Thermalright Ultra-120 Extreme: Lapping Part1

DirkVoodoo

EDIT: I realise I should point out that there is in fact no lapping in part 1, its just a boring introduction...like "the phantom menace" or "the DaVinci Code (movie, not book)".

Right, Just thought I would let you all know a few caveats about the Thermalright Ultra 120 Extreme.

I was initially sold on the stellar performance achieved by the anandtech crowd in their review. Hoping to see the same performance on an E6600, I took the plunge and bought the cooler with a 1200rpm Scythe Sflex fan.

Initial results were disappointing. I was getting 33/35C at idle and into the high 40s on stock settings, in stark contrast to the overclocked X6800 from the review. Granted, there is some error for setup/additional cooling, but I never had odd temperatures with the stock intel HSF. I applied AS5, reseated it, tried with the thermalright supplied TIM (managed to spray most of my load over the heatspreader, nothing a well practiced mop-up can't fix ), AS5 again. I tried with AS and thermalright application instructions, no joy.

So, consulting google led to the conclusion that these coolers are quite notorious for having extremely convex bases. There have been some successful attempts at lapping, so it is possible, but the majority of users experiencing this problem (still not sure what % of overall extreme owners this is) are afraid to mess with an expensive cooler. One chap was even afraid that lapping would completely remove the base and wear through the heat pipes, best of luck to him!

So, I would just urge my fellow boardsies to take into consideration that if you fork out for high end air cooling (on dual core at least, don't know about its performance on quad) and expect high end results, you may have to lap this sucker.

I will post again in the next few days, with something similar to L31mr0d's lapping guide, except sans photos and probably with some borked hardware. Stay tuned for the not-so-thrilling conclusion.


FURTHER EDIT: It should also be noted that the new 775 retention clip is easier to install but inferior to the regular clip. It affords the heatsink a little too much freedom to move around, and has been suggested that in tower cases this causes an uneven pressure distribution on the heatspreader thanks to one Sir. I. Newton. Some remedies bandied about include pennies and washers to increase clamping force.

pgleeson

Hey man, like you I also got the Thermalright Ultra 120 extreme, AS5, E6600 and the 1600 version of the same fan u got....heh...coincidence eh?

Anyway, I'm getting 21C idle and 28-30 load. So I think I must have gotten one of the good ones. Although, I am concerned with the looseness of it. Has it been long since you bought it?

DirkVoodoo

No, I got it last week. Have you checked your surface for curvature? What method did you use in applying the TIM? What case are you using, etc. Results seem to vary greatly for this HSF, unlike the Tuniq tower which seems fairly consistent.

pgleeson

There's not much curvature on my one, it looks smooth enough. I coughed up the big cash for the case and got a Stacker 830 Evo so that should certainly be taking a few degrees off the temps. I didn't lap it, but I was tempted - I decided to just try it the way it was first and if I had gotten crap temps then I would have. I applied the TIM according to the instructions on the arctic silver site.

L31mr0d

DirkVoodoo wrote:

sans photos

what? what use is a guide without pictures, pictures being a thousand words and all that. Take some pictures dammit, I want to see that P160

Seriously though, I didn't even know about the P160 until I seen your sig, looks like that will be my next case. Does it have insulated side panels like the P180? Also it might help knowing what your temps where with your last heatsink.

DirkVoodoo wrote:

No, I got it last week. Have you checked your surface for curvature? What method did you use in applying the TIM? What case are you using, etc. Results seem to vary greatly for this HSF, unlike the Tuniq tower which seems fairly consistent.

Have you checked YOUR surface for curvature? Before you just go ahead and lap it it might be an idea to make sure that it is actually curved. Also have you thought of the possibility the IHS on the e6600 might be curved also?

I'd compare both surfaces to something that you know to be flat (i.e the edge of a machinists ruler) and see if they actually are curved. You can then use this as your gauge when lapping. Also remember when you get it flat, you can stop lapping there, only go for shiny if you have the patience as it will take 3 times longer to get it shiny than to get it flat for usually only a marginal drop in temps. Then main drop in temps will come from getting it flat.

Also what method and grits are you using to lap?

DirkVoodoo

Demands, demands...

I'm trying it out first on my Zalman 9500 AM2 cooler. Following your methodology. Using a selection of grits up to 1200. The base of this already had a very nice surface finish, but a quick test with a flat edge showed it be concave.

Don't worry, I will be checking my Ultra-120 for curvature, as well as the IHS. I'm also going to try remounting it first with the penny trick, to see if increased clamping force is my problem.

And the P160 has no insulation, some resonance noise does occur, but its a very cool case, in the thermal sense.

EDIT: Talking to some of my colleagues, I'm considering making use of the surface engineering lab here in Belfield to see if I can improve on the current method.