I have seen mentions of "lapping" copper emitter pads, Sinkpads, Noctigons etc. what exactly is lapping, why is it nessesary and how is it done?
Take a piece of glass, lay down some 400-grit wet/dry sandpaper, and lightly sand the bottom of a MCPCB then check the pattern it leaves. You might be shocked at how not-flat it is.
What you do is use a coarse grit paper to make it flat (220 grit or so), then use successively finer and finer grit to get rid of the scratches left by the previous grit. How far you should take it depends on how it's going to be used - if it's going to be soldered, lapping isn't needed at all. With epoxy, it just needs to be flat so the layer of epoxy can be as thin as possible. If it's held down by screws or clamped in place by the reflector and you're using a paste thermal compound, the shinier and flatter you can make it, the better.
BUT, unless you also have a way to get a similarly flat surface on the thing the MCPCB sits on, it's mostly pointless (most pills make it really hard to get the MCPCB mounting surface in contact with a flat sheet of sandpaper). You can make it shiny without making it truly flat, but that's mostly pointless too. The flatness is more important than the final surface finish (within reason of course).
lapping is getting the two mating surfaces as flat as possible. you want to get all imperfections out to get the best thermal pathway with as little thermal compound as possible
Lapping is grinding the surface so that is is completely flat. This ensures best contact and thus best thermal transfer.
It’s normally done by taping a sheet of sand paper on a flat surface and swiping the MCPCB across many times.
You proceed to higher grit sandpaper so that it becomes smooth and mirror like.
Not exactly necessary, sometimes this is done to remove the nickel plating of some MCPCB’s as well.
I sorta lap my CPU’s if I feel that they’re not flat.
Good question. It should create an interesting discussion.
Of course, we know lapping is done to create 2 flat true surfaces in order to have maximum physical contact.
Makes sense, but I wonder if having two true but rough surfaces creates better thermal transfer when a thermal compound is used to mate the two surfaces. It is well known that rough surfaces emit more heat. A rough surface has more surface area than a smooth surface. So two true but rough surfaces against each other should have more surface area to emit and receive heat.
I know that solder/thermal compound will not conduct heat as well as aluminum or copper, but the increased surface area and small amount of compound (short bridge) may perform better. I don’t know this, but it may be the case.
I too was wondering if a “brushed” surface would be better than a highly polished one.
The thermal compound doesn’t conduct heat as well as aluminum or copper, and having two rough surfaces are much harder to get maximum contact area. It’s easier to get the two as flat as possible, and the thermal compound to replace any minuscule air gaps.
In a perfect world, it’d be best to have two perfectly mating surfaces and no thermal compound at all, but that’ll never happen so that’s why the thermal compound is used.
I suppose two polished surfaces with compound would create a sort of surface tension.
Thermal conduction relies on physical contact, which is easiest to guarantee with two flat smooth surfaces. Theoretically, two zig-zag surfaces that perfectly mate will give a better conduction as the contact surface area is increased. However, reality sets in and making two perfectly mating zig-zag surfaces is not easy. Easiest is two flat surfaces.
I recommend lapping the star and pill to remove any bumps that may be on those surfaces. Pills are really a pain because most have a lip around the edge. You need a rod with a flat end to get at the mating surface. A plate glass surface is best, but just a clean, smooth desktop will do (I often use a large aluminum heatsink). Use a randomish circular sanding motion.
As an eye-opening test, paint the surface of the star (or pill) with water-based paint and set the star on it. Then, look at the bottom of the star. The places where paint transfers to the star is where physical contact occurred. Usually the result is not much contact at all. Wash the paint off before it sets. 
This is where the Thermal Interface Material (TIM) comes to the rescue. It fills the gaps to improve contact. TIMs usually have no where near the conduction of the star or pill materials, but they are much better than air.
It does, and that’s why I ‘wiggle’ the star around some to spread the TIM underneath. A good portion will squeeze out the sides, which is a good thing.
I have a few different types of TIM, and I find the thicker, pastier ones harder to get thin.
Or Poor Man's Machinist's Dye, also known as a Sharpie. :p
I find that Sharpies dry too quick. Paint has issues too, as it doesn’t go on in an even coat.
Pills surfaces are surprisingly non-flat. Probably should not be surprised though… most budget lights aren’t exactly precision machined 
With the Sharpie, color in the area with dots, not scribbles. Even when dried it will still transfer.
Sharpies are great for stuff like fitting a reflector into a light it wasn't meant for, to see where a cut needs to be made when you can't see what's hitting where.
Thanks Comfy, I’ll have to try that again. Maybe I didn’t apply it right. 
So, how DO you lap the emitter end of a pill when it is countersunk? I filled in the emitter end of a Defiant AAA light, but now I’m trying to figure that out. Here’s what it looks like now:
That piece that I added is kind of rough, and I was hoping that I might be able to use some kind of wheel on a Dremel, but I don’t see any that are completely flat (they all have a screw in them).
I used to have to lap hydraulic components to ensure a perfect seal at high pressure. We lapped on a machined piece of stone (granite?) with a full sheet of very fine wet-or-dry sandpaper and heaps of kerosine to stick the paper to the stone, and to stop the sandpaper clogging. I was taught to apply zero downward pressure to avoid sanding on an angle, just the weight of your hand is enough. I was also taught to sand in figure 8’s, rotating the component 1/3 of a circle every few figure 8’s to ensure even sanding as well. Hope this makes sense?!
I got Roche F8 from FastTech just so I could swap out the emitter. The pill and body are integral and deep. Dead center, under the star was a deep conical depression. It looked like the spindle of the lathe was attached to that spot (directly under the emitter). There was nothing to be done but fill it with arctic silver. There were screw holes, but no screws. I put some in , tightened it down and it seems to transfer heat well.
As I understand it, lapping is what you get at certain, uhhh, “gentleman’s clubs”. You have to pay extra for it, and the surface seldom stays flat… :party:
Would [quote=relic38]
I find that Sharpies dry too quick. Paint has issues too, as it doesn’t go on in an even coat. Pills surfaces are surprisingly non-flat. Probably should not be surprised though… most budget lights aren’t exactly precision machined !http://i00.i.aliimg.com/img/pb/088/585/407/407585088_409.jpg!
[/quote]
bb
Would dykem work?
I just want to share this information:
Our Noctigon boards are actually lapped before the plating process to remove grooves that exist on the copper substrate.