How to check flatness of CPUs and coolers - INK and OPTICAL INTERFERENCE methods

[BoggledBeagle]

I see that Arctic published nice illustration of the Intel CPU mounted in the socket with normal mounting mechanism and with mounting frame:

 

[BoggledBeagle]

It seems that Arctic is pushing mounting frames even with their cheap new Freezer 36 cooler:



I think it is a good decision, nobody needs banana CPUs...

 

[lemonadesoda]

Has anybody done a study of HOT lapping vs. COLD lapping.

Afterall, if the CPU is going to be operating at, say, 80C, then it should be lapped at that temperature. Difficult to do. Possibly easier to do at 50C. But there is little point lapping at 20C and thinking that the die and housing will remain 100.00000% flat during thermal expansion.

Then, everyone is lapping their CPU IHS but forgetting to lap the heatsink to the same level of effort and flatness. Again, at 80C.

Personally, I think there are diminishing returns. Trying to save 1-2 C by uberlapping just isnt 'worth it. Only do it with banana CPUs and cheap coolers that have poorly finished CPU contact heat sinks. In the end, what does that 1-2 C save you? Energy? No, CPU uses the same energy. Overclock? I doubt you are going to get any material OC on a regular CPU and PC setup. Time and money better spent buying a better CPU in the first place.

Anybody tried putting gold leaf between the CPU IHS and the cooler heatsink instead of thermal paste? I wonder how that bling would work out.

 

[ThrashZone]

Hi,
Yeah forming a gold shim would be more useful than ink hehe

 

[BoggledBeagle]

Has anybody done a study of HOT lapping vs. COLD lapping. Afterall, if the CPU is going to be operating at, say, 80C, then it should be lapped at that temperature.

Why? I demonstrated here that mounting mechanisms of the socket have HUGE effect on the shape of the IHS (at least with Intel LGA 1700 CPUs). Any changes in shape caused by temperature heat expansion would be thousand times smaller than this.

Anybody tried putting gold leaf between the CPU IHS and the cooler heatsink instead of thermal paste? I wonder how that bling would work out.

Why??? Gold has worse heat conductivity than copper and gold leaf is impossible to work with, it would break into unmanageable heat conductive pieces floating around in the computer!

 

[lemonadesoda]

Why? I demonstrated here that mounting mechanisms of the socket have HUGE effect on the shape of the IHS (at least with Intel LGA 1700 CPUs). Any changes in shape caused by temperature heat expansion would be thousand times smaller than this.

Oh, so you are advocating NO LAPPING AT ALL because the mounting mechanisms bend the CPU/IHS anyway. That's actually a really useful takeaway from this whole thread. Don't waste your time lapping.

Alternatively, breakdown an old mainboard, mount the CPU in the mechanism, lap it at 60C, then transfer to working mainboard, for the ultimate mechanism at temperature match. But I guess that also doesnt help, since each mechanism is mounted fractionally differently, and so you lap to the old mechanism stresses, not to the new.

Why??? Gold has worse heat conductivity than copper and gold leaf is impossible to work with, it would break into unmanageable heat conductive pieces floating around in the computer!

Err, is your IHS copper? Is your heatsink copper? I don't think you will find copper leaf anywhere near as thin as gold leaf, and remember thermal resistance is related to the THICKNESS of the material. If the transition layer is very thin and very soft, just an example, gold leaf, then it's ability to squeeze, bend, and be nanometers thin, is actually something worth considering. And you don't have to deal with thick thermal paste which actually is far thicker than gold leaf. Plus, remember, there are some people here that like LEDs on their PCs. WTF really, each to their own. Just as some like other bling gadgets, if a sheet of gold leaf works just as well, then for some, blingy unique, each to their own.

There is a reason old microprocessors had so much gold in on their heat spreading caps, and there was a whole industry in recovering that gold in recycling.

 

[joemama]

Anybody tried putting gold leaf between the CPU IHS and the cooler heatsink instead of thermal paste? I wonder how that bling would work out.

Here: https://www.reddit.com/r/hardware/comments/939glu Gold leaf seems to work worse than thermal paste

 

[BoggledBeagle]

Here: https://www.reddit.com/r/hardware/comments/939glu Gold leaf seems to work worse than thermal paste

Notice that the person needed to crumple the leaf.

Gold leaf itself is inredibly thin (0.1 micrometers) and useless as such.

 

[BoggledBeagle]

I tested Noctua NH U12A Black version. It appears to be a newest revision - packaging style is different than the style visible in reviews on the internet.



Tha base is very roughly machined. This rough machining will ensure a pretty thick layer of thermal paste even if the overall shapes of the mating surfaces would match perfectly. It is hard to justify.

The overall shape is strongly cylindrical. If the cooler was mounted in conventional vertical position, I think that this shape could match a shape of Intel CPU deformed my the normal mounting mechanism. Flat CPUs of Intel CPUs mounted with the frame probably will have a poor contact.

You can see the thick layer of paste on my CPU mounted with the frame.





I also do not understand, why is the cooler mounted off axis - you can see asymetrical mounting bar above in the ink photo and also off center impring of the CPU on the base..

Cooling wise it seems to be able to cool just few tens of watts more than a small cheap Arctic 34 cooler with 4 heatpipes.

Overall I am strongly disappointed with this cooler.

I think Noctua should use much finer machining, and should sell TWO VERSIONS of the cooler, one with really flat base and the second one matching well deformed Intel CPUs - the current version may be close to this one?

Of course it would be interesting to see how would the performance of the cooler improve if I lapped it, but my expenses connected with computer stuff lately have been very high and I do not want to effectively destroy the cooler - making it unreturnable or difficult to resell. So unfortunatelly it is not happening. There is also the problem with asymetrical mount, which I do not understand, do not like and I am unable to make a symetrical mounting bar myself.

I will return this cooler.

 

[delshay]

I tested Noctua NH U12A Black version. It appears to be a newest revision - packaging style is different than the style visible in reviews on the internet.

View attachment 338684

Tha base is very roughly machined. This rough machining will ensure a pretty thick layer of thermal paste even if the overall shapes of the mating surfaces would match perfectly. It is hard to justify.

The overall shape is strongly cylindrical. If the cooler was mounted in conventional vertical position, I think that this shape could match a shape of Intel CPU deformed my the normal mounting mechanism. Flat CPUs of Intel CPUs mounted with the frame probably will have a poor contact.

You can see the thick layer of paste on my CPU mounted with the frame.

View attachment 338685

View attachment 338689

I think Noctua should use much finer machining, and should sell TWO VERSIONS of the cooler, one with really flat base and the second one matching well deformed Intel CPUs

That's a interesting comment. Will user(s) pay extra for extreme flatness where cutting edge technology is used to measure flatness. I certainly will pay the extra for extreme flatness or custom material.

 

[BoggledBeagle]

I contacted Noctua with questions, I will post answers if I ever get some.

I believe the strategy of making shapes on cooler bases if profoundly wrong. You make a bulged cooler and what is the result, when this bulge meets flat or bulging CPU? The result is very thick layer of paste and poor cooling, you can see the thick layer above.

Just make the coolers flat, or at least make a flat version for people who know they need one!

Will user(s) pay extra for extreme flatness

You do not need extreme flatness, normal flatness will suffice. The normal flatness will not be more expensive, they can make flat surface the same cheaply as they machine shapes on the cooler bases.

I contacted Noctua with questions, I will post answers if I ever get some.

I will post the information I got so far, the positive outcome is, that I learned, that

Producing different coolers with different base convexity was not deemed necessary until now; however, we are currently looking more into this option for future projects.

Other than that the information is not very satisfactory so far.

I will post my messages in Italics.

Hello, I just published findings about your cooler here:

https://www.techpowerup.com/forums/...ical-interference-methods.318042/post-5216826

Please answer my questions:

Why is the base machined so roughly?

Why is the cooler mount asymetrical?

Why is the shape of the base cylindrical?

Why do not you make two versions of the cooler, one with truly flat base?

I may publish answers to these questions.

Thank you

Stepan ***


Greetings from Noctua!

Thank you very much for contacting us.

We would also like to thank you for your questions!

1. The base is not machined roughly, the tiny grooves on the base of the cooler are by design to increase contact surface area. Furthermore, a mirror finish to the base of the cooler was not beneficial to cooling performance in our testing.

2. Due to differences in mounting, the CPU can appear to be asymmetrical; however, this is not the case. Images like the one shared in the forum post you linked can be a result of differences in mounting pressure on the screws of the mounting bracket of the cooler.

3. We would kindly ask you to elaborate on this question.

4. Producing different coolers with different base convexity was not deemed necessary until now; however, we are currently looking more into this option for future projects.

Kind regards,
Lucas ***
Noctua Technical Support


Hallo Lucas,

gruss aus Prag und danke für antworte!

Regarding your answer 2:

You can clearly see in the photo in the thread asymetrical mounting bar (the attachment points are not symetrical to the center of the base), and also the paste imprint of the CPU on the base is clearly off center. The question was, what is the reason for this asymetrical mounting? I hope I do not have to draw lines and measurements on the photos, you can do it yourselves if you have problems seeing it on the pictures.

Regarding my question 3:

You can clearly see in the ink photo, that the shape of the base is not flat, but it could be best described as cylindrical. You can see very unsatisfactory contact with the CPU, where the paste left thick residues, the imprint of the paste closely matches the ink photo, where the thin layer of paste was exactly in the place there there was thin layer of ink and the area was less blue. I stress that I closely followed your installation instructions and all screws were tight.

So my question was, what is your reasoning behind making this particular shape of the base. As I wrote in the thread, this shape might work well will Intel cpus bent in the socket, but it fails badly when the Intel CPU is mounted with the frame, and if would probably also not work well with AMD CPUs which due to the thick heatspreaders probably stay much flatter in the socket.

Stepan


Greetings from Noctua!

Thank you very much for getting back to us.

As to your question 2: Apologies, I misunderstood as per the mounting bracket. The mounting bracket is asymmetrical to ensure 100% RAM compatibility of the cooler. As for the base imprint, this can be due to a number of factors, the most common of which is different pressures applied to the screws on the mounting bracket. To ensure that this does not happen, one would have to use a torque screwdriver and alternate every quarter turn when fastening the screws of the bracket.

And as to question 3: We would kindly invite you to have a look at our FAQ entry on the topic of base convexity.


Kind regards,
Lucas ***
Noctua Technical Support

Hallo Lucas,

what are you talking about? See my picture. The placement of the cooler on the CPU is the result of the cooler mounting hardware, not improperly turning screws when mounting the cooler!



The resulting force applied by the cooler should be in the middle of the CPU IHS! You have special AMD mounting brackets for the case you want a different cooler placement on the IHS. If you explain that the cooler is offset because of the RAM compatibility, something seriously wrong happened in the design phase of the cooler.

The picture in the link is so ein Quatsch!



See my picture instead! THE NOCTUA OWL IS CRYING !!! Or is it sweat because it overheated? You decide.



Stepan

I hope Lucas will not take my joke with the owl as an ending of our conversation, because there are questions still unanswered.

 

[Shrek]

Metal plates will generally have residual stress from their manufacture, especially if they were extruded aluminum. This will ease with time, but distortion will result. So if you really want flat, the metal should be stress releaved first before machining.

It's just not worth the trouble.

 

[BoggledBeagle]

Metal plates will generally have residual stress from their manufacture....

I believe that they first machine the copper base and then they solder heatpipes on it. This heat stress may cause some unwanted warping.

They state they manufacture bulging cooler bases intentionally, see the FAQ screenshot.


NEW POST: (this adding to old posts is really annoying)

I tested DeepCool Assassin IV cooler.






The base has a truly flat spot in the middle. It is also very finely machined. You can see it is flat even with the black ink. Then it slopes to the sides. This arrangement will work acceptably (or well) with both banana and flat or bulging CPUs.


You can see my bulging CPU will mate with the cooler flat part well.



I tested the Cooler with 240W power draw:



I also found that this AIR COOLER can cool 300W, which is remarkable:



Not more than that. I wonder how could the cooler perform if it was flat on a larger area, it seems it could cool 350W. Or 400W? Who knows... Each heatpipe has probably some limit of heat it can handle.

Numerous reviews stated that Assassin IV cooler is better than Noctuas. I think you can see the reason why here. In comparison with PRECISE Assassin base the Noctua base is a rough donkey hoove.

 

[BoggledBeagle]

I received another message from Noctua from a member of the design team:

Greetings from Noctua!

My name is Dawid I am part of the R&D Team, and I've been looped into this conversation to clear things up.

"You can clearly see in the photo in the thread asymetrical mounting bar (the attachment points are not symetrical to the center of the base), and also the paste imprint of the CPU on the base is clearly off center. The question was, what is the reason for this asymetrical mounting? I hope I do not have to draw lines and measurements on the photos, you can do it yourselves if you have problems seeing it on the pictures.
what are you talking about? See my picture. The placement of the cooler on the CPU is the result of the cooler mounting hardware, not improperly turning screws when mounting the cooler! "

It appears that there was some miscommunication here, as my colleague was thinking that you were asking if the asymmetrical fastening bracket was the cause for the uneven thermal paste spread. After your latest replay, we understand what your original question was.

We take the design-stage of our products very seriously by performing multiple test-runs with different designs.

Yes, the NH-U12A is indeed not completely centred on the CPU.
This decision was made during the design process of the fin-stack design. The goal was to maximise the cooling performance of the included NF-A12x25 as well as the motherboard compatibility of the cooler. Our testing, on multiple platforms, has shown that a slight offset from the centre of the IHS yields virtually no performance loss in addition to being out-weighed by the performance gain of the current fin-stack design.

"You can clearly see in the ink photo, that the shape of the base is not flat, but it could be best described as cylindrical. You can see very unsatisfactory contact with the CPU, where the paste left thick residues, the imprint of the paste closely matches the ink photo, where the thin layer of paste was exactly in the place there there was thin layer of ink and the area was less blue. I stress that I closely followed your installation instructions and all screws were tight.

So my question was, what is your reasoning behind making this particular shape of the base. As I wrote in the thread, this shape might work well will Intel cpus bent in the socket, but it fails badly when the Intel CPU is mounted with the frame, and if would probably also not work well with AMD CPUs which due to the thick heatspreaders probably stay much flatter in the socket."

The base of our coolers is optimised to work well with both Intel- and AMD-based systems. This means that the current base geometry of our coolers is meant to fit well the widest range of possible IHS shapes. This information is based on the respective socket specifications by AMD and Intel, which also specify the surface flatness of the IHS (including the tolerance).

By using a CPU contact frame, your socket no longer complies with the socket specification. This is not something we can design our products for, as your system is effectively one of a kind.

Kind regards,
Dawid ***
R&D Engineer

My response:

Hello Dawid,

thanks for your response.

I am afraid that I do not understand completelly your response about the off center placement of the cooler on the CPU with relation to finstack, because the finstack is connected with the base of the cooler with heatpipes. You can easily correct the off-center placement of the cooler base on the CPU by SHAPING THE HEATPIPES DIFFERENTLY.

While the fact that my CPU mounting method in not according to official socket specification is true, the other important fact is that a lot of people who want to avoid bending the CPUs use exactly the same mounting method. You simply should not ignore this fact.

I have been informed that "Producing different coolers with different base convexity was not deemed necessary until now; however, we are currently looking more into this option for future projects."

I believe it would be highly beneficial if you produced versions of your coolers with TRULY FLAT bases for people who have flat or convex CPUs. This base plate is a separate part of the cooler and you can easilly just solder this flat base plate to an existing cooler with no other modifications needed. Except of the base plate itself there is NOTHING YOU NEED TO CHANGE. There is no reason to wait, you can do this change in a few weeks, literally.

So I would like to know, if you will do this?

You can see in the thread, that a cooler from a competing manufacturter with just a small portion of the base really flat performed well on my CPU, it could cool 300W. I believe your cooler WITH TRULY FLAT BASE can perform well too.

Stepan

 

[BoggledBeagle]

I bought an experimental Noctua NH U12S Redux cooler.

The base is notably worse than the base of the U12A cooler:






I decided to grind the nipple drowning in the sea off. Unfortunately I had just one sheet of coarse emery paper and I gave up trying to flatten the whole base.





The end result is very nice:



You can see that the layer of paste is much thinner now on almost whole IHS area:

 

[sneekypeet]

I bought an experimental Noctua NH U12S Redux cooler.

The base is notably worse than the base of the U12A cooler:

View attachment 339234
View attachment 339235
View attachment 339236
View attachment 339237

I decided to grind the nipple drowning in the sea off. Unfortunately I had just one sheet of coarse emery paper and I gave up trying to flatten the whole base.

View attachment 339239
View attachment 339238
View attachment 339240

The end result is very nice:

View attachment 339241

You can see that the layer of paste is much thinner now on almost the whole IHS area:

View attachment 339242
View attachment 339243

What is the thermal difference?

 

[freeagent]

The crying owl lol

 

[TPCEA]

Following the discussion in another thread, I would like to post some practical information how to check flatness of the CPU heatspreades or cooler base.


The INK METHOD

Simply place a drop of ink (I used black fountain pen ink) on the surface and press against a glass. I used a 10mm thick plate of glass, which I use for lapping, which is really stiff and does not deform, you may get some useful info even with a lot thinner glass.

You can see the results:

View attachment 330476

View attachment 330477


The lapped surface in nearly perfectly flat, except the two right corners.
The CPU excluding a few milimetres on the edges and upper corners is also acceptably flat and I would not feel any need to lap it for normal use. You can even clearly see the grinding marks after it has been ground flat.


The OPTICAL INTERFERENCE METHOD

Under certain conditions you can see a rainbow pattern (with broad spectrum light) or light and dark pattern (with monochromatic light) between a reflective surface of the object and flat glass.

If you had a green light with wavelenght of 0.5 μm (micrometer), each neighboring dark (or light) stripe would mean, that the distance between the two surfaces changed by 0.25 μm between these stripes.

I quickly lapped and polished the above photographed CPU (and destroyed a valuable historic artefact) and got this picture:

View attachment 330501View attachment 330502View attachment 330574

You can see, that in the indicated area there is something like 8 stripes from top to bottom. That should mean that in this area the heatspreader is flat within 8 x 0.25 μm, that is two micrometers! (two thousands of a milimeter)

• Post the results using ink method of your brand new CPU and cooler base, of the CPU before and after spending some time mounted in the socket with normal mounting mechanism.
• Post your results before and after lapping cooler base and CPUs.
• If you had a small piece of glass, you can easily test the CPU while being mounted in the socket.
• Post the difference between stock mounting mechanism of the LGA1700 socket and the mounting frame. (In my experience an LGA1700 CPU remains slightly bent even after being removed from the socket with stock mounting mechanism, which may affect your results.)

WARNING: Lapping a CPU voids the warranty.

ANOTHER WARNING: Do not decide to lap the cpu only after seeing in out of the socket, you need to see what is happening with it when it is mounted in the socket!

EDIT:

If somebody wondered, why the interference pattern does not look very nice, it is because the surface itself in less than perfect - the plating started to wear off and how I said I lapped it just very quickly. I also did not have a proper monochromatic light and no special equipment. This sort of stuff can require special lens to capture it, special shape of the light beam etc.

However it proves clearly, that you can even with a very dirty quick lapping job get to 2 micrometer flatness on a large portion of the CPU.

Here is a picture of how the surface really looks:

View attachment 330566

EDIT 2:

I should add that I put the CPU under the middle of the glass piece and left it press against the CPU only by its weight (2.44kg). In reality coolers press the CPU with much larger force. While changing the force you can observe changing of the pattern, it also depends on how you hold or support the CPU. So if the shape of the CPU looks like "an unimpressive hill", this shape could improve under the pressure of the cooler.

I do not believe this optical interference method has much practical use for a normal consumer, the ink method is so much more easy and convenient...



The overall result - the thickness of the gap between the CPU and cooler base and resulting thermal resistance of this thermal interface depends on many factors: shape of the mating surfaces, flexibility of the mating bodies, contact force, mounting of the CPU and cooler, TIM, etc.

EDIT 3: For anyone wanting to post pictures in this thread: can you PLEASE INPUT THUMBNAILS in the text, not whole pictures? I have no idea why the forum inputs large pictures as a default.

Would it be useful to press the IHS against an ink pad, and then against fine paper? What bond or whatever would be best?

 

[BoggledBeagle]

What is the thermal difference?

I screwed the temp measurement up.

If would be nice if somebody else than me did something. This was meant as an inpirational/educational thread, I really do not need to do everything myself!

The nice thing is that you get a lot of useful info just with a small piece of glass and a drop of ink, you do not need to buy laser measurement apparatus worth tens thousand dollars. Steve from Gamers nexus mentioned how much their apparatus costs, I already forgot, I think it was something between 30 and 50 grand.

 

[BoggledBeagle]

The Owl is silent.

It does not want to commit to making truly flat cooler bases. It is a lazy bird. Ein fauler Vogel.

 

[delshay]

@OP

Welldone!

I worked it out found my lapping to be incorrect & So I'm going to explain which I think is the professional way to lap a cooler. If anyone thinks I'm wrong please correct me.

Below is photo of an incomplete lapping of a vapor chamber. When I shine torch behind it & place a mirror on the cooler no light gets though, but the lapping is not complete let me explain why. You can see I've drawn an imaginary box on the cooler. That box when lapping with a single finger, you must not stray too far away from it, otherwise you will have roll-off on all sides, ( in other word's all the edges will be lapped lower when compared to the center of the cooler..

Now if you stay true & always lap in my imaginary box regardless of any direction,. you should end up with a near perfect flat surface. Move the pressure point too far away from that box, you will tilt the lapping. What i'm trying to say here is, treat the cooler as if it had a very small CPU die in the center & always lap here & never outside it.

Getting back to the photo below, if I had the holy grail of perfect flat surface, there will practically be no thermal paste between die & cooler, as all will be pushed out over the side. As i'm lapping this cooler none of the outside perimeter got a mirror finish even thou my target/mounting pressure is still that tiny imaginary box. This tells me the lapping is still incomplete because at some point the edges will show a mirror finish even thou my target is still that imaginary box.

NOTE: When when lapping to that imaginary box, movement is restricted in all direction, around 3 - 4 mm. Stray too far away from that imaginary box you will tilt the lapping ..not forgetting whatever method is used to lap the cooler, complete coverage of the cooler must be maintained at all times.

EDIT: & always use a single finger. Do not use two fingers as this will change the mounting pressure & tilt the lapping also ...Please feel free to correct me if I am wrong here & I want to see plenty of comments here.

.

 

[BoggledBeagle]

@OP How are you lapping the cooler? You have lapped it incorrectly as the sides have thick thermal paste.

I already wrote:

Unfortunately I had just one sheet of coarse emery paper and I gave up trying to flatten the whole base.

The curve to which Noctua shapes the bases is very strong and If I wanted to get rid of the remaining curved parts at the sides of the base I would need to remove A LOT of material and I had no tools (enough coarse emery paper) and patience to do so.

I was also avoiding the word "lapping" because the end result is not perfectly flat even on the limited area, but it is MUCH FLATTER than what I started with.

You can see the comparison of acceptably flat areas after and before my attempt to flatten it:



I think my area is at least 6 times larger.

 

[delshay]

I already wrote:

Unfortunately I had just one sheet of coarse emery paper and I gave up trying to flatten the whole base.

The curve to which Noctua shapes the bases is very strong and If I wanted to get rid of the remaining curved parts at the sides of the base I would need to remove A LOT of material and I had no tools (enough coarse emery paper) and patience to do so.

I was also avoiding the word "lapping" because the end result is not perfectly flat even on the limited area, but it is MUCH FLATTER than what I started with.

Sorry, You were doing it correctly I edited my posting. I've zoom it on your photo & I can see you are moving in the right direction, so it was my mistake. I'm also learning how to improve my lapping skills

 

[BoggledBeagle]

Below is photo of an incomplete lapping of a vapor chamber. When I shine torch behind it & place a mirror on the cooler no light gets though, but the lapping is not complete let me explain why. You can see I've drawn an imaginary box on the cooler. That box when lapping with a single finger, you must not stray too far away from it, otherwise you will have roll-off on all sides, ( in other word's all the edges will be lapped lower when compared to the center of the cooler..

Now if you stay true & always lap in my imaginary box regardless of any direction,. you should end up with a near perfect flat surface. Move the pressure point too far away from that box, you will tilt the lapping. What i'm trying to say here is, treat the cooler as if it had a very small CPU die in the center & always lap here & never outside it.

Getting back to the photo below, if I had the holy grail of perfect flat surface, there will practically be no thermal paste between die & cooler, as all will be pushed out over the side. As i'm lapping this cooler none of the outside perimeter got a mirror finish even thou my target/mounting pressure is still that tiny imaginary box. This tells me the lapping is still incomplete because at some point the edges will show a mirror finish even thou my target is still that imaginary box.

NOTE: When when lapping to that imaginary box, movement is restricted in all direction, around 3 - 4 mm. Stray too far away from that imaginary box you will tilt the lapping ..not forgetting whatever method is used to lap the cooler, complete coverage of the cooler must be maintained at all times.

EDIT: & always use a single finger. Do not use two fingers as this will change the mounting pressure & tilt the lapping also ...Please feel free to correct me if I am wrong here & I want to see plenty of comments here.

This seems to be a good tutorial how to reach true flat surface, unfortunately I believe sometimes it can be hard to achieve, when you are trying to flatten a large unwieldy object as an air cooler.

One more huge complication is when you start grinding or lapping a very curved object, as was this cooler base. It is very hard to achieve grinding in plane that you need, as there is no reference plane to start with. It might have helped if I soldered/glued some guides in the corners of the base before I started.

You can see in the ink photo, that the edges of the flat area are not parallel and the plane is skewed.

I think I managed to remove a material in this way, approximately:

 

[delshay]

This seems to be a good tutorial how to reach true flat surface, unfortunately I believe sometimes it can be hard to achieve, when you are trying to flatten a large unwieldy object as an air cooler.

One more huge complication is when you start grinding or lapping a very curved object, as was this cooler base. It is very hard to achieve grinding in plane that you need, as there is no reference plane to start with. It might have helped if I soldered/glued some guides in the corners of the base before I started.

You can see in the ink photo, that the edges are not parallel and the plane is skewed.

I think I managed to remove a material in this way, approximately:

View attachment 340123

You still have what looks like nickel plating on the cooler. If so you have not completed full lapping of the cooler because we should see "all copper". Given that you say you ran out of emery paper I assume you will finish it later.

I''m using sandpaper & my method in lapping is different as it's GPU cooler which has four screw sockets, so it can't be lapped on a flat surface. Anyway lets' talk flatness because that's the driving force behind this thread. From what i'm seeing here in early lapping I am 100% right what I said in post #96. No light can be seen with a torch place behind the mirror in most direction as I shift the mirror around the cooler.
What I'm after is to understand what cause the edge of lapped cooler to be slightly lower than the center. What I said in post #96 has fixed this. No light getting through in most direction right up-to the edge.

But like I said, if the target/mounting pressure is always the center where the overdrawn box, at at some point it must affect the perimeter. It was starting to see improvement towards the perimeter as I was lapping, but I got tired because it's a very slow lapping process because I can't move no more than 4mm back & forth, so it takes a long time. .


LAPPING DIRECTION ( EDIT )

Most of my lapping must be done East to West, The reason for this is the shape of the "GPU/HBM die" & the length of the coldplate being longer in this direction. Lapping my heatsink mostly East to West will reduce roll-off around the perimeter significantly also, again due-to it being longer in this direction.. You should be lapping to the shape of the IHS if the coldplate is equal all side.

At the end of the day I want extreme edge to edge flatness & I'm on the right track for achieving that goal thanks to this thread..