Best air CPU Cooler?

[McArra]

Originally posted by: VinDSL

Originally posted by: McArra
I rather put a thermal sensor as I did LOL

That's fine!

So, what have you concluded, putting a thermal sensor as you did?

You must be joking :laugh:

I've just put a Thermal sensor, in the lower part of the core, it never goes above 38ºC when MoBo says 58ºC... give the core itself 10ºC more, it can't be hotter.

 

[Noubourne]

You manually placed a thermal sensor in the lower part of the core?

Wow. Where did you get your engineering PhD?

 

[McArra]

No, sorry, I don't know how to call it.
http://img122.imageshack.us/img122/9389/tempsensorto9.jpg

Iknow it isn't the very best place, but is nice for a reference

 

[Howard]

hahahahahaha

That's probably one of the coldest spots on the IHS, assuming you even get good contact with it.

 

[VinDSL]

Maybe you should forget the HSF and try submersion cooling instead...

 

[McArra]

Originally posted by: Howard
hahahahahaha

That's probably one of the coldest spots on the IHS, assuming you even get good contact with it.

And the only place where you can put the thermal sensor, and the reason why I put another 10ºC extra to the results.

68ºC would mean a thermal resistance of 0.30ºC/W, which is too much for the Big Typhoon, and assuming a high power consuption of 150 watts, that's worse than a Zalman 7000al/cu.

So let's guess a much more normal thermal resistance of the Big Typhoon of 0.22ºC/W at worst, then the delta would be of 33ºC, giving a total temperature of 55ºC at full load.

Taking into account, that a new 9 blade fan at the same rpms made the temperature 2 to 3ºC cooler, it seems like the limit was in the cooling itself rather than the contact.

So I think, my conclusion of a bad MoBo thermal sensor is well made taking thermodynamics into account

 

[Howard]

Originally posted by: McArra

Originally posted by: Howard
hahahahahaha

That's probably one of the coldest spots on the IHS, assuming you even get good contact with it.

And the only place where you can put the thermal sensor, and the reason why I put another 10ºC extra to the results.

68ºC would mean a thermal resistance of 0.30ºC/W, which is too much for the Big Typhoon

You can't work backwards like that. The thermal resistance is relevant only to the system where it was measured.

The motherboard should be reading the temperature from the sensors in the CPU anyway.

 

[JEDIYoda]

Originally posted by: sbuckler

Originally posted by: JEDIYoda

Originally posted by: sbuckler
Using a TT will not make any difference - BT is just as good. Your choice of heat sync is not a problem. It's either a problem with how you mounted it, a problem with concave a heatspreader on the cpu, or a problem with the temp probe.

How can you even say that????
to be just as good implies that the temps will be as good or better...
When in fact there is noway to be 100% certain of that statement based on way too many variables to even make such a misinformed conjecture!!

Look at corsa's post - he has a test showing a TT getting identical results to a BT:
http://forum.abit-usa.com/showthread.php?t=103462&page=4

Anyway the point being the difference between them is too small to matter - but the op is getting temps that are way too high. Recomending he switch heat syncs to another one that may or may not be a degree or two cooler isn't going to fix his problem - the fact he is using a BT not a TT is not the source of the problem, something else is.

Sorry real world applications verses tests by a website are NOt the same!!
Had you read MY post you would have understood that!

When the Ninja came out every site that reviewed the Ninja put it heads and tails above all other heatsinks.
Yet the truth be told when people jumped on the Ninja band wagon and bpought the Ninja they came to a sudden realization-- real world applications are not the same as being tested in a controlled enviroment!!

Then if you look at NAVIG`s guidelines or the manner of the way he tested the heatsinks there is noway it was based on factual data...or even in a controlled environment thus....
His testing methodology is at best Bogus....

Then let me leave you with this taken from the link you posted...
The evening temps were fairly constant over the last 2 days (about 50F), so I decided to generate several hundred temps for you.Also keep in mind the Abit bios reads approximately 10c above reality (reality being an external probe drilled into a heatsink base) as documented elsewhere.

Then we have this -- btw which really says it all.....
Load was generated by dual Prime95 at approximately 8 minutes or so (when it would pass the first round of calculations). The system was turned OFF between every test to cool it down. All applications were Arctic Silver Ceramique with about 1 hour of burn in time for each heatsink. After removal, all my applications generated smooth thin and even coatings on the cpu.

Sure it very nice that somebody would go to all that trouble but again.....
its all conjecture based on no regulated environment....
such as making note before every test of the ambient temperature...
the system was turned off after every test..ok...for how long?
Then the burn is was for ABOUT 1 hr........

I have all the heatsinks listed including the Tuniq...the ultra 120 the XP120 the SO120 the Zalman 9500 as well as the 9700 as well as Ninja and the Big Typhoon and a myriad of other heatsinks....

Trust me when I say real world applications really do vary sometimes by a huge amount not just a few degrees.....

From one PC to the next what is supposedly the best for a given system could very well be shizza for a different system!!

 

[McArra]

Originally posted by: Howard

Originally posted by: McArra

Originally posted by: Howard
hahahahahaha

That's probably one of the coldest spots on the IHS, assuming you even get good contact with it.

And the only place where you can put the thermal sensor, and the reason why I put another 10ºC extra to the results.

68ºC would mean a thermal resistance of 0.30ºC/W, which is too much for the Big Typhoon

You can't work backwards like that. The thermal resistance is relevant only to the system where it was measured.

The motherboard should be reading the temperature from the sensors in the CPU anyway.

The sensor is usually wrong, like the voltage sensor, which is always wrong.
Of course you can go backwards, you can do it using safe values.

 

[corsa]

ok.....TT BT heat pipes under normal conditions do not feel warm to touch, disconnect the fan from MB, stress test and feel the pipes where they bend......they should get hot.
If they are u have bad temperature sensor, if they are still cool then u have bad heat sink/mounting........next
ohh and when u see lots of these ???? & !!!! .....ignore the post

 

[McArra]

I want the think you are smoking.

 

[JEDIYoda]

Originally posted by: corsa
ok.....TT BT heat pipes under normal conditions do not feel warm to touch, disconnect the fan from MB, stress test and feel the pipes where they bend......they should get hot.
If they are u have bad temperature sensor, if they are still cool then u have bad heat sink/mounting........next
ohh and when u see lots of these ???? & !!!! .....ignore the post

true....I agree....after all I wouldn`t want you to learn something about how to properly test a heatsink...being specific rather than vague in your testing methodology!! Oh....sorry I forgot?????????????????????????????????????????????????????????????????????????
???????????????????????????????????????????????????????????????????????????????
???????????????????????????????????????????????????????????????????????????????
???????????????????????????????????????????????????????????????????????????????
?????????!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!???????????????????????????????????????????????????????????????????????????
???????????????????????????????????????????????????????????????????????????????
???????????????????????????????????????????????????????????????????????????????
???????????????????????????????????????????????????????????????????????????????
????????????????????????????????????????????????????????!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!??????????????????????????????????????!!!!
!!?!?!??!?!????!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!?????????????
????????????????????????????????????????????....etc etc etc.....

 

[Rubycon]

Originally posted by: Howard

Originally posted by: McArra

Originally posted by: Howard
The IHS might be concave, unfortunately.

Preasure shoul correct that, I guess.

Nope.

Actually enough pressure would. Unfortunately it would crush the cpu core and ruin the socket.