How hot is too hot? AMD T-Bird

[undercover]

I got a 1.0ghz BlueCore T-Bird and ran the multiplyer to 12.0. I now run at 1.2ghz and the 2 temp prongs I have running to it read 36c and 40c MAX. I have run Counter-Strike on it for hours and it hasn't gotten any hotter.

If I decide to run it a little faster, or maybe my apartment get's warmer, when should I cut it off? How hot is too hot? I have 2 digital read outs on the front of the case, so I am always watching it.

Anyone?

 

[Viperoni]

Are you running at default volage?
That temp is quite low, which is probably due to the innaccuracies in your temp probes

Enter Mike!

 

[Viperoni]

The max they should EVER get is about 50c, IMO

 

[Hanky]

Mikewarrior2 will most likely say something more, but as far as I know the temperature readings on Socket A boards via thermistors are always inaccurate, while it depends on the board (and many other circumstances) HOW inaccurate they are. So it would be interesting to know which board you own, but I think one can already say - even when adding...let's say 10°C to your temps - your temps seem to be alright.

 

[RabeaticSquirrel]

First of all where do you have the thermal sensors located? Do you actually have them touching the core or do you have them sitting on top of the HS/Fan? I'm not too sure that your temp readout is correct, of course I could be wrong.... I have an 800 T'Bird @970 under full load it pushes 42C, and that's with a Hedgehog 32CFM delta and artic silver. If your running a T'Bird 1000@1200 you should be pushing a higher temp. Unless that is some type of new core. I'm not real sure what you mean by a Blue Bird. Now don't get me wrong, its not entirely impossible to acheive those cooling temps with conventional airflow, it just seems odd under full load that you're pulling only 40. None the less, to answer you question, don't get above 50 to 55C. Anything above....say......50C and I would begin to worry, 55 as an absolute MAX!!! But, if it does get that hot, don't panic, remember the T'Birds have a total MAX operating temp of 90C. Of course if your CPU hits 90C, you can probably call it quits in about 15 minutes. Like I said, that is the TOTAL MAX operating temp. If this confuses you, let me know and I'll try to explain further.

 

[RabeaticSquirrel]

let me ask a little more detailed. What do you mean by the "2 temp prongs?" You said you have two digital readouts on the front. That leads me to believe that you have two digital therms, with the thermal sensors on the core. Explain a little further so we can better understand you setup. That would help us diagnose your problem a little better, before MikeWarrior2 gets here.... J/K Mike....

 

[undercover]

When I say Bluebird I mean it's got the blue core. It's unlocked (brindged). I have a hedgehog with the High Output Delta fan.

I have 2 probes going to the bottom of the heat sink. Bottom meaning, on the part of the heatsink that is right on top of the CPU.

They aren't the ones from the mobo manufacture.

 

[undercover]

I would like to get the prongs closer to the actual CPU but I have a copper ship in between the hedgehof and the chip. The wire to the prong is kinda fat so I think it would compromise the connection between heatsink and cpu.

How should I go about getting it closer?

I also use Artice Silver if that helps.

 

[RabeaticSquirrel]

so your probes are on top of your CPU HS? They're not actually on the bottom (finished side that touches the CPU core) of the HS (heat sink)? IF this is true...read on. If not....ignore and answer the above question please.

My suggestion would be to remove the copper shim. Go buy some thermal epoxy and attach one of the thermal probes to the CPU PGA (pin grid array) the gray area around the CPU Core. Don't cross any bridges or touch any other pieces sticking up, and try to get the sensor probe to touch the side of the CPU core. Hopefully someone will read this and provide you with a picture, I unfortunately don't have one to show you. CAREFULLY reattach your heatsink, adding even ample pressure all over the heatsink to avoid breaking your CPU. Once you re-attach the HS and fan (of course you would need to clean the HS bottom and CPU core with isoprohol alocohol to remove the old thermal goo) with your new coat of thermal goo, make sure the probe isn't sticking in between the CPU core and HS. An easy way to see that is, when you power up the machine if it drastically jumps in temp in like increments of 5+ degrees reaching 50C before you get into windows, then you've got a problem. Go back and check to make sure you didn't get anything in between the core and HS. If not, happy monitoring. Let me know if I can be of any further assistance.

 

[Mikewarrior2]

Because you're not touching core, you're nowhere near representing core temp(keeping in mind that thermistors touching cpu core read a bit less than full core temp). IN actuality, you're measuring Heatsink/Shim temp.

Depending on your ambient case temp, you can either be only off by a bit, or can be off by a lot.(CPU is probably around 20-24C over ambient caes temp for you).

I don't think you'll be able to get the probes close unless you take off the shim. NOt sure if you want to do that or not.

Can you give us a Socket-thermistor reading(From MB for cpu temp) just for reference? And what MB will you be readnig the temp from?


Mike

 

[undercover]

Ok, probe location:

---------------
000000000000000 <------- Fan
---------------
|||||||||||||||
|||||||||||||||
|||||||||||||||
||||||||||||||| <------- Heat Sink
|||||||||||||||
||||||||||||||| <------- 2 Probes &quot;Float&quot; Right here
|||||||||||||||
_______________ <------- Bottom of heatsink
XX <------- AMD Athlon 1ghz
_______________ <------- MSI K7T Pro2a Mobo


Does that help?

 

[RabeaticSquirrel]

Ok that's what I thought. You won't get a true reading of your CPU core with the probes floating in the HS like that. You can most likely expect that the actual temp is around 5 - 15 C higher than what those probes are reading. 15 being a very high, but it is possilbe. Try my earlier suggestion (which Mikewarrior2 suggested as well). If you want to try it, take the copper shim off and try to make contact with the CPU core....

 

[undercover]

Do you think the pressure the clamp on the hedgehog applies will be enough to flaten the wire? Because I would hate for the cpu to burn up cause the HS isn't touching it.

 

[RabeaticSquirrel]

NO NO NO NO. let me see if I can explain this better....hang on for a few minutes let me think.

 

[undercover]

I know what you mean. Like make the end of the prong touch the cpu.

BUT

the actual FLAT part of the prong is like ALMOST half and inch I am guessing. The space between the edge of the CPU and the CHIP is that length or longer. So the FAT part of the wire that connects to the PRONG mak get under the HS, therfore adding space between it.

Get me?

 

[RabeaticSquirrel]

This is how it should go:

+++++++++++
+++++++++++ <------- This is the Fan
+++++++++++

###########
###########
###########
###########
########### <------- This is your Heat Sink
###########
###########
########### <------- You said you had two probes floating here.....
###########


++++ <------- This is the actual CPU core itself
=========== <------- This is your actual CPU (PGA, Pin Grid Array, the gray area around your CPU core)

Now we both understand how you have your CPU/HS set up.....and of course you have the copper shim around the CPU core and on top of the PGA....

This is what I mean......

++++++++++
++++++++++ <----- FAN
++++++++++

##########
##########
##########
##########
########## <----- Heat sink
##########
##########
##########
##########

++++**** <----- CPU core, one of your probes should lye flat on the PGA, just touching the CPU core
========== <----- PGA, gray area around your CPU core

Probe is listed with ****

See what i mean? The probe should never go in between your CPU core and Heatsink. It should be between the gray area surrounding the CPU core and your heatsink just touching the edge of your core. MAKE SURE IT IS NOT IN BETWEEN THE CORE AND THE HEATSINK!!!! Got it? Let me know if that confuses you and I'll be glad to try and explain again....

 

[RabeaticSquirrel]

Yeah I see your point. Sorry. You may have to strip off the heatshrink around the two wires on your probe to make it fit, and if it in any way happers the ability of the HS to lye flat on the core, then trash it and get another probe. What kind of probe is it? Is it like a digital thermometer probe, like the ones that millisec.com sales?

 

[undercover]

I understood that from the begining ;-)

I am worried about the thickness of the prong

I'll just have to play with it

thanks!

 

[Mikewarrior2]

You're gonna have a real hard time keeping the heatsink flat, even if you strip off the wire wrap.

but if you aren't having problems dont' worry about it?

BTW, can you check your CPU temps from either VIA hardware monitor or the MSI program? I'm guessing it is around 40C as well?


Mike

 

[MeloManiac]

My TBird 900 with a ChromeOrb is around 50 degrees idle, 56 full load, and I hear about lots of people getting that kind of temperatures, so I don't really understand how anyone can get 40 with a 1k clocked at 1200.... BTW how can I tell if I have a 'bluecore' or whatever, and what does it mean? Anyway I'll get a silverado 50 so I can get the temp down and maybe overclock it.

 

[Daudi]

WOW, those are high temperatures. I've got my 1000mhz tbird @ 1153mhz and I run about 42C core while running RC5 full load for a few hours..36-37C at idle..
blue core tbirds (usually found in the 1ghz+ range) have copper interconnects and stuff which are able to overclock alot better
green cores have aluminum ones I believe and aren't as good.
look at your cpu core, if it's blue then you have a blue core, if it's green then you have a gree core..pretty simple..

 

[Mikewarrior2]

c-orbs don't run well hot on amd CPUs. This is the performance chart on a c2/internal diode readng(some 9C higher than a FOP32-1).

Heatsink Tests: Internal Diode versus Socket-Thermsitor


Expect AMD temp differences between heatsink(core temp wise) to be greater than these temp differences.



Mike