been on for about twenty minute now so not very long. OK well here goes the crash test!
January 7, 2004 1:22:21 AM

51 degrees isn't bad. mine idles at 51runs normally between 52-54 and full load at nearly 60 and its perfectly stable. using a volcano2. with the stock cooler it used to be >63 and was terribly unstable

:cool: :eek:  :redface: :frown: :lol:  :mad:  :eek:  :smile: :tongue: :wink: :evil: 
January 7, 2004 1:25:21 AM

OK it crashed didnt actually have to play a game it did it on the whilst surfing the web. Trying to find a diablo2 item pack! Anyway here are the results.

Motherboard (case) temp starts at 41 then goes to 42 and then 43 and stays there.

The CPU fluctuates around 54 and 55

Hmm from when I started the logs the Vcore varies from 1.79 and 1.81 but never goes to 1.80? This is with a setting of 1.5 or atleast thats what it says, I think it has defaulted to 1.75.

+3.3 stays on 3.47 +- 0.02
+5 varies between 4.75-4.87
+12 has a general trend of rising up starting at 12.6 and then rising up to 13 before the crash.
-12 -12.6 +- 0.3 for the whole time. no trend
-5 -5.6 +-0.2 for the whole time.

Nothing really odd about that.
Although it doubt as I said before that its the PSU, when I has the core V at 1.51 it didnt hang once in about five hours, it WILL crash again if I dont get this typed fast enough!

SO it is something to do with the CPU and mobo, either getting to hot, which is unlikely as I have been told it can get a lil bit higher then 55 without crashing, so I would guess something else is to blame. Is there anyway to test if its the mobo?

Again cheers for any help Matt
January 7, 2004 5:06:50 AM

I have an Asus A7V133 and in order to change the vcore, I have to use the jumpers. Changing them in the BIOS does not seem to work when Windows XP loads. Might want to give that a try.

My PC setup is a little different, but this might help you:

Athlon Thunderbird 1.33 GHz
Asus A7V133

CPU temp 52C
MB temp 38C
+12V 11.734
+5V 4.945
+3.3V 3.504
VCore 1.792

My vcore is set at 1.75, and system runs stable. On another note, I had the exact same problems you mention when I tried to overclock. Have you touched any other settings in the BIOS?
January 7, 2004 5:17:20 AM

Quote:
Now I have set the core voltage to 1.5 volts to lower the temperature, with everything set to 5 volts (it made a 2 degree difference from 46-44 going from 12 - 5 with the case open at a core of 1.7v) I now have a temperature of 51 degrees and all im doing is writing this!

Sorry, I just noticed this in your first post and am a little bit confused. Were you able to successfully change the vcore? Also, what do you mean you have everything set from 12V to 5V? Are you talking about the I/O Voltage Setting (VIO)? Because that should be at 3.45/3.56V, I think.
January 7, 2004 12:09:45 PM

When I say 12 to 5 V I was talking about all the case fans (yes I have one sucking air in at the front and two blowng out at the back) and the fan on the Zalman with its variable speel control. 12v was meaning fast on teh zalman and 12 on the case to provide max cooling. I like them on 5v as I like the PC running next to silent.

The PC has run stable for three years and has only started to crash now with a stock heatsink, the zalman didnt alter this.

About Vcore, for some reason after it crashed and I posted the results, it was then running at 1.55v so I left the comp on all night and I am writing now from that same boot 11 hours later. So there you have it, using a core of 1.5/.55 volts make it stable and on 1.75 more like 1.80 is crashes.

It crahsed at 55 54 degrees also. I know the chip can get hotter then this without freezing because it used to when we 1st got it.
January 8, 2004 12:56:41 AM

Oh, I see. Have you tried memtest86 yet, as someone else suggested? If your ram checked out fine, I would try adjusting your vcore through the jumpers, instead of BIOS. It's strange, but my Asus mobo's vcore has to be changed through the jumpers. BIOS changes give correct vcore in BIOS, but not in Windows XP.

Sounds to me your mobo is damaged, but if running at 1.5V is stable, might as well just lower the vcore through jumpers. Good luck.

Edit: By the way, have you run Prime95 with vcore at 1.5V? Might let you know just how stable your system really is.<P ID="edit"><FONT SIZE=-1><EM>Edited by knight0 on 01/07/04 10:00 PM.</EM></FONT></P>
January 8, 2004 1:29:09 AM

The rise to 13 volts before the crash seems significant. Since you have another computer in the house, why not swap psus for a bit. If that doesn't do it, I would take a serious look at the area around the socket. Are the top of the capacitors rounded? Is there black dust? Black dust is from carbon tracking (a type of short circuit) but can be cleaned. Rounded capacitors can be replaced by someone who is good at soldering. Good luck.
January 8, 2004 8:30:39 AM

I concur with endyen, especially about trying the other PSU if you can. You 5V is a little low I reckon and another PSU should help there.

The other thing is, for someone with apparently decent airflow in their case, your system temp is quite high, so perhaps the poor voltage control your board seems to have means it's supplying too much voltage to the chipset? Last mobo I had with a system temp. that high was my aforementioned Asus board.. (Current <i>overclocked</i> Nforce2 mobo system temp is ~28C IIRC, with just 1 case fan at the back)

Try a new PSU, and if that doesn't help, perhaps the mobo is on its way out? maybe some components could be replaced as endyen suggests, or it might be easier to just buy a new mobo...

---
<font color=red>Those of you who think you know everything are annoying to those of us who do.</font color=red> :wink:
January 8, 2004 8:43:46 AM

I think its time to try the other tests, the systemp has always been high with this mobo and processor I dodnt realise it was anything to worry about. Replacing a few capacitors is not a problem to me whatsoever, I build my own hifi equipment and fabricate my own PCB's etc so thats ok. So naturally I am familiar with dead components having killed a few in my time anyway, none of the caps looked dead when I last saw them altho I will re open the case and check everything once more.

Its interesting with the vcore setting, for some reason it doesnt set it right in win 2k when you first boot up, but if I reboot its fine? After though I have been into the bios and reset it altho I didnt change any settings.
January 8, 2004 10:58:49 AM

Ok i set the CPU tester going on 1.5 volts, then left it for about 30 mins and played daiblo2 LOD for about two hours, I set the memory in the tester to 128mb, anyway nothing went wrong yet. Im assuming that the thing is stable, when I was playing I got a temp of 54 degrees and just recently 55. 55 has been enough to crash it on 1.75 volts but not on 1.55, so far.
!