To lap or not to Lap?

[-=Vect0r=-]

I can understand why lapping a slightly convex/concave CPU can improve temps as the flatness increases the surface contact with the heat sink thus increases the heat transfer.

Presumably therefore, it follows that if you lap the chip you need to lap the heatsink in order to maximise the surface contact.

Currently running 1.5V through chip in sig and load is roughly 59C max.

Would like to achieve 3.6GHz on air, I can currently do it but temps and vCore go through the roof and are unacceptable.

Would it be worth lapping my chip/TT to achieve a higher clock or would it solely improve temps at current clock?

 

[IAmATeaf]

Personally I'd think that you'd not get much of an improvement as a full load temp of 59 is really good for a vCore of 1.5v. On the flip side you might shave a degree or 2 and lapping the CPU and HSF certainly makes you feel good

 

[platinum87]

youd prob get a 5 degree temp drop, i got 5-8 degree drop so im happy, now at 59 degrees load, which i think is good against the quad core database there is only 2 guys at same clock speed who are equal to or cooler than mine.. (not including water)

 

[Makunouchi]

Isnt it also to do with removing the slightly less conductive material ontop to get down to the copper stuff underneith?

 

[-=Vect0r=-]

youd prob get a 5 degree temp drop, i got 5-8 degree drop so im happy, now at 59 degrees load, which i think is good against the quad core database there is only 2 guys at same clock speed who are equal to or cooler than mine.. (not including water)

That's impressive! Were you able to hit 3.6GHz before you lapped? Were you able to lower the vcore/temps after the lapping?

 

[platinum87]

That's impressive! Were you able to hit 3.6GHz before you lapped? Were you able to lower the vcore/temps after the lapping?

i have not modified anything since i have lapped.

i bascially put the vcore to 1.35 / 3.6ghz, it didnt boot to windows, so i jumped to 1.4050 and its stable under prime. (i do my oc in bios, therefore i dont like wasting time going up using smallest increments ! ) eventually i will check my max stable when i have time, until them just leave on current settings

 

[-=Vect0r=-]

Cheers, might leave off for a while. Sounds like the primary aim of lapping is to reduce temps rather than voltage.

I was hoping I could also reduce the vcore and thus achieve higher clock with a lower vcore.

As temps aren't the limiting factor me I think I will leave the lapping game to a future upgrade if required.

 

[-=Vect0r=-]

Couldn't resist having a go.... Here's the result and the obligatory gratuitious reflection shot

My temps have not really seen a demonstrable improvement. I think I have gained 1/2 deg at most, although I should have taken a CoreTemp temp log before & after to make a direct comparison. Hindsight is a wonderful thing!

I also changed the TT orientation to front to rear rather than out through the top. I also strapped a Yate Loon to it to further reduce temps