Thermal paste is just as evil as thermodynamics. It's not in the Bible, so it's the work of the Devil! You and all of your voodoo thermodynamics.

I'm already seeing HUGE differences between unlapped and lapped. Even with the first start up, and no time for things to settle in, I'm seeing a peak temp of about 51C under load, running a heavy Prime95 cycle, listening to music and 2 Chrome browsers open. And my case is wide open, and one of my fans isn't installed. Once I close up the case I bet it drops a few more degrees, since the flow is all messed up right now. I can't recall the exact number, but I used to have to watch my temps very closely when running Prime95, because the temp would creep up close to the max recommended temp. Not even close now. In a few days the temp will drop a few more degrees once the paste cures, or goes through a few cycles, or whatever voodoo magic **** it does.

My FX-8350 was far worse than I ever thought it would be. I had been into copper in the center and edges for a long time, and I could still read the laser etched AMD logo in a low spot. And I must have some older version of the Hyper 212. I guess the non EVO version, since my heat pipes have large gaps between them. Not ideal, especially since there are 4 pipes, and the two center ones split the center running die in the processor. So the hottest spot is sort of missed, and only the two center pipes catch the edges of the hot spot. Might have to look into an EVO version.

Mine:



EVO:

 

Me thinks you were using too much paste before.

 

Do you have the 212 that has the aluminum heat sink around the heat pipes with the 2mm gaps between the heat pipes? If so thats the 212 plus, the 212 evo is the one i have.

 

Yeah, I have the Plus. I think it's an earlier version. Doesn't look nearly as efficient.

 

I wasn't using much. A very thin bead down the center, flattened out with a credit card edge, and excess close to the edges wiped away. Maybe a bit more than ideal, but not what I would call a lot of paste. I used even less this time though, just barely enough to see when spread out. So it may be partly due to the paste amount.

 

Dont do the spread method. Thats old school and has been proven to be full of fail.

Put a single blob about the size of a pea in the middle and use the heatsink to spread it. It does a much better job of keeping it even. The amount you use is critical though. If its squeezing out the sides youre using too much.


Also, here is some cool benchamrking on different compounds

http://www.tomshardware.com/reviews/...k,3616-18.html


Dont sweat the heat-sink design so much. That direct heat-pipe contact deal is just a marketing gimmick. A layer of aluminum between the pipes and the CPU would have a negligible effect. Thats how the used to build them and thats now heat pipes still are used outside of the enthusiast aftermarket.

 

^ This.

It's how the OEMs do it. (Or you can run a thin line across the major axis of the CPU, which is to say, roughly aligned with the orientation of the die.)

 

That's what I did. A rectangle, thin bar centered over the die. Almost like an OEM pad of compound, just rectangular not square, and much much thinner. But by the time I got done ******* with the aggravating mounting system for the heat sink, the **** was probably spread all over the processor anyway. It'll be coming back off in a few days anyway, when my new fans arrive. So I'll clean it up and do it again and hopefully not fumble around with the mounting system too much. The screws weren't aligning correctly for some reason. I guess I bumped one of them and moved it to a different slot by mistake.

 

Hell. I think I'm just going to let my ASUS 990FX R2.0 handle the overclock. Set in performance mode, it throttles between 1550mhz at idle, and when running any programs it jumps to ~4375mhz. It also plays with the FSB and does a nice job of kicking the memory in the *** and gets it overclocked pretty well, with tight timing, good stable voltages. I've been running it Prime95 for max temp and it hasn't gone above 52C in 45 minutes to an hour. That's good enough for me.

Yep. I'll hang with this for now. If I find myself needing more, or bored, I'll go back to manually tweaking. But this is doing a surprisingly good job. It has underclocked my memory by 200mhz, but I can live with that very negligible issue. I would like to see the temp drop to below 50C, which I think it will after a week or so. Plus it's a little warmer in the house than usual.

 

as I feared, these new weak *** fans aren't up to the task like the old monsters were. Somehow the idle temp is down 2 degrees, averaging around 29C, but the under load temp has gone up 9-10 degrees to around 60C. At least it's still stable, so I'm not too concerned. It'll last until my next build, which should be due in the next year or two, going by my usual PC build schedule.

 

I posted the article 2 pages ago but i take it u didnt see it.

Thermal Paste Comparison, Part One: Applying Grease And More - Everything You Wanted To Know About Cooling A CPU

 

I guess I missed it. Good read. I just wasn't thinking on such a microscopic level. This makes me think even more about my old Hyper 212 Plus not being a very good cooler, at least not on the level of the EVO. All that contacts the processor is the heat pipes themselves. The metal pieces of the plate between the pipes are raised a good .010" and contribute nothing to cooling, or if anything trap hot air in. Or I suppose with enough compound it would fill the gaps, but I don't want to use so much. The EVO is MUCH better in that way, since the heat pipes cover the full surface with no gaps. Probably completely worth the upgrade.

I wish I could find just the mounting plates for the fan of the EVO, so I could mount a push/pull fan setup. Thought I guess zip ties would work the same.

 

Actually there were some tests i saw when i was looking at my setup. The push+pull setup was only 2 degrees better than the push setup.. a pull setup on the other hand was 1 degree better than the push setup.

Run it on a pull setup and leave it like that. IMO the extra fan noise for the single 1 degree improvement isnt worth it.

 

It's always been a pull, but any improvement for a little noise is worth it for me, since the noise is absolutely zero bother to me. I sleep with white noise, and my computer used to be my white noise, until I swapped the fans back to less powerful ones that I could barely hear. I had to go buy a new box fan. My old setup was by most peoples standards REALLY loud and unacceptable. At one point there were 7 fans, 4 of which were 3000rpm, 133cfm monsters and one small 80mm at 3000rpm that gave off a pretty high pitch hum.

Anyway, going to order an EVO, and another matching fan for a push/pull. I bet I drop at least 4 or 5C in the swap over.

Also, what do you think about blocking off the sides in a push/pull configuration? Surely the push fan pushing air out the sides, while the pull fan is trying to pull air back in through the sides causes some turbulence and probably slows movement of air through the fins of the cooler. I'm thinking covering those sides and sort of making an enclosure would help a bit. Could be completely wrong though.

 

Could be worth trying. I havent wasted any time on it since i'm satisfied with the performance now. 51 degrees at 4.5 gigs is fine to me considering before with the old stock cooler it ran at 62 degrees at only 4.1 gigs.

 

With these shitty new fans, I'm seeing 62-63C max at 4.4, which is not at all acceptable, considering I was more than 10C cooler with the old fans and no lapping. I wasn't anticipating such a loss of performance with these new fans. I did have it stable at 4.3 and only 52C with these fans, but I wasn't satisfied with the tune. New cooler and fans, and then try again. If I could get 4.5 stable at less than 55C I'll be happy.

Not sure why I always like the hard route. Same with cars. Start with less than ideal performance mods (n/a) and try to squeeze every drop of performance out of it. I could just throw water cooling on it (turbo) and be done, and get better numbers for not much more money than I've already spent on cooler and fans.

 

Dude, maybe you need thermal compound, lol.

 

? I have thermal compound. AS5. It's just these new "silent" fans. I fart a stronger breeze than these things. I was expecting a little more performance out of them than when I got. I have another use for them, so it wasn't a complete waste of $16.

 

Oh I thought you were not going to use it because of your lapping.

 

No. I was thinking I was going to try it without just to see what would happen, but after the discussion I decided it wasn't even worth the trouble of trying and having to take it back apart to put some compound on it anyway.