liquid metal tim on copper plated aluminum.

[Bill Dimwit]

Has anyone tired liquid metal tim on a copper plated aluminum heat sink?
Dose it work or will it end up like none plated aluminum?

Thanks.

 

[trents]

As long as you don't have scratches the breach the integrity of the copper plating I think it would be okay. How thick is the copper plating? Is it just a coating or is it an actual plate?

 

[Bill Dimwit]

As long as you don't have scratches the breach the integrity of the copper plating I think it would be okay. How thick is the copper plating? Is it just a coating or is it an actual plate?

I'm not sure how thick it is. But I agree with you thoughts. Its a zalman cnps2x. It surface is far from flat so I'll lap it and opt not to use liquid metal to be safe.
But the question still remain, has anyone given it a shot before?

 

[trents]

Unless you lap the CPU lid too you will not gain anything. The cooler bases these days are slightly dished on purpose because the CPU lids are also. They want most of the stress to be placed on the edges of the heatsink where the vertical walls are. This is to keep from damaging the die which is very small now and underneath the center of lid. It's not like it was in the old days when lots of people were lapping. Gains are minimal to nil. If you still want to lap, I would check temps before and after to discern if you accomplished anything.

 

[Bill Dimwit]

Unless you lap the CPU lid too you will not gain anything. The cooler bases these days are slightly dished on purpose because the CPU lids are also. They want most of the stress to be placed on the edges of the heatsink where the vertical walls are. This is to keep from damaging the die which is very small now and underneath the center of lid. It's not like it was in the old days when lots of people were lapping. Gains are minimal to nil. If you still want to lap, I would check temps before and after to discern if you accomplished anything.

I just finished up lapping it to 2k grit. I did a before and after.
Before with a full load with F@H the CPU max load temps were 63-65c. After lapping the max load temps with F@H is 48-52c. Core 3/4 run much cooler. The surface of the heat sink was so bad it didn't even sit flat on the table. It was sloped to one side too. Both are something I seen in reviews of the heatsink.

I may lap it a second time and take more time to do it right when (if) I get a 7700.



Edit:
May bad, forgot to plug in the system intake fan when I tested after the lap.
Post lap temps are 46-50c. Very impressive. I don't think Id ever seen that much of an improvement before before. And its not a bad apply of tim. I tested with the same Tim before hand with 2 mounts to be safe.

 

[EarthDog]

I've never seen lapping take 20C off temps...I wonder how many curious results you are going to get Bill.

Also, you were not clear if you lapped only the heatsink or both the sink and IHS. Seems like just the heatsink?

Unless you lap the CPU lid too you will not gain anything. The cooler bases these days are slightly dished on purpose because the CPU lids are also. They want most of the stress to be placed on the edges of the heatsink where the vertical walls are. This is to keep from damaging the die which is very small now and underneath the center of lid. It's not like it was in the old days when lots of people were lapping. Gains are minimal to nil.

Source? Some heatsink bases were designed to be convex and CPU dies were still in the middle of the IHS (that hasn't changed). I don't recall hearing a passing mention of that changing. Dies didn't move and some heatsink bases are still slightly convex. Also consider, there is a business now for better IHSes. With that in mind, are the gains nil to minimum?

Also, dished, I am assuming you mean concave. If both the heatsink base and CPU IHS are concave, that would make for poor contact between it and the base.

 

[Johan45]

I think your testing method is flawed from the start by using F@H as your guage. The WUs change and so does the workload so you can't do an apples to apples comparison like that.
I have been down the lapping road before and unless that cooler's plate had the shape of a table spoon there's no way it shaved 20° off

 

[trents]

Yep, I meant to say the IHS is concave. That becomes obvious when you begin to lap them as the copper underneath the nickel plating begins to show through first around the edges. I think this was not so some years ago, around the time the TRUE cooler came out and so many people were lapping their bases. The base of the TRUE was obviously convex out of the factory. You could set it on a table top and see that. If the base of the cooler were convex to mate with the concave top of the IHS it should create the same contact effect as lapping both to make them flat. Why would it help? That's why I say I don't think the IHS some years ago was concave like it is today and I think I remember reading somewhere Intel began doing that to protect the smaller die and. It creates a small gap at the center so the edges take the pressure. No, I don't have a source to cite but I remember reading it about the time Skylake came out if my memory serves me correctly (and it doesn't always do that) in conjunction with the thinner PCB used beginning with that generation.

 

[Bill Dimwit]

I think your testing method is flawed from the start by using F@H as your guage. The WUs change and so does the workload so you can't do an apples to apples comparison like that.
I have been down the lapping road before and unless that cooler's plate had the shape of a table spoon there's no way it shaved 20° off

I tested using the same WU both before and after, both for the first 5% on the WU. I also tested with the CPU burner option in furmark and the temps were within 1-2c of the folding. I just did not say anything about the furmark CPU burner test as its not as real world as F@H and I never been a fan of synthetic tests of any sort. But I ran it to have a second data source to confirm my findings.

The surface was pretty bad. The heat pipe was higher then the base plate by about 1/4 mm and more oval in shape, the base plate was sloping to one side. When you sat it on a flat surface you could see a gap on one side of the base plate that was almost 1mm thick. I can't say I ever seen a worse heat sink surface in my life. It was so bad I just about returned it, but after seeing some reviews saying the same things I just lapped it. A good lap of both a CPU and heat sink will drop around 5c, the drops I saw here are not at all typical.

Yep, I meant to say the IHS is concave. That becomes obvious when you begin to lap them as the copper underneath the nickel plating begins to show through first around the edges. I think this was not so some years ago, around the time the TRUE cooler came out and so many people were lapping their bases. The base of the TRUE was obviously convex out of the factory. You could set it on a table top and see that. If the base of the cooler were convex to mate with the concave top of the IHS it should create the same contact effect as lapping both to make them flat. Why would it help? That's why I say I don't think the IHS some years ago was concave like it is today and I think I remember reading somewhere Intel began doing that to protect the smaller die and. It creates a small gap at the center so the edges take the pressure. No, I don't have a source to cite but I remember reading it about the time Skylake came out if my memory serves me correctly (and it doesn't always do that) in conjunction with the thinner PCB used beginning with that generation.

Everything you say sounds about right from my experience lapping heat sinks and CPU's.

I've never seen lapping take 20C off temps...I wonder how many curious results you are going to get Bill.

Also, you were not clear if you lapped only the heatsink or both the sink and IHS. Seems like just the heatsink?

Just the heatsink. I plain to sell the CPU when I upgrade to a 7700 and don't think a lapped 6700t would sell to well.
One side of the heatsink was about 1/2 to 1mm higher and the the heat pipe was oval in shape and about 1/4mm higher then the rest of the heatsink. It was just a plain mess. I really wish I had a good camera to take before and after photos. But my old phone has a crummy camera.


EDIT:
I know its ms paint, but it should give you an idea as how bad it was. The black is the heat sink and the red is a flat surface. The black spot in the center is the heat pipe.
View attachment 199167

 

[EarthDog]

I just did not say anything about the furmark CPU burner test as its not as real world as F@H and I never been a fan of synthetic tests of any sort. But I ran it to have a second data source to confirm my findings.

In cases like these, its best to run something synthetics and not variable like F@H. I see you did your best to minimize it, but why play games with F@H and not use a simple straight up test.

As far as the heatsinks go, trents, I have't really seen much of that. I recall concave ihs, and convex heatsinks, but I do not recall things changing due to crushed dies.

 

[Bill Dimwit]

In cases like these, its best to run something synthetics and not variable like F@H. I see you did your best to minimize it, but why play games with F@H and not use a simple straight up test.

I really was not playing games with F@H. I got a new WU, folded it to 5%, shut the system off before it saved the WU progress and retested after lapping/applying new tim.
The only thing I do with the system that will load it up is F@H so I wanted to test with it for the most part.
I used the CPU burner in Furmark is confirm the findings. It was within 1-2c of the max folding temps.
So witch ever way you look at it the temps were about 15c lower after the lap with both.

 

[trents]

Bill, if the heat pipe was holding the base off of the IHS then I can easily see a 20c temp improvement. Otherwise, in my experience lapping gives more like a 2-6c improvement and that was in the old days. I haven't tried it in recent years. When I have thermal issues I just delid. That's a much more efficient solution IMO. But like I said, if the heat pipe were preventing good contact then that's another story.

 

[Bill Dimwit]

Bill, if the heat pipe was holding the base off of the IHS then I can easily see a 20c temp improvement. Otherwise, in my experience lapping gives more like a 2-6c improvement and that was in the old days. I haven't tried it in recent years. When I have thermal issues I just delid. That's a much more efficient solution IMO. But like I said, if the heat pipe were preventing good contact then that's another story.

I agree. It really was bad.
The next biggest drop I seen first hand was 8c. 2-6c is about normal when lapping both a CPUand heatsink. Drops are much lower when lapping a die, not that you want to do that with anything remotely modern.

I have lapped just about every heatsink I bought. I buy most of my stuff second hand, don't care much about warranty's and have stacks of sand paper. So why not lap them all? It as a free 1-3c drop for me. More often then not the cheaper older heat sink will see the biggest improvements.
I haven't lapped to many newer heatsinks so I cant say if the 2-6c drop still holds true. The newest CPU I have lapped was a 1366 6 core xeon. And I tend to only lap newer heat sinks if there cheaper or really need it.

Once I upgrade my CPU I'll try lapping and delidding.