Solid pill improvement with copper core - useful?


I got a lathe a week ago and I’m thinking about some idea I had.

If I make a solid pill out of aluminium, but add a copper core (let’s say 10mm) as some type of “heatpipe” - will this improve the heatsinking?

I just ordered 15mm and 10mm copper - mainly to make those spring-caps used in many tailcaps.

Like this:

Using a copper star (no, not soldering together the two copper parts…) should work fine.
Or not?


This sort of concept is how many entry-level PC cooling systems work—but I would make one modification to your proposed design.

Below is an image of the stock Intel cooler (the design has been pretty over the years)

As you can clearly see, it has a copper core, surrounded by an aluminum heatsink. The copper is press-fit into the aluminum.

If I were you, I would allow the copper core in your design to go all the way through to the top—that way the star is transferring heat most efficiently to the core, and then being radiated out to the aluminum.

I’m surprised more people at BLF don’t do this sort of thing—it can offer much of the performance of a full copper block at significant cost savings!

Yeah, and by doing so you would keep the possibility of soldering a copper MCPCB onto your heatsink.

It goes all the way to the top - on top there is the LED-Star :wink:

Didn’t think of those CPU coolers. Of course…

OK, then I know what to do.
I hate copper because of the price and the bad machinability (at least with my hobby-lathe without liquid cooling). Aluminum is cheap and easy to work with.

I will get some drills up to 9.9mm (only got a set 1-6mm) and then turning the copper to a press-fit.


How do you want to connect the MCPCB to the heatsink?

Ah, I see it now—I thought that was part of the pill at first! Copper can certainly be a bear to work with, I’ll give you that much. Good luck with this project.

That draft is just a quick sketch to visualise my idea :wink:

With a copper MCPCB, tha temptation to solder it together is just extremely high - I think, someday I will try it.

But I have mostly non-copper stars, so I will use some normal heat transfer paste.

New lathe aye and no pictures? BLF rules, no pictures or it did not happen, ok, please can we have pictures of your new lathe?

Maybe. Maybe not.
I have to clean the mess here up and get a bit more free time, then I can make a thread about it :wink:

Just finished 3 pieces of aluminum for a fake SolarStorm X3 light without any heatsinking.

Congratulations on the posession of the lathe! I hope you will make many nice things with it :-)

But where is comfy when people are in need of getting disappointed?... Oh well, I will do his job for now :

Ok, it is very unlikely that you will see any output difference using a pill like this compared to a pure aluminium one. The only place where copper makes a noticable difference is directly, say within half a millimeter, under the led, any further from the led the heat is already so spread-out over a large surface that any well heat conducting material will do: aluminium, brass, those are really well conducting materials!

:-) And I will play the devils lavyer then...

But what djozz said is still not tested in a light box!

It is what many of us think. It is however not tested AFAIK.


I know that the copper isn’t useful 20mm under the LED - but it’s easier to make :wink:

500mmx10mm copper (E-Cu) costs 7€ - so it’s not a problem to use 20mm of that.

First I must admit that copper makes you feel good :-) , I love using copper for making things!

A direct comparison using a lightbox will immediately settle the discussion but is not needed.

Look again at my XM-L ledboard comparison, there is no difference in light output between a copper and an aluminium Sinkpad. These are differences happening extremely close to led where all the heat still goes through a very narrow channel and yet they have no effect, how would you expect any effect so much further away where the 'heat channel' is 100 times wider?

I am not sure djozz.

I just know that when soldering the wires on large brass or copper pills where the star is reflowed to the copper/brass then sometimes the soldering iron gets stuck to the wire pad because the solder solidifies. That has never happened on anything else I've build using other materials. And I have tried a lot of TIM approaches.

But maybe there is no difference at all. Could be the power levels are too low. Certainly a diffence between the 15 W and up of a XM-L2 and a 60W soldering iron. It is certain that the alu alloy that the sinkpads are made of is carefully selected for its thermal properties. And they do a great job. No protests there.

I’m digging deep into the guts of BLF - but I found this old thread and I will use it :smiley:

I have now a lathe, aluminum, copper and a lot of tools that are needed to make nearly everything I want.

I made a test with a SK-68 as a working light

I used 10mm Copper and drilled a 9,9mm hole. I grinded the copper down so I could use it perfectly and glued it in with Fujik - I don’t know if that is necessary, because the fit is very tight and you can’t see a gap between the two materials.
I was surprised how good copper is to drill and work with - I imagined it worse than that.
Well, it worked pretty flawless and this try is a full success.

And pressed into the bored out original pill

I can’t say if anything has improved, because the Nichia 219 is only getting 600mA (2W), because the host otherwise overheats. :stuck_out_tongue:

I have a question though - What is better:


I’ve been thinking about this for quite a time - I think the second one will be better, because the heat has to go to the outside to the surface while the first one has too much thermal mass.

I will use that with a single 18650 light @ 3*XP-L on Noctigon, so maybe 15W to be disposed.
15mm Copper this time.


seems counter-productive to drill it out.
I’m seriously jealous of your lathe :heart_eyes: . The work looks great.

It’s a cup shape, not a solid core if the one in front of me is any indication (you bet they wouldn’t waste more copper than needed on a cheapo HS).

Try putting some WD40 on the piece and on your tools (definitely works on files, don’t know about lathe tools).

I believe you - are you going to repeat the test with quad MTG2s at 6A each? :stuck_out_tongue:

This is a bit off topic, but I wonder if cutting groves around the area (to work as thermals ) would work. When I solder to copper (90% copper alloy coins to be exact) I end up having to drill holes through the material to attach wires.

As far as I understand it, less thermal mass -> less thermal capacity -> the entire thing gets hotter (more °C per mm³) compared to a larger body. Only exemption is having airflow (or other coolant) to the thusly increased surface area.