Will Lapping Equalize Temps

rulocal

rulocal

rulocal

Associate
Joined
2 Oct 2005
Posts
507
Location
London
I am getting significant difference in temps on my Opteron 170 with Tuiq 120 HSF. Under load temps are 57 and 47. Will lapping help? Alternatively should I consider removing the IHS. If I remove the IHS, will I have to make any modificantions to fit the Tuniq or just be very careul when screwing down the unit?

Cheers
 
lapping will sometimes equalise/lower temps, de-capping will allow you to equalise (and will reduce temps) as long as you can get a good mount.

i duno about how that HS mounts.. best wait for another to say if it works or check the Decapping threat to see if anyone else has decapped with that cooler.
 
I presume you are running a program that loads both cores to get those temps? Rather than a single threaded app only loading one core?
 
PinkFloyd said:
I presume you are running a program that loads both cores to get those temps? Rather than a single threaded app only loading one core?
Click to expand...


Yep. Tried with Orthos and Dual Prime. I used to get a 3 degree difference with my Opteron 165 but 10 degrees is ridiculous. It is particularly annoying as this seems a good chip that could well be capable of very good speeds. I primes for a while at 3Ghz but when the temp gets high on the hot core it fails. This is without a huge boost to Vcore.
 
PinkFloyd said:
Fair enough, was just checking. I get exactly the same issue on my quad, one core is about 3-4 degrees hotter.
Click to expand...

I think it is quite common, in fact usual, to see a difference but 10 degrees is a huge amount and I believe severely limting my overclock and would love to do something about it if possible.

Any ideas?
 
well, back in the day we were hanging HUGE AX7's and HEAVY SLK800's off of the much much smaller AXP die, and those didnt even use the mounting holes (the AX7 definately only clipped onto the socket.

http://slo-tech.com/slike/293.jpg <who remembers this bad-boy combo? (review http://www.frostytech.com/articleview.cfm?articleID=1274 )

if its a worry - carefully make a shim out of layers of electrical tape. this is what ive had todo to get good contact with my decapped chip and the RBX waterblock (the koolance block used to fit fine without a shim!)

uneven contact could definately hold back an OC. i used to be able to run 2920mhz @ 1.5v @ ~69/~80C load. lapping took it down to 1.475V @ ~68/~75c decapping got ~40/~43c @ 1.425V (mine was only just in contact with a thick layer of goup under the IHS)
now its decapped i can also run at 3100mhz @ 1.65v with ~50/~55c load...
 
Last edited:
Amonlym said:
well, back in the day we were hanging HUGE AX7's and HEAVY SLK800's off of the much much smaller AXP die, and those didnt even use the mounting holes (the AX7 definately only clipped onto the socket.

http://slo-tech.com/slike/293.jpg <who remembers this bad-boy combo? (review http://www.frostytech.com/articleview.cfm?articleID=1274 )

if its a worry - carefully make a shim out of layers of electrical tape. this is what ive had todo to get good contact with my decapped chip and the RBX waterblock (the koolance block used to fit fine without a shim!)

uneven contact could definately hold back an OC. i used to be able to run 2920mhz @ 1.5v @ ~69/~80C load. lapping took it down to 1.475V @ ~68/~75c decapping got ~40/~43c @ 1.425V (mine was only just in contact with a thick layer of goup under the IHS)
now its decapped i can also run at 3100mhz @ 1.65v with ~50/~55c load...
Click to expand...

This sounds a bit complicated. I don't even know what a him is. The Tuniq screws down onto the chip and until you have screwed it down does not simply sit on the chip, so I am less worried about crushing the chip and more worried about a poor contact.

If I do go ahead do I simply put a nit of compound directly onto the core and slowly lower the Tuniq in the normal way.
 
You must log in or register to reply here.
Back
Top Bottom