1765 Lumen P60 Quad XPG Silver Plated Copper Heatsink Direct Drop-in, in developement

[VanIsleDSM]

I must have calculated the angle wrong, I can't remember what the decimals were, but from your measurements I came up with 9.xx degrees I think, I don't have it in front of me.

So.. it should be no big deal, take a little more material off at the taper and it should sit right in.. I know it's not going down as far as I designed it to, because the contact, the lowest part of the driver, is supposed to drop in past the spring perch that holds down the 18650, so it will make contact when the material has been removed to allow it to do so.

I should have my 6P Defender soon, I ordered from a place only an hour and a half ferry ride away, so it shouldn't take too long, and then I can perfect the fitment.

I am just curious about whether there is a gap with the Malkoff and the Solarforce, as without the grounding spring, the small gap ensures that the head when screwed on, pushes the drop-in onto the spring perch to make good electrical contact. If there is no gap, there would be no pressure, just the fit of the heatsink on the walls.

I think it is prudent to either design this drop-in with a small gap, so it is pressed against the spring perch for contact, which means it would have to be large enough to have a small gap on the Solarforce, which would mean a slightly larger gap on the Surefire, or to use the grounding spring to apply pressure for current transmission and go for no gap.

I would like to make the drop-in with no gap, as I kind of have a fetish for OEM quality and fit in all of my hobbies. This requires me to make the custom spring, and will add some expense, maybe $10 or so overall vs no spring?

How important is "the gap" to all you CPFers out there? What would you like to see done?

 

[Kestrel]

Is there a gap with the Malkoff on the solarforce host, or does it only have a gap on the surefire?

Very good question. My Solarforce host is so deep as to not securely retain an M60 with the relatively deep SolarForce head/bezel or the somewhat shallower SureFire Z44 head/bezel. For this type of configuration, I have to place two washers in the SolarForce to elevate the M60 somewhat & provide a positive reference surface:

Regarding the 'quad' in this host: This 'spring shelf' is too wide (measurements provided on page 2 of this thread if I recall) for the 'Quad' spring to seat on in its current configuration, so one of these washers still has to be used to give the spring enough of a shelf to reference off of. Utilizing one washer for this purpose (0.033" thickness) results in a secure assembly with a resultant 'quad gap' of only ~0.004". A very good fit. I do dislike these outer springs very very much - it makes for very difficult compatibility between hosts & drop-ins, some of which are designed for this spring (SolarForce) and some of which are not. 😡 Definitely the bane for us 'SureFire LEGO' folks.

BTW, even this fitment does not permit a 'button-top' AW 18500 to make an electrical connection with the 'quad', although the forward end of the cell is fully seated and is resting against the thin washer. The top of the 'button' is approximately flush with the top of the washer. :sigh:
Since I keep on running into this issue, I would very much recommend adding a small copper 'post' to the bottom of the drop-in to provide reliable electrical connection to various button-top/flat top & 'free-floating'/'shelf-retained' LiIon cells. Something like 0.100" - 0.150" would be perfect IMO - the spring in the tailcap will take up pretty generous tolerances in this regard.


Here's the "test bench", with four separate light sources for taking (somewhat) decent photos: 😎

FM measurements (listed below) I value highly and are pending:
(& do not give the same difficulties that I had with the SolarForce host above)

While I would like to get these done, I've been LEGOing, measuring, photographing, uploading, & posting for the past three hours now. I'm tired now & am done with this for the evening. 🙂

FM host / SureFire head/bezel
Malkoff gap:
Quad gap:

FM host / Solarforce head/bezel
Malkoff gap:
Quad gap:

 

[Kestrel]

I think it is prudent to either design this drop-in with a small gap, so it is pressed against the spring perch for contact, which means it would have to be large enough to have a small gap on the Solarforce, which would mean a slightly larger gap on the Surefire, or to use the grounding spring to apply pressure for current transmission and go for no gap.

I would like to make the drop-in with no gap, as I kind of have a fetish for OEM quality and fit in all of my hobbies. This requires me to make the custom spring, and will add some expense, maybe $10 or so overall vs no spring?

How important is "the gap" to all you CPFers out there? What would you like to see done?

Good questions. In my opinon, this drop-in would be best without a spring for using in the SureFire hosts. Holding tight tolerances with the heatsink while utilizing the slight compressiblity of the optic could pretty much eliminate the gap compared to the Malkoff units - those are 100% incompressible, that's for sure. However, the slightly deeper SolarForce head/bezels change the situation enough to provide difficulties in this regard, since these begin to be more popular on the good non-SureFire hosts such as the FiveMegas.

In addition, the deeper 'well' in the SolarForce/etc hosts pretty much require a spring for positive contact, as we've now seen.

The 'recessable' spring was an interesting idea, but getting it to work would entail it being small enough in diameter to fit in the relatively narrow-throated SureFire hosts, while it being large enough so it positively references off the SolarForce/etc somewhat wider spring shelfs. :shrug:


FWIW I just live with the 'Malkoff gap' in my FiveMegas and SureFires, and can use a washer in my SolarForces (resulting in a 'Malkoff gap' in those as well). I don't use any sprung LED dropins, just a narrow-spring SureFire P90 in my SolarForce - so that unit is the only one that doesn't get the gap. <shrug>
I have a Malkoff in a G2L, there is pretty much no gap in that combo, although there is some variability (& some flexibility) in the nitrolon G2's.

I just thought of one significant disadvantage to your units having a 'gap': Any significant frontal impact would simply destroy the plastic optic riser posts, which the head/bezel is essentially referencing off of & creating the head/body gap. The Malkoffs are OK in this regard - a monolithic piece of brass as a front reference surface.

Edit: You know what might be worth it? A 'solid' unsprung drop-in fitted perfectly to the SureFire (& fitted reasonably well to the FiveMegas) - therefore, good thermal contact on the side, the angle, and the bottom reference edge. (however, I believe that there is a Malkoff patent on this type of configuration, so something needs to be different! My guess is that the Malkoff patent uses the heatsink to conduct heat to the head/bezel assembly. Your drop-in doesn't do this - the optic is in the way of this thermal path - so maybe this isn't an issue.) For an additional $10, a precision aluminum insert could be supplied to fit to the 'larger well' SolarForce/etc hosts, with an interior made to duplicate the original SureFire interior. That will result in the best thermal fit possible to a wide variety of hosts. This completely dispenses with the spring (and all the associated difficulties) and maximizes thermal contact in all of the popular hosts.

 

[Kestrel]

I must have calculated the angle wrong, I can't remember what the decimals were, but from your measurements I came up with 9.xx degrees I think, I don't have it in front of me.

So.. it should be no big deal, take a little more material off at the taper and it should sit right in.. I know it's not going down as far as I designed it to

BTW, don't take that M60 bevel location (that I placed immediately adjacent to the quad's bevel location in my earlier picture) as gospel, in all likelihood it doesn't mate to the angle in the Surefire hosts, and there is probably an air gap in that angled region between the M60 and the SF 6P body.

I hope that makes some kind of sense, I'm getting very :tired:, I stayed up way too late last night posting my initial impressions on this unit, so I don't want to say up too late tonight... 😉


Edit: OK, that's it for tonight - I've made a number of edits to my last few posts with a few ideas - clarifications - new thoughts, etc. After this evening's session I'm a little burned out on this, I'm going to try to take one or two days off.

 

[VanIsleDSM]

Haha.. Don't overwork yourself!

I'll get it all figured out once I have my Surefire host. I appreciate your hard work and your ideas, they'll definitely make the final product better!

I had already considered making a piece to drop into the clones that gives them the same profile as a genuine 6P. Would be nice to also include a sleeve to make the inner diameter of the clone the same as the 6P, but the material would be too thin, wouldn't be possible. Copper tape or something will have to be used.

There will hopefully be another option for a host too, to make an improvement in that area. I'm working on a prototype that will have fins for better heat dissipation, and a silver plated copper sleeve where the drop-in sits to transfer the heat to these fins to be dissipated. Sleeve will also include gold plating on the spring perch for electrical contact. When you're working with 5A, it makes all the little things add up to a big difference. I can't say for certain yet if everything will all work out with the build, but if it does they'll be available with the FET switch, UCL lens, and anodized either red, green, blue, yellow, or black (custom colours also an option). The silver plated copper sleeve also allows the use of liquid metal instead of AS5, to provide a far superior thermal contact.

 

[BigHonu]

A HUGE 'Thank You' to Kestrel and VIDSM for working this stuff out! Looking forward to the results!

 

[Noctis]

Edit: You know what might be worth it? A 'solid' unsprung drop-in fitted perfectly to the SureFire (& fitted reasonably well to the FiveMegas) - therefore, good thermal contact on the side, the angle, and the bottom reference edge. (however, I believe that there is a Malkoff patent on this type of configuration, so something needs to be different! My guess is that the Malkoff patent uses the heatsink to conduct heat to the head/bezel assembly. Your drop-in doesn't do this - the optic is in the way of this thermal path - so maybe this isn't an issue.) For an additional $10, a precision aluminum insert could be supplied to fit to the 'larger well' SolarForce/etc hosts, with an interior made to duplicate the original SureFire interior. That will result in the best thermal fit possible to a wide variety of hosts. This completely dispenses with the spring (and all the associated difficulties) and maximizes thermal contact in all of the popular hosts.

Wouldn't it be easier to wrap it in copper tape? I bought myself a buttload of that and AS5 for my other drop-ins for the sole purpose of maximizing thermal/electrical contact between the module and the host. I figure other people would have done the same.

Also slightly off-topic here, but I ended up with a ton of abrasives for metal. Would heat transfer be improved if I were to polish the interior of the host or drop-in to a mirror finish? Or would that completely mess up the fitting of both parts?

 

[Kestrel]

OK, after a short break & thinking about things some more, I think that my best route for me to contribute in the short-term (i.e. messing with this prototype in my hosts) might be the following:

A fully seated Malkoff M60 of X1 inches in total length rises X2 inches above the top of the host (when the head/bezel is removed), and creates a gap of X3 inches when the head/bezel is screwed down onto the host.

The current Quad prototype is Y1 inches in total length and rises Y2 = (X2+a, since it protrudes 'a' more than the M60) inches above the top of the host, creating a head/bezel gap of Y3 = (X3+b, since it creates a larger gap than the M60) inches. Therefore, the quad really needs to have a total overall length of X1 minus X3 to get no gap whatsoever. The shortcut for this situation is to take the X2+a length and turn it down to a length of X2 minus X3. Makes sense? :thinking: So I find 'a' and 'b' by just taking a couple of comparative measurements. Problem solved, the exact total length (Y1 new) and seating depth (which is related to Y2 new) (defined by the drop-ins angle/bevel), for the next quad prototype will then be known, resulting in the correct length to result in zero gap when placed in the SureFire 6P. :thumbsup: I'll try to get some measurements on this tomorrow - I'm hoping that I could provide some info so you don't have to do trial-and-error lathe work to get a fully-seating drop-in in your SureFire host that is currently in transit.

I know for a fact that the FiveMega P60-type hosts get a somewhat greater 'Malkoff gap'. So a zero gap in a SF 6P would still result in a small gap in the Fivemegas. I've LEGO'd a fair bit with my stuff and unfortunately this seems to be pretty much unavoidable. 🙁

Then there is that issue with the deeper SolarForce/Ultrafire hosts & their head/bezel assemblies - a unit with zero 'gap' in the SF 6P will get 'end-shake' (& poor electrical contact) in the deeper Solarforce hosts. Reading the above posts, I understand that a full-length sleeve would be too thin to turn down from tube stock. How about a short spacer ring that will fit on the bottom spring shelf of the Solarforce? It could be tall enough to stabilize the quad's side-to-side motion (there is much more delta between the two diameters down there). More importantly, however, a small aluminum ring will prevent the quad from going too deep into the SolarForces, and would be of precise enough of length to result in a near-zero 'quad gap' in the Solarforce hosts and the SolarForce head/bezel assemblies. I think I could figure out some measurements for that simple part.

Life is just so easy once that gnarly non-SF large outer spring is gone, isn't it? :naughty: While the spring in the quad prototype seated pretty well in the SF 6P, it could completely slip off the wider shelf on my SolarForce and slip into the battery chamber. Making it wider would require a wider heatsink recess for it, 'fattening' the bottom of the drop-in and causing it to interfere with the ID just above the SF 6P spring shelf.

OK, I have a plan. I hope to get OAL/bevel location measurements by tomorrow or Friday, and measurements for a Solarforce 'spacer ring base' a few days later.

:huh:

Edit: The cool thing is that I'm pretty well positioned to optimize drop-in fitment using my:

• SureFire aluminum hosts (new and old production)
• SureFire nitrolon host (slightly different interior measurements (i.e. where it really counts, where the P60 sits) than the SureFire 6P)
• FiveMega host (somewhat different measurements than the 6P)
• SolarForce host (considerably different measurements than the 6P)

So hopefully, the best overall fit to all of these different hosts could be achieved. While the Malkoff M60 is a great unit, it provides a small gap in the SF 6P, somewhat less of a gap in the SF G2/G3 nitrolons, considerably more gap in the Fivemegas (yes I've measured & compared that), and zero (actually sort of a 'negative gap' if you think about it) in the deeper Solarforces (resulting in end-shake and poor electrical contact when not using a spacer washer or something), which I have also experienced. I agree with the OP's goal, minimizing the cosmetic gap (even more important when you are 'leading with your chin' i.e. the vulnerable plastic optic on the front of the drop-in which the head/bezel assembly may be referencing off of) while maximizing the variety of hosts that can achieve good fitment & electical contact with this unit - very important and a constant minor headache for us SureFire LEGO-ers. :thumbsup::thumbsup:

The thing is, is that this gap is necessary for the Malkoff drop-ins & the related tolerance stacking in the host/dropin combos, as the Malkoffs are pretty much incompressible. The plastic optic in your quad is noticibly more compressible, this may be enough to minimize any relatively minor resultant gap (due to tolerance considerations) pretty significantly.

OK, I've been working on this post for slightly over an hour, and it's getting late again. 😉
Take care,
K

 

[kodama]

So after following this thread or trying to follow :thinking:, is the problem trying to get a perfect fit in the SF host or make it a universal fit for clones too? I was at a hobby shop the other day and seen some sculpey and thought it might be good for getting dimensions for the drop-in. Figured you could mash it into the drop-in cavity, heat it up a bit in an oven then pop it out to take measurements off of. I've never really worked with sculpey so I don't know at what rates it might expand or contract when heating. Guess you could pour some molten lead in there if you're hard core but prob ruin the finish .
http://en.wikipedia.org/wiki/Sculpey
http://www.sculpey.com/products

 

[Kestrel]

is the problem trying to get a perfect fit in the SF host or make it a universal fit for clones too?

This is the difficult aspect, the OP is trying for both - an optimum fit (mechanically, thermally, and cosmetically) for the popular SureFire 6P, but is still trying to get adequate fit to the other hosts such as the deeper and wider SolarForce/UltraFire hosts. And I'd really like adequate fit to the FiveMega hosts as well - the FiveMegas are some really well-thought out, top-quality bodies that fill very good niches IMO. :huh:

I was at a hobby shop the other day and seen some sculpey and thought it might be good for getting dimensions for the drop-in. Figured you could mash it into the drop-in cavity, heat it up a bit in an oven then pop it out to take measurements off of.

If it was just working the dimensions, I've got that taken care of. Back in the first two pages of this thread, I posted a full suite of precision measurements & partial engineering drawings for my SureFire & SolarForce hosts and also both head/bezel assemblies.

The difficulty here is that the head/bezel assembly makes the situation more complicated - it's not just the dimensions of the host cavity that is of concern. For example, the SureFire Z44 head/bezel is rather shallow compared to the also-popular Solarforce head/bezel assembly, and a drop-in that fits well in a SureFire or FiveMega host may end up being a poor fit (to the point of malfunctioning electrically) if this combo is topped up with the popular SolarForce head/bezel assemblies.

For example, on CPF:

• Most SureFire hosts use the SF Z44 head/bezels, but the inexpensive (& slightly deeper) SolarForce head/bezels are also used on occasion - they often have those nifty stainless steel bezel rings, and they occupy a very good price point.
• Some of the nitrolon SureFire hosts which came with the SF nitrolon head/bezels may be using the aluminum SolarForce head/bezels for better heat handling - those head/bezel assemblies are considerably less expensive upgrades compared to the SureFire Z44's
• Some FiveMega hosts have the SF head/bezels on them, while some have the SolarForce/etc ones. For my FiveMega, it's a toss-up, whichever one I might have handy.
• Most SolarForce hosts have the SolarForce head/bezels, but I doubt that all of them do.

With somewhat different measurements for all the above situations, it's a real headache - that's eight total combinations just up there. If it was just me, I might prefer just saying, 'SureFire only' (& still getting good fitment with the Fivemegas), but even I ended up eventually purchasing a SolarForce host - I really wanted a single-CR123 P60 light for an inexpensive 5mm LED mod, but the short SureFire 3P's are collector items & very expensive, and the FiveMegas, while excellent & worth every penny, are also pretty pricey. For the 'mess-around', secondary-use types of lights, I'd expect that a fair number of CPFers have a difficult time justifying a complete collection of "expensive" hosts (their words, not mine LOL).


Overall, it's a pretty difficult set of goals.

 

[kodama]

This is the difficult aspect, the OP is trying for both - an optimum fit (mechanically, thermally, and cosmetically) for the popular SureFire 6P, but still trying to get adequate fit to the other hosts such as the deeper and wider SolarForce/UltraFire hosts. And I'd really like adequate fit to the FiveMega hosts as well - the FiveMegas are some really well-thought out bodies that fill very good niches IMO. :huh:


The difficulty here is that the head/bezel assembly makes the situation more complicated - it's not just the dimensions of the host cavity that is of concern. For example, the SureFire Z44 head/bezel is rather shallow compared to the also-popular Solarforce head/bezel assembly, and a drop-in that fits well in a SureFire or FiveMega host may end up being a poor fit if this combo is topped up with the popular SolarForce head/bezel assemblies.

Overall, it's a pretty difficult set of goals.

Ahh, I see the dilemma now! I guess other than making different sized drop-ins, is to make a standard drop-in but use different sleeves tailored for different hosts. This would probably cost quite a bit more :shakehead.

 

[HarryN]

Very nice project.

I push things sometimes as well, but I really don't think you will benefit from using pure copper or pure Al vs another alloy.

I agree that the thermal and electrical properties are better in the pure materials, but you will likely be dismayed by the challenge of machining them as accurately as the alloys, which is looking like a big deal in this project.

 

[VanIsleDSM]

I'm not using pure aluminum. Just pure copper.

If you don't think that cutting the thermal conductivity of copper in half by adding a few impurities will make a difference then I'm not sure which forum you have 3000 posts on Harry!

I've already machined pure copper, and it is a *****. But it's also the best material to use for the heatsink, other than pure silver.

Kodama, I'll be making a little tapered piece to fit into the clones to give them basically the same inner profile as a surefire, and then some copper tape around the drop-in and it should fit nicely.

 

[don.gwapo]

Kodama, I'll be making a little tapered piece to fit into the clones to give them basically the same inner profile as a surefire, and then some copper tape around the drop-in and it should fit nicely.

Thanks for that coz I got an L2M for this and were patiently waiting that this drop in will be done soon.

 

[old4570]

Tolerances being what they are ? + - .005 at least and then possibly the odd one that goes beyond that .. Making a perfect fit wont work ..

You need a large sample , and then depending on the variables , make to the smallest size [ ones size fits all ]

Im sure 5 mega might have tighter tolerances than Surefire .. If your going to semi mass produce [ production run ] then you will need to go for a Universal fit [ even for SF ] and that means more measurements .

Good luck with this , its a really nice drop in ..

Solarforce L2 Measuring down to the step ..

0.910inch
0.910inch
0.909inch
0.908inch
0.909inch

 

[VanIsleDSM]

Ok.. I've stayed up way too late figuring this out, but finally after racking my brain and laying awake at night envisioning possibilities, I'm pretty sure I've come up with the ultimate way to make the electrical connection, so it's reliable in any host, provides very low resistance, no need for any gaps, and the drop-in won't be referenced off the bezel in case of a drop and strike on that area that could damage the optic.

I think it's a fairly unique system, it uses multiple springs of sorts. I mocked up a test unit just using spring steel, but I've got some beryllium copper on the way for the final product.

Still waiting for my 6P.. I got the order confirmation a while ago, no shipping notice, but not all places notify you of that. I sent them an email today, hopefully hear back soon. I don't want to play the waiting game any more! Takes so long to get everything together...

 

[h22]

WOW! Verry nice work, add me to the list.

If there is a pre pay or sigh up list get me on it!

Personally, as this is probly the biggest baddest drop in, I wouldn't mind spending extra $$$ on a bored surefire host either.

I'm still waiting on a custom mod from Download, a single SST-50, that is rumored to be over 700lm OTF, but this thing is crazy.

I wouldnt mind the higher current 4, but if the price isn't too bad I may go for both!

Joe

 

[HarryN]

I'm not using pure aluminum. Just pure copper.

If you don't think that cutting the thermal conductivity of copper in half by adding a few impurities will make a difference then I'm not sure which forum you have 3000 posts on Harry!

I've already machined pure copper, and it is a *****. But it's also the best material to use for the heatsink, other than pure silver.

QUOTE]

Hi - There is nothing wrong with reducing the thermal path, its just a matter of understanding the balance. ( and I applaud your push for the limits BTW)

- Inside the XP-G, the thermal path is 6 C / W, so for 3ish watts, nearly 20C
- The interface from the XP-G to the thermal path is probably another 5 C
- The copper heat spreader path is probably 1 - 2 C for pure copper, 3 - 4C for imperfect copper.
- The interface of the drop in to the body is perhaps another 5 C, even with AA and a tight fit.
- Path from head to air + Body - probably 10 - 20 C

Total thermal path (not including flashlight to air and body) - is approx 35 C, so the 1-2 C difference for copper is relatively small by comparison.

Go for it - nothing to loose from doing it !!!

 

[VanIsleDSM]

Ok, an update to where I'm at on this, received my teflon wire today, thanks Mudman!

The 6P should be here wednesday, they only had the floor model left, and they sent me an email asking me if that was ok, but I never received it, so it was never shipped.

So I have time to get the heatsink fitting properly into the genuine 6P, and then the only thing left is the beryllium copper for the spring, I have everything else. The copper spring material should be here next week sometime.

You never know for sure... but I think this next prototype will be the one, and it should be done pretty soon!



Edit: Somehow I didn't see your post at first Harry. I agree using pure copper is definitely not the biggest difference, but it does help, especially multiplied by the power levels being used here. I think your guestimates are probably pretty accurate, except for the LED to thermal path. With solder, a MCPCB, and then some AA to the sink, yeah, probably more than that, but directly soldered to the heatsink I would figure less, around 1C I'd guess. I haven't mentioned yet, that I made a jig to actually clamp the XPGs in place while they are soldered, it's a necessary part of the process for other reasons, but as a nice side effect it will also ensure that the solder layer is as thin as possible, further increasing the efficiency of the thermal path, if only slightly again... they all add up!

 

[Kestrel]

I think this next prototype will be the one, and it should be done pretty soon!

That's cool. I've been way too busy to do any more fiddling with prototype #1 since I last posted, my apologies. I gather that now you're going in a slightly different direction anyhow.
Thanks for all the info BTW, pretty neat to see the process.