Overheating problem with T-bird 800

gimps

Senior member
Nov 6, 2000
383
0
0
I recently purchased a T-Bird 800/Abit KT7-RAID mobo combo from www.monarchcomputer.com
I'm having horrible overheating problems. It has a Tai-Soli HSF on it. Once Windows gets loaded, my CPU is at 144 degrees Farenheit! I had to unlock my CPU just to under-clock it to 500 MHz. Now, it's only 127 once Windows loads.
Is the only thing to do would be to get a new HSF? If so, what HSF would you guys recommend, that fit on the Abit KT7-RAID mobo?
Thanks.
 

Renob

Diamond Member
Jun 18, 2000
7,596
1
81
Try takeing it of and putting it back on be sure to clean of the thermal compound and put on some new compound and make sure it is a nice snug fit.
 

Tetsuo316

Golden Member
Mar 14, 2000
1,825
0
0
sounds to me like one of three things:

1) faulty readings. can be determined by touching the heatsink after you power down/unplug/and ground yourself. does it feel really, really hot? download all the latest utilities and bios updates from abit's website.

2) they installed the hs/fan without the thermal compound, or forgot to remove the tape on it or something. remove the hs/fan by pressing down on the tabs (usually it's really hard to do this, so i recommend using a standard-head screwdriver-just be careful). check and see if there's any bubble-gum or paste-looking substance on the hs/fan. if there's paste, then just put the fan back on. if there's bubble gum, make sure there's no plastic on top of it and replace the hs/fan.

3) the hs/fan isn't touching the core of the cpu. take the motherboard out of the case, and hold it up toward a light so you can see between the hs/fan and the cpu. is the core (the center of the cpu that sticks up and is blue/green) making good contact with the hs?

let us know if this helps or not.
 

FreakyOCR

Senior member
Oct 15, 2000
954
0
0
I might add another option.

You chipped a corner off your die, and it still runs. I chipped a corner off my 800 and it doesn't run, but damn, I'm surprised it didn't melt the frickin' socket!!! I powered down, about 10 secs later, I had the HSF off and I touched the die and burnt my finger. (Like touching a frying pan)
 

gimps

Senior member
Nov 6, 2000
383
0
0
The heatsink does not feel hot. Could be faulty readings, but I really don't know :(
My computer has been running VERY unstable since I built it. I have all the latest drivers installed, the newest BIOS for the mobo. I am lost as to why it is so unstable.
If I erase the lead off the L1 bridges, to bring it back to it's locked state, it seems to run a bit more stable. Yet, the temp readings are insane. I want to believe the readings, tho. Because after roughly 5 minutes of Q3'in, my computer resets. While I'm there, I go to the BIOS to check the temps. It showed my CPU at 140 degrees!:Q
I can't do much, because of the problem. Is it possible that it's a faulty CPU, or something? Windows loads fine. I can run mIRC, and ICQ with no problem. But if I'm staying busy browsing the web, my whole computer freezes (even ctrl-alt-del doesn't work), and then after roughly 10-15 seconds, it restarts. I have tried running it at 800 with the VCore upped to 1.80. Still nothing, which shouldn't be the problem, considering AMD says it only needs 1.70 (even tho my mobo says the defauly is 1.75).
Is it possible that it's a faulty RAM chip? If so, then why would it not make a difference, whether it runs at 100 MHz, 133, or even 138 MHz? Or whether it's CAS2 or CAS3 doesn't make a difference?
This is the first AMD system I've built, but the 5th system I've built. I am very stumped, and have no clues as to what I need to do.
I have taken the HSF off a few times, and there is a bit of thermal compound on the DIE itself, but nowhere else on the chip.
I'm guessing the HS is touching the die, cuz there is a bit of thermal compound on the HS where it contacts the die.
Any help/suggestions/comments of any kind would be great!
As I said, I am very stumped. The Tai-Soli HSF is supposed to be sufficient for up to 1 GHz, yet I can't get my system running stable at 800 MHz, and not even at 500 (which is what I'm running it at now)!

Thanks!
 

gimps

Senior member
Nov 6, 2000
383
0
0
Well I think I may have gotten somewhere. I got a Microsoft IntelliMouse Optical, and was using the USB connection. I decided to try it in the PS/2, and disable the USB. After disabling the USB, SB16 Emulation, and Creative Gameport, my system is finally running stable. One question, now.. any idea what's up with the Abit KT7 temp gauges? If they are incorrect, is it the software or hardware that's showing the incorrect temps? Clocking my CPU back to 800 (it is an 800 T-Bird), it showed a temp of 150 degrees Farenheit, as soon as I got into Windows. Anybody have anything to say on this?

Thanks again!
 

flashcadillac

Junior Member
Aug 15, 2000
20
0
0
What is the voltage setting on the chip? I just received an T-Bird 800 with the stock voltage on the chip set to 1.75 vice 1.7....I'm tempted to cut the bridge that takes it that high and default to 1.7 The Really weird part is that the motherboard (a new Tyan S2390) is reporting a vcore of 1.81 to 1.82, and so does Motherboard-Monitor. Temperature seems to be between 45c and 47c.

Should I cut the bridge to drop it back to 1.7? What should I use to cut it?
 

Argile

Junior Member
Nov 6, 2000
12
0
0
I wouldn't worry about the voltage on the t-bird. I have an 800@900 right now at 1.8v and its running 35C idle, 38C under load. You may just want to get some better cooling, although 45-47 isn't high.
 

flashcadillac

Junior Member
Aug 15, 2000
20
0
0
Yeah, I feel a little better after reading others' experiences with the chip. I'd like to know WHY the dang motherboard doesn't give it 1.75 (as the chip requests)....why the 1.81/1.82?

Have had a heck of a time with the new box (updated from a Tyan 1590), I finally ended-up pulling the promise ultra66 and using the onboard ide sockets (supposedly 66 as well). The CPU utilization according to HD-TACH dropped from around 6-7% to 0 when I did that. Now if I can only stop the random lockups (probably also the promise was the cause...)

Thanks again.
 

FreakyOCR

Senior member
Oct 15, 2000
954
0
0
My chip is at 1.80v and it reports 1.86v Look Below:::
It's also idleing at 26C, at 37C under stress
 

flashcadillac

Junior Member
Aug 15, 2000
20
0
0
I have a probably-dumb question...how are you generating the sig describing your rig...do you have a reporting tool to show settings?
 

FreakyOCR

Senior member
Oct 15, 2000
954
0
0
My KT7-RAID came with the via hardware monitor. That tells me the temp via the thermistor under the cpu. I also know what my voltage is because I set it at that.
 

flashcadillac

Junior Member
Aug 15, 2000
20
0
0
Anyone know where I can get the latest Via hardware monitor....I have a Tyan S2390 mb that doesn't have such a thing, yet I noticed that my old Tyan 1590 driver-disk had it (for the the 1590's chipset only).

Thanks in advance.
 

Netma

Member
Oct 16, 2000
169
0
0
I had stability problems similiar to Gimps. I switched from the pencil trick to Loctite and now its rock stable. Be sure you remove all the pencil lead.
As for the Temp problems, Could you give more info about your cooling. What case are you using and how many fans? Do you have a front fan blowing in, etc? My brother has a KT7-Raid-CPU combo from Monarch with Tbird 900Mhz and Taisol fan. Never exceeds 114F.
 

HumblePie

Lifer
Oct 30, 2000
14,665
440
126
Gimps I'm having the exact same problem as you. Just recently purchased a 700 tbird and bought the taisol cek734. put on some artic silver and the damn thing just keeps going up and up in temps until it crashes. I'm running the cpu at 600 just so I can read email without the damn thing crashing. I'm beginning to suspect a bad batch of cpu's were made and we got the crap of the crop. I used the locite rear window defogger repair kit to unlock the cpu and it makes no difference. The pencil trick wouldn't do a dang thing at all no matter how much lead, the type, or how clean the lines were. I'm at my wits end and sending the damn thing back for a replacement. I've tried the HSF on a friends and swapped cpu's from his board to mine... he has an abit kt7 as well and his doesn't overheat and become really unstable like mine does.
 

MortaniuS

Senior member
Oct 12, 2000
654
0
0
If your heatsink doesnt feel hot that COULD mean its no tdoing its job. I would take the heatsink of and clean it off totally, the person who installed it may have left the plastic on and it melted.
 

Mikewarrior2

Diamond Member
Oct 20, 1999
7,132
0
0
FreakyOCR,

Your temps are probably off. In order to achieve 37C full load core temp, you would need an ambient temp of 13C(55f). Most cases average ambient temp in the 20s, so your core temp would most likely be in the 44-50C range.


Mike
 

FreakyOCR

Senior member
Oct 15, 2000
954
0
0
TRY SETTING THE VOLTAGE TO 1.7 IT MAY SAY DEFAULT 1.75 BUT IT"S NOT!!! just a thought...
 

Mikewarrior2

Diamond Member
Oct 20, 1999
7,132
0
0
And like i said before, your temps are off significantly. Which Kt7 bios are you running?

Your cpu core temp(full load) is probably between 45-50C.


Mike
 

FreakyOCR

Senior member
Oct 15, 2000
954
0
0
Mikewarrior, you must be wrong, there is quite a few peeps who say their core temp under load is around 37C. I am not the only one!
 

Mikewarrior2

Diamond Member
Oct 20, 1999
7,132
0
0
You saying that proves your ignorance!! First off, NO AMD CPU IS CAPABLE OF MEASURING CORE TEMPURATURE.

SECONDLY, I have stated numerous times that it a problem with the method of temp measurement with socket A mb's. In tests comparing a socket thermistor versus internal diode, the temp differences can be anywhere from 10-15C off from each other, with the obvious nod going to internal diode's for accuracy purposes.

Thirdly, Why don't you look at this based on MATH AND PHYSICS. THis is indisputable.

For example, your cpu, at 1ghz @ 1.8Volts, puts out roughly 57W of heat. Your heatsink, the FOP-38 has a thermal conductivity of about .4 C/W.

So, we take 57W Multiplied by .4C/W and we get 22.8C. In order for your cpu to be running at 37C, you would need an ambient at the very highest of 14.2C. Which is 57F. Now bear in mind that these numbers represent the absolute best temps you could possibly get. Since you are using thermal grease or the built-in PCTC, you might have a cpu that is running 24C over ambient temp. IN which case, your AMBIENT CASE TEMP would have to be lower than 57F.

This may sound pompous, but almost every single overclocked Duron/t-bird with load temps under 40C is a pretty sure-fire bet for a misreading. The sheer number of people saying their durons run at 1ghz at 38-32C is huge, and you know what, they are all wrong.

And like i've said before, you aren't the only one with a claim that their cpu is running under 40C. Lots of people claim that. And in the majority of cases, they aren't lying, but they are posting erroneous data from the mb sensor, and in most cases, erroneous data that is physically impossible with conventional cooling.


Mike