Best thermal pathway for heatsinks

As long as you have to have either a custom-made or an intentionally matched heatsink, I have to ask, why use a module? May as well just add a few features to the custom-made heatsink.

It is much easier to align and swap-out the diodes...

Not everyone is able to press their diode directly into the heatsink...

But aren't they already assuming everyone is able to press-fit the module into their heatsinks?

With flashlights, there are P60 hosts that accept p60 drop-ins. These are cool because you can swap out whatever drop-ins you want. However, they have pretty bad thermal performance, and you have to wrap it in foil to even attempt to use any high power drop-ins.

Using a copper module with a plain jane set-screw heatsink is akin to using a high performance P60, built on copper slug with LED direct soldered to copper, 12AWG wiring, gold plated springs in an Ultrafire P60 host. The host is designed to accept ANY p60 (kind of like generic heatsinks designed to accept all aixiz, despite any possible tolerance issues), and therefore, the drop-in makes unacceptable thermal contact with the host. No matter how awesome the guts are, if the guts can't dissipate heat to the body, you are limiting the performance of the device.

I propose that we use the already-standardized aixiz modules' thread pitch and size as the interfacing part on the heatsink, rather than a >12mm hole where the module is held in by a tiny setscrew. You can still use a setscrew to ensure that the user doesn't unscrew the module instead of just the lens, but a setscrew design does not provide optimum heat dissipation.

Did anybody try my thought exercise? Perhaps it was too complicated. This one is simpler, but not quite the same. Close enough:
Imagine heatsinking your aixiz module by simply pushing it down real hard onto a flat plate of metal. That is the way a setscrew design works. Setscrew pushes on aixiz module, aixiz module contacts heatsink along a tiny line.

Instead of using a setscrew, why not just screw the module directly into the heatsink? (and apply some epoxy if needed + thermal paste)

Or maybe even use copper seal tape for the threads (if that's a thing). Just thinking in terms of reusablilty and better contact

I'm not sure if threading a heatsink to accept the front of the modules is an easy task, especially when dealing with copper. But I'm sure there are many talented machinists.

I'm also thinking that the thermal performance would be slightly worse than pressing the module in with a vice, as there would be significant air gaps if the module is allowed to rotate to screw in.

ryansoh3 said:

I'm not sure if threading a heatsink to accept the front of the modules is an easy task, especially when dealing with copper. But I'm sure there are many talented machinists.

I'm also thinking that the thermal performance would be slightly worse than pressing the module in with a vice, as there would be significant air gaps if the module is allowed to rotate to screw in.

Click to expand...

Actually, having tried to machine copper before..it's extremely soft to work with.
You've got to take extra care when machining it especially doing threading. It dings and scratches very easily.
:can:

Cel said:

Instead of using a setscrew, why not just screw the module directly into the heatsink? (and apply some epoxy if needed + thermal paste)

Click to expand...

Heh, sometimes I get a bit long-winded!

ryansoh3 said:

I'm not sure if threading a heatsink to accept the front of the modules is an easy task, especially when dealing with copper. But I'm sure there are many talented machinists.

I'm also thinking that the thermal performance would be slightly worse than pressing the module in with a vice, as there would be significant air gaps if the module is allowed to rotate to screw in.

Click to expand...

It's hard to know without doing experiments, I guess. Both designs have benefits, and it is hard to guess at which one would outperform the other. My main thought is that maybe not everybody has a benchtop vice. What methods of fairly easy assembly will still allow us decent performance?

If I had the means to press parts, I would go with a direct press heatsink, and buy a diode press tool to aid with this, or use those nice copper modules pressed into a mating heatsink. If I didn't have the means to press parts together, then I'm stuck with using a crappy set-screw setup. Hence the idea to interface heatsink to aixiz using female threads. A diode back contacting surface could also be realized if you cut threads directly into the heatsink. The threads would squeeze the back of the diode onto the heatsink. These solutions ought to create a better performing system than our current easy solution of setscrew and hole that can still be assembled by folks without specialized tooling.

Right now I'm just throwing ideas out, but wouldn't it be amazing if we could all agree on what the next new "standard" should/could be? We should be trying to expose all the pros and cons of each solution in terms of ease of use, performance, and manufacturing feasibility. For example, it could be that cutting threads into a heatsink can't be done on lathe, which would mean it would be much harder and expensive to do that, which would be a major con. I think performance will have to be sacrificed a bit for usability, but if you have a super efficient diode and want to push it to the limits, you'd probably use a better solution anyways; something like a diode pressed directly into a heatsink, or a copper module pressed into a mating heatsink.

This is the new one I have coming in.


[/URL][/IMG]
[/URL][/IMG]


The drive goes in with a press fit ring to hold drive in to make ground .You can set the drive and test it before you add module no need to screw module so wires will not get damaged . This one will fit 20mm or 17mm driver or small square driver. If you use the square driver then add 17 or 20mm blank to make +and -
I have copper ones coming in too. Will fit a lot of host some C5 Q5 -MX900

I keep working to make my heat sink better and better the hole is 12mm so you can use the Aixiz module

Nice heatsink design, I like that it will fit several hosts, excellent work.
When will you offer them for sale? Please provide more details on the copper one.
Love copper's thermal properties even though it is a pain to machine.

Should have them in about 3 weeks $35 on copper

COMMANDR said:

Nice heatsink design, I like that it will fit several hosts, excellent work.
When will you offer them for sale? Please provide more details on the copper one.
Love copper's thermal properties even though it is a pain to machine.

Click to expand...

I just got a module back half with a 3.5W 445nm diode in it and used it to bring my DominatoR to life. If any out there have a dead DominatoR and want to bring it back, I highly recommend getting any 9mm diode, and maybe even some 5.6mm diodes made with one, the short one on the left in post 27 and again in 29. it is a very tight fit though, Perfect fit but very tight, I had to press it in with considerable force. future extraction from the DominatoR heatsink, if required, may be a PITA. but you will have a better heatsinked diode than the original, which is good in my case because the new diode in it is the new, more powerful one.

for those interested, I did this because my PL-E pro 4W is now doing only 1.3W, the diode degraded already, more crap luck for me. I'm preparing to send it in. I decided to have the 4W 445nm pro repaired using a new driver and a 1W 520nm diode instead, for a few reasons.

1. the 445nm diode puts off more heat than the 520nm diode, and I think the above setup may provide better heatsinking than the pro. it is just a guess, I have not seen the pro's guts, but the DominatoR heat sink is over a troy lb of copper.

2. the pro does have better beam specs, because they used additional lenses to correct the beam some, causing power loss. my version of the pro 445nm is driven at 3.7A. honestly, since this is my burning laser (well, master burner, among burners), beam specs are not as important to me. I'll take the divergence hit, and get a power boost (almost 1W using a single element g lens), as well as driving the diode 200mA lower, at 3.5A. so in theory, better heatsinking and lower current will lead to longer life, this way I'm not replacing too many $300 diodes. and like I said, this DominatoR 2.0 burns better than the pro did, by a little.

3. the pro has the BE, since the raw divergence of the 520 is as good as the last gen 445nm 9mm diodes, it'll be sweet to have a BE to increase the range of such a bright laser. I figure good to have the brightest laser have the widest beam.

@ BB, Moh ran a GB for modules, without threading on the back, so yes they do exist. I even have a selection at home in my draw.

Ps IIRC the dimensions were also discussed in his thread :beer:

rhd said:

Anyone know what the rear thread pitch of standard Aixiz modules is? We should make a better back.

Click to expand...

M11x0.5

eta: oh, this was posted earlier. I hadn't read up to that point yet. Oh well

grainde said:

@ BB, Moh ran a GB for modules, without threading on the back, so yes they do exist. I even have a selection at home in my draw.

Ps IIRC the dimensions were also discussed in his thread :beer:

Click to expand...

I got in on that one, or one of them, but unfortunately all the modules in the order were switched to threaded types to simplify the order. Now that DTR and others have full-copper backs though, it doesn't matter so much anymore.