Fitting M@g Heatsinks

[Alan B]

I purchased some P7 heatsinks here on CPF. They are nice, but not large enough diametrically to be snug in the M@g body tube (they are a close fit, but not snug). So I am considering ways of making them slightly tighter.

I am planning to use a d2Flex, and I may want to disassemble the light later on to access the PCB so I don't really want to glue the sink into the tube solid.

I have seen some use a little bit of epoxy so it can be broken, but I wonder if other methods might be better. One that occurred to me was to lightly knurl the OD of the heatsink. I might be able to do that in my lathe, but have no experience with knurling.

Another simpler possibility is to put a few punch marks into the OD of the heatsink which will raise a little material and make it tight. I do plan to put some heatsink compound (non hardening) between the sink and the body tube.

I would really like to put some rods between the sink and the M@g switch and make them into one unit, but that is more precision work than I want to go through right now. I've got a few of these to do for friends.

Thoughts, suggestions, other ideas??

 

[jasonck08]

You could put a thin sheet of copper or aluminum around the heat sink to make it fit snug...

 

[Alan B]

You could put a thin sheet of copper or aluminum around the heat sink to make it fit snug...

Thanks for the suggestion.

The space is approx 0.004" diametrically (measured with digital calipers), so 0.002" shim stock would be about right. If we make it too tight we won't be able to get it in or out, though. The aluminum shim stock I found is "red", wonder if that is oxide or ??. Wonder what the heat transfer is through two more interfaces, and there won't be much room for thermal grease, which might be okay.

Another thing I thought about is to turn the sink down a bit and make a fitting sleeve, but that is a bit of work and at the end of the day probably would be either too tight to readily get out or too loose to reliably stay in. I guess I'm thinking that a little bit of interference is needed, but not the whole surface, if it is to be readily removable.

 

[65535]

The color coding for .002" shim stock is red, it's paint.

I might try drilling a diagonal hole int eh heatsink by the lip, tapping it and using a setscrew to hold it in place. If you use shim stock with a setscrew it should be pretty close. With shimming you don't actually have to make an entire sleeve just use the shim to center the piece.

 

[Mirage_Man]

I usually use thermal grease like Arctic Alumina. Once it sets in it shouldn't move very easily.

 

[tx101]

I am not a machinist so just ignore me if this sounds stupid

Make a shim (more like a collar) and thread the inside hole
Turn down and thread the heatsink so that the heatsink
now screws into the shim

The only problem with this is that your wires going from the D2flex
to the Mag switch will get twisted when you screw the sink in

 

[precisionworks]

measured with digital calipers

The best digital calipers in the world (Starrett, Mitu, etc.) are no better than +/- .001", which is why they are not used for precision fits. You may want to mic the OD of the heatsink, then measure the ID where it needs to go - an inside mic, or expanding bore gage is the least expensive way to do this.

With as little room as you already have, the digital caliper introduces a huge amount of error and variability.

 

[will]

Just some thoughts -
You might dimple the outside diameter of the heat sink. That will give you some extra tightness. This is most likely the easiest to do.

Another way might be to create a burr on the outside diameter of the heat sink. Put it in the lathe, hold a blunt edge against it and push in, this should raise and edge.

if you want something to give tightness and be easily removable, you would need to do something with a screw and a tapped hole, if there is room. Think of something like an expansion arbor, as you tighten the screw it will expand the diameter. This could be near the edge of the heatsink.

Also - keep in mind that not all Maglites will have exactly the same inside diameter....

 

[Alan B]

The best digital calipers in the world (Starrett, Mitu, etc.) are no better than +/- .001", which is why they are not used for precision fits. You may want to mic the OD of the heatsink, then measure the ID where it needs to go - an inside mic, or expanding bore gage is the least expensive way to do this.

With as little room as you already have, the digital caliper introduces a huge amount of error and variability.

Thanks for all the comments.

I'm sure you are right. These are Mitutoyo, and the measurement was made with the same caliper at the same time, but that's the best I can do. No micrometers over 1" here.

 

[f22shift]

the thermal glue they have on dx, fujik?, dries like a caulk. holds but non permanent.

 

[Alan B]

I purchased some P7 heatsinks here on CPF. They are nice, but not large enough diametrically to be snug in the M@g body tube (they are a close fit, but not snug). So I am considering ways of making them slightly tighter.

I am planning to use a d2Flex, and I may want to disassemble the light later on to access the PCB so I don't really want to glue the sink into the tube solid.

I have seen some use a little bit of epoxy so it can be broken, but I wonder if other methods might be better. One that occurred to me was to lightly knurl the OD of the heatsink. I might be able to do that in my lathe, but have no experience with knurling.

Another simpler possibility is to put a few punch marks into the OD of the heatsink which will raise a little material and make it tight. ...

Thoughts, suggestions, other ideas??

The punch marks worked very nicely. I recommend them.

 

[LukeA]

I'd take a length of heavy duty aluminum foil about 9 inches long, get a very light coating of AS5 on on both sides, wrap it around the heatsink (twice) and install. Heavy duty aluminum foil is about 0.001 in. thick, so it'll fill the gap. Regular foil is 0.0005 in. or so thick, so the length heeds to double.

The above, of course, is true if your 0.004 in. measurement is accurate.

[didn't read the last post prior to posting this, but I think it may be helpful to someone someday, so I'm leaving it.]