I'm thinking of the ek supreme. It is very intentionally convex, to ensure that a lot of pressure is applied to the centre of the die. Presumably EK know what they're doing, but I'm still tempted to run a mill over it/lap it. What do you think?
Would you lap an intentionally bowed waterblock?
- Thread starter JonJ678
- Start date
What cpu are you putting it on, ive mainly used intel c2d/c2q as the last few years, and lapping certainly helps on air cooling (lapped TRUE black) ive heard thermalright press release's saying that they make their heatsinks concave to cater for the cpu's ihs, my previous q6600 was hiting 81c on 2 cores at 3.8ghz prior to lapping, lapped the cpu and the TRUE black to 2000 grit and got it down to 68-68-62-62.
Very seriously considering watercooling myself, im in need of a new case as my current akasa eclipse is getting on a bit, thinking of a silverstne tj07 with a 120.3 rad and a basic cpu loop to start with.
Very seriously considering watercooling myself, im in need of a new case as my current akasa eclipse is getting on a bit, thinking of a silverstne tj07 with a 120.3 rad and a basic cpu loop to start with.
I'm using an akasa omega with an eclipse side panel, lovely case 
It would be going on a q9550 ideally. I've been fighting to rma that for over 5 weeks so far though, so at present it's on a 7750be which is quite cold enough anyway. If the retailer decides to flatly refuse to deal with me, and small claims gets me nowhere, it'll end up on an i7 in about 6 months time.
I've got a lapped true sat on my desk at the moment, lapping it helped significantly. 2000 grit must have taken you a while
The problem is, I can understand the reasoning to making it convex. The mounting pressure deforms the base, flattening it against the processor. The issue is that I don't believe this happens. Looking at the block now, it's about 6mm thick copper with channels cut out. I just cant see the pressure from mounting it deforming it even slightly, and it's really quite bowed. Which would mean lapping processor and block would improve things significantly. But if I'm wrong, I might write off a 40 quid cpu block, which is a bad thing for a student to do.
It would be going on a q9550 ideally. I've been fighting to rma that for over 5 weeks so far though, so at present it's on a 7750be which is quite cold enough anyway. If the retailer decides to flatly refuse to deal with me, and small claims gets me nowhere, it'll end up on an i7 in about 6 months time.
I've got a lapped true sat on my desk at the moment, lapping it helped significantly. 2000 grit must have taken you a while
The problem is, I can understand the reasoning to making it convex. The mounting pressure deforms the base, flattening it against the processor. The issue is that I don't believe this happens. Looking at the block now, it's about 6mm thick copper with channels cut out. I just cant see the pressure from mounting it deforming it even slightly, and it's really quite bowed. Which would mean lapping processor and block would improve things significantly. But if I'm wrong, I might write off a 40 quid cpu block, which is a bad thing for a student to do.
Yep jon the akasa is a very good case, eclipse 62 with window here, globalwin eclipse on my second rig (gonna get it powdercoated in white) my previous q6600 was an average 1,2750 vid chip, took it to 3.8ghz after lapping it and my TRUE black, my current q9550 is prime stable at 3.6ghz on very little volts (not blend stable unfortunately) When i lapped my old q6600 it didnt take that long tbh mate, im a painter and decorator (15 years in that game) 2 hours or so to do the cpu but the TRUE black took about 4 hours, took a while getting through the nickel, they looked so nice when they were done it seemed a shame to cover them up
I can well believe the TRUE took four hours off you. Its not so convenient to lap either, stripped skin off my knuckles doing so. It took me about an hour, but I used a lapping table
My q9550 would pass small ffts for hours yet fail blend almost immediately when overclocked, only thing that helped seemed to be more vcore. Strange. Led to a limit around 3.7 when I wouldn't go over 1.3625.
No other thoughts from the forum on lapping waterblocks? It helps lots with air coolers, but seems to be a topic of debate with water
My q9550 would pass small ffts for hours yet fail blend almost immediately when overclocked, only thing that helped seemed to be more vcore. Strange. Led to a limit around 3.7 when I wouldn't go over 1.3625.
No other thoughts from the forum on lapping waterblocks? It helps lots with air coolers, but seems to be a topic of debate with water
Most coolers that have an intentional bow are CONVEX not concave. A CPU's IHS is normally concave and hence top coolers try to counter that by making there coolers convex.
Lapping a convex cooler will be extremely difficult and I would only ever consider it if you also intend to lap the CPU.
Lapping a convex cooler will be extremely difficult and I would only ever consider it if you also intend to lap the CPU.
Cheers w3bbo, well seen i always get confused with convex/concave (few beer in me doesnt help either) and good point about lapping both, i done my old q6600 and TRUE black and it made quite a difference, allowed me to get to 3.8ghz on reasonable voltage with temps below 70c, very seriously considering doing my current q9550 to try and even the temps a bit.
Hehe, np m8. Lapping can certainly help some people but others won't notice a difference. It depends how bad the Concave/Convex is in the first place and also what the finish is like on the CPU.
I religeously lap all of my CPU's/heat sinks, more often than not there is a big benefit but other times less so.
Get that 9550 lapped! I expect a full report including pics and before/after temp comparisons
I religeously lap all of my CPU's/heat sinks, more often than not there is a big benefit but other times less so.
Get that 9550 lapped! I expect a full report including pics and before/after temp comparisons
Ah, so it's more an attempt to get the two curved surfaces to mate nicely than a deformation under pressure thing? That makes sense.
Lapping the convex waterblock will be easy enough in practice, as Id run a milling bit over the surface first. My current processor isn't lapped, but its predecessor was and I'll probably lap the next one I get. It appears i7 runs exceptionally hot, so if lapping both will improve the state of affairs I'm inclined to do so.
Any guesses at whether lapping both would make a significant difference w3bbo?
Cheers both of you
Lapping the convex waterblock will be easy enough in practice, as Id run a milling bit over the surface first. My current processor isn't lapped, but its predecessor was and I'll probably lap the next one I get. It appears i7 runs exceptionally hot, so if lapping both will improve the state of affairs I'm inclined to do so.
Any guesses at whether lapping both would make a significant difference w3bbo?
Cheers both of you
..............
Any guesses at whether lapping both would make a significant difference w3bbo?
Cheers both of you
Ofc it would.
in an ideal world both surfaces would be flat, perfect surfaces with no pits or high/low points.
Obv' this is not possible, and that's what the use of the thermal compound is, to make up the differences.
The smaller the differences are, the lower thermal restiance.
That is the bottleneck(thermal barrier) of the energy(heat) transfering from the IHS into the waterblock.
Before i installed the q9550 i tested it with a brand new stanley blade to see how level the ihs was, looked fairly good to the naked eye.Hehe, np m8. Lapping can certainly help some people but others won't notice a difference. It depends how bad the Concave/Convex is in the first place and also what the finish is like on the CPU.
I religeously lap all of my CPU's/heat sinks, more often than not there is a big benefit but other times less so.
Get that 9550 lapped! I expect a full report including pics and before/after temp comparisons
Current idles.
Prime small FFT.
Bit of difference between cores, but not as bad as my q6600 was prior to lapping, im also toying with the idea of bodging some sort of bracket for my TRUE black, due to its weight i think its bending downwards slightly and not getting a good contact, ive took the side panel of the case and pushed it up slightly and seen the temps even up a bit.