Temps for FX-8150

[ccokeman]

I'm just going to go with the FX series having bugged temperature sensors.

 

Weirdest thing I have seen in a while though but three different board gave me three different readings. Go AMD!

[Waco]

 

Weirdest thing I have seen in a while though but three different board gave me three different readings. Go AMD!

At least it's not as bad as my lying Gigabyte board in terms of vcore. I had mine set at 1.39 volts...and lo and behold I check my vcore randomly with CPU-Z and it's reporting 1.56 volts. No wonder I thought my chip was golden...

 

I had to drop down to 4.5 GHz and "1.33 vcore" which gives me 1.43 volts under load in Windows. My chip is pretty crappy after all...it's stock vcore is 1.36 volts. I'm pretty sure that's one of the highest I've ever seen...

[Drdeath]

Overclocking 101:

 

If both matting surfaces aren't perfectly flat, your not going to get optimal results from your cooling solution.

 

You should always mount and then remove your cooler to check for TIM coverage before even starting up your computer for the first time.

 

Lap or return the cooler for a new one for best results.

 

 

??? Funny. There is no completely flat surface and the lid on CPU's have a slight bow. Lapping gets ya 2-5(More like 2 under load) degrees and with Sandybridge not worth it.

[Drdeath]

I think your right about it "flattening out" I think it is probably (as i said deforming to some point ) ...like halving the beginning radius. another possibility is that the heatspreader is being pushed in a bit as well. Although that theory seems unlikely merely because i wouldn't think a company that manufactures heat-sinks would take the risk of changing the shape of the CPU. Although they keep building heatsinks that exceed the motherboards cantilevered weight recommendations.

 

In the case of the Silver Arrow, after mounting and then removing to inspect the spread of the TIM, there is roughly an area of 3mm of almost no compound right down the middle of the convex ridge. I would like to see the research as well. It seems like a simple enough concept. The more pressure between two metal surfaces, the better the transfer.At some point though, they are trading surface area for better contact. Whatever the combination of events happening when you clamp it down, it sure seems to work.

 

 

Too much pressure is not good for Intel sockets. 70Lbs of torque is optimal but who uses a torque wrench?

[red1776]

Too much pressure is not good for Intel sockets. 70Lbs of torque is optimal but who uses a torque wrench?

 

 

In the case of the Silver Arrow , the intel mounting bracket uses a pressure wheel (dial) that goes from 40 to a max of 70lbs according to Thermalright.

 

 

 

A lot of the reviews say things like "we tightened it all the way .....and observed no ill effects" ..or the like, but to me it feels like it's going to pull the socket through the board. I have stopped somewhere around 50-60lbs ( i am estimating) and have not realized any discernible change in performance.

 

I have read that people in the metallurgy field continue to say that around 70 psi is optimal like DR has said. This article regarding mounting pressure and lapping is interesting, although it is going on 4 years old and I wonder how the prevalent use of heat-pipes (running through or direct contact) has changed the lapping for flatness vs high spot) aspect. It would seem that it has to some degree with the use of convex mounting bases.

 

Interesting read nonetheless.

 

http://www.overclockers.com/heatsink-mounting-pressure-vs-performance/