Not complaining but my baby has never got down to the 60's before.
Two days ago I took her back apart and did a paint brush, Q-Tip, alcohol, and canned air kind of clean up. When I say took her apart, I mean everything was taken out and taken apart to be cleaned. Well I didn't take apart the hard drives. But I did figure out how to get the shroud off the video card. Good thing. It was pretty trashed out.
During the day, when she's getting used, she stays in the mid 80's jetting up to the low 90's. She gets a lot warmer as seen in the pic when she's real busy like running a virus scan.
What's odd, and I'm sure I brought this up before, is the cores heat up unevenly. So to get the 411 on what's up with that, Google and I got busy using the search term "Intel i5 cores heating unevenly". The results made for several hours of interesting reading.
It appears it's a quirk of the heat spreader Intel puts over the chip itself. The chip as far as I can determine, is a rectangle running north/south when sitting in it's socket. That's with the yellow alignment arrow down and to the left. It appears from others jacking the spreader off that Intel uses a thermal paste over the chip then bonds the heat spreader cap to the circuit board with a very strong adhesive. Some of the braver souls that got the heat spreader off resorted to some seriously off the wall methods to break the bond.
A few reported doing a through cleaning of the factory paste and replacing it with some familiar pastes like AS5 to a liquid thermal interface I had never heard of before. Then bonding the heat spreader back on. Here's the interesting part.
AS5 had no better results than the factory paste. Minor drops that didn't justify all the trouble they went through to get it apart. The liquid only gave a slightly better return in a 10 degree drop. Leaving the spreader off and just installing the heat sink directly to the chip didn't work well at all in most cases. Since the chip is so small that would make sense. If you're using something like the CM Hyper 212 evo, there would not be enough contact surface area. But to me if you were using a heat sink with a solid slug in the center it might work better. It would have to cover the entire chip for it to have a chance to work.
And all of what I read was on air cooling. I didn't see anything on using liquid cooling with the spreader off.
Now my personal observations tell me it's not worth buggering up a $229.00usd CPU chip getting it open. The end results seriously doesn't justify the risk of destroying the chip. If you got a 20 to 30 degree F drop then maybe. But from what I read getting those sort of temp drops seems out of the question.
From here is just what I've seen first hand in building a lot of systems.
Did you ever notice when building a P4 system that the heat spreader on the CPU what flat? The reason I bring this up is because if you look at the heat spreader on the current line of CPU chips, it has a slight dome shape. In my mind this would account for the uneven distribution of the heat off the chip going to the heat spreader. If it was flat with the thermal interface then maybe the heat distribution would be more evenly spread out. Makes me wonder why they changed the spreader configuration.
It would seem to me if you applied enough pressure to flatten out the dome shape eventually the stress would snap the chip.
Let's touch on lapping for a sec. Going back to the surface of the spreader you may have noticed the swirls in the metal. Scratches in the surface actually give you more surface area for the heat to dissipate from. I just can't get my head around polishing those scratches out. When you take them off it reduces the effectual surface area. If anything adding some more scratches would increase the surface area. You'd have to use a thinner thermal interface to get down in them, but to my way of thinking it would be more efficient.
Ok enough of my babbling on. What do you gals and guys think?