Sold/Expired FS: empty LED tower for Surefire M3T/M4/M6/KT1/KT2/KT4 head(accept 17mm driver)

[Justin Case]

Unless there is compelling data showing that copper has a real (not just a theoretical) advantage vs aluminum, I would not want to pay the price premium. Of course copper has an advantage in terms of thermal properties for the material. But you have assess performance as a system, and that means introducing real world interfaces and gaps where the tower contacts the TH. I personally would not goop up those interfaces with thermal compound. Very messy and easy to extrude some compound onto the reflector. And there is also the LED/tower interface, which most likely won't be a reflowed contact, but instead will use thermal epoxy.

A long tower that can be shimmed at the flange is an interesting approach. Certainly it is flexible wrt various LEDs.
One size tower for any LED. Just use the right shim. Downside is that the shim is a loose part. A shim glued to the tower stem won't get lost. Downside to a glued shim is the introduction of more thermal resistance at the LED/tower interface.

Note that if cost really is no object, IMO you'd design the tower using the ARCMania/AW design, not the Netkidz. The Shoppe's 14 mm drivers are more efficient than any DX or KD 17 mm driver I've tried, plus The Shoippe offers boost, boost/buck, and buck drivers.

 

[tx101]

Not sure copper would be a good idea especially if this follows the Netkidz design

if you look at a cross section of the Netkidz design, the negative spring sits inside
the Tower. Would the copper wall of the Tower be strong enough to hold the spring
in place without deforming ?

 

[notrefined]

thanks for pointing that out....I've never seen any of these, so I wouldn't have known

 

[precisionworks]

Perhaps the most attractive approach, both from a performance standpoint as well as pricing, is to use copper as the post and aluminum as the main body/spring contact. Doing this would raise the price just a little over an all aluminum tower, as copper in near net size could be used - instead of turning down a huge chunk of copper & putting most of it into the dumpster.

I know machining but am in the dark (pun intended) in this area. What's the consensus on a hybrid copper/aluminum tower?

 

[nbjly]

That sounds perfect.

I'm thinking that would make it easier to have different length posts? Or even one that was extra long and could be finished as needed? Or even one that could be be adjusted for focus via the threads? /Jeff Goldbloom off/

 

[tx101]

Contact springs ..... the negative contact spring on the Netkidz Tower is smaller in diameter than those found
on a typical Chinese-made P60 dropin, 19.10mm vs 20.90mm ( 0.75 inches vs 0.82 inches)
Height is roughly about 12.75mm (0.5 inches)

Positive contact spring, diameter 7mm (0.275 inches), height 11mm (0.43 inches)

These measurements are taken from an already build Tower.

 

[Justin Case]

Will a copper stem offset the potential thermal resistance at the new interface with the base of the tower assembly. I don't know the answer to that. But other variables such a fit of the stem to the reflector opening walls, waste heat generation by the driver might dominate over using copper vs aluminum for the stem.

You probably need to assess the wall thickness of the base to make sure you'll have enough thread engagement with a swappable stem (assuming you use a threaded attachment).

You may want to compare the filament height for a genuine SureFire MN lamp to determine the exact stem height needed for various emitters of interest. Of course, that means you also need to estimate LED die heights for those emitters.

IMO, old emitters like the Seoul P4 and Lux V ought to be ruled out and thus their stem heights also ignored. IMO, the emitters of interest are the XP-G and XM-L, and maybe the Rebel and MC-E. But my guess, with no actual supporting data, is the the XP-G and XM-L will be by far more popular, and I'd probably design for those two for simplicity.

An adjustable stem height tower sounds good in theory except the practical choices of emitter are fairly limited. So the actual benefit of such a tower is minimal IMO. There might be some value in adjustabilty for fine focus adjustments. But my limited tests suggest that good focus is fairly insensitive to hgitting an exact position. As long as you are in a narrow range, focus seems fine and that range seems adquate to deal with LED and TH reflector tolerances.

Perhaps the most attractive approach, both from a performance standpoint as well as pricing, is to use copper as the post and aluminum as the main body/spring contact. Doing this would raise the price just a little over an all aluminum tower, as copper in near net size could be used - instead of turning down a huge chunk of copper & putting most of it into the dumpster.

I know machining but am in the dark (pun intended) in this area. What's the consensus on a hybrid copper/aluminum tower?

 

[nbjly]

The most successful designs I've been seeing are bare LED's reflowed directly to their copper heatsink. The center post has limited mass so it would make sense to produce it from the best material reasonably obtainable. The cavity on my KT head is deep enough to allow the base to be much thicker, about 1/4 inch, that would accommodate the threaded/pressed post. A shorter positive spring would be necessary - anyone know of a source? The ground springs are readily available. The post could have as large a diameter at the base as the raw stock would allow for more surface area, any increase in diameter would be a lot more. Either a press fit with the appropriate goo or fine threads would be the best way the bond the post with the base? A tiny hole in the bottom for air to escape(wires down the sides)?

I agree only two lengths are necessary XP-G and XM-L (until the next best is available!) that seems better than shims.

 

[Justin Case]

The most successful designs I've been seeing are bare LED's reflowed directly to their copper heatsink. The center post has limited mass so it would make sense to produce it from the best material reasonably obtainable. The cavity on my KT head is deep enough to allow the base to be much thicker, about 1/4 inch, that would accommodate the threaded/pressed post. A shorter positive spring would be necessary - anyone know of a source? The ground springs are readily available. The post could have as large a diameter at the base as the raw stock would allow for more surface area, any increase in diameter would be a lot more. Either a press fit with the appropriate goo or fine threads would be the best way the bond the post with the base? A tiny hole in the bottom for air to escape(wires down the sides)?

I agree only two lengths are necessary XP-G and XM-L (until the next best is available!) that seems better than shims.

The issues with designing a tower for reflowing an XP-G or XM-L to the top are that it requires yet another machining step for the top face of the stem, subsequent connection of the hookup wires to the LED will be very challenging, and the stem is now sub-optimal for those who don't want to reflow a tiny XP-G or XM-L to the tower (but would rather glue an LED that is reflowed onto an 8mm board).

Dimension changes to the tower base may need to be checked with compatibility with fitting into the neck of the SF flashlight body.

It's not clear to me that the springs are readily available, at least in the sense of finding them on DX, SB, or KD or harvesting them from a P60 drop-in. The center spring -- maybe. But the outer spring is an open question. The NetkIdz design uses a spring with a different OD vs a P60.

Tower focus height should be the same for an XP-G and XM-L.

 

[nbjly]

I was just saying Copper was better - I'll be using PCB's. If a copper center post only raises the cost somewhat, it should be worth it.
It would be interesting to build a reflowed one to see if it could be run harder. But there's lots of other more efficient platforms to experiment with - I'm just trying to update a couple of favorite lights.

There are outer springs available, but they would require a P60 diameter base, or shoulder. A fat blocky P60 I have fits in everything I have to try here (Surefire), except the LU60 collar, which can be bored.

So just one length/height then? Sounds better. I just didn't want a stack of stuff on top.

Outer springs:

http://flashlightconnection.com/Der...ng-for-Dereelight-P60-style-Drop-Ins-p40.html


So what's everybody wanting to do with them, anyway? I want to put an XR-E R2 on one and an neutral XM-L on another

 

[Justin Case]

It's unclear to me that a copper stem would give any performance gain. Also, if the tower were designed for a more efficient 14mm driver from the Shoppe vs. a cheaper 17mm driver, you'd probably generate about a half watt less waste heat. And you'd have more metal mass left in the base of the tower for heat sinking.

Will Quiles built a tower with a larger diam base that allowed a standard P60 spring to fit over it. He also built an XR-E tower. IMO, an XR-E is more hassle than it's worth. You have to do a lot of fine grinding/filing of the ceramic substrate to get the XR-E to fit into an SF TH opening. That leaves very little of the solder pads left for the hookup wires. The soldering isn't as hard as with something like a bare XP-G, but you also won't have the generous sized solder pads that you'll have with an 8mm board for the XP-G or XM-L.

And for all that effort, Will found that the XR-E hot spot brightness was less than that for a Lux V.

 

[DUQ]

I say we just keep it simple, stick with 6061 and keep it as close to the original design as we know it already works.

Maybe we should look into tweaking the design to accept more popular 14mm circuits. I would rather spend less on the tower and more on a driver.

 

[tx101]

Maybe stick with the 17mm design but include
a sleeve that will allow a 15mm driver to be used

 

[precisionworks]

Either a press fit with the appropriate goo or fine threads would be the best way the bond the post with the base? A tiny hole in the bottom for air to escape(wires down the sides)?

With either an aluminum base/aluminum post or an aluminum base/copper post, the press fit is none too secure & I'd be uncomfortable with it. An easier approach that still allows height adjustment it to cut male threads at the bottom of the post, and tap matching female threads into the base. Once optimum height is found, the post would be locked in place with either a hex nut (easiest way) or a set screw.

The thermal conductivity of copper is almost twice that of aluminum, so it would seem to be beneficial to make the post out of copper.

I say we just keep it simple, stick with 6061 and keep it as close to the original design as we know it already works.

If this were easy, someone would already be making & selling them. Machining a tower is not unlike boring a light - there are different approaches to doing both, and each approach will yield a specific result.

I would rather spend less on the tower and more on a driver.

Any M4 project will not be cheap, no matter how its done.

SF M4 - about $150-$200 USD
bore for 2x18650 - about $60
buy a pair of AW18650's - $40
Plus purchase tower, emitter, driver, etc.

 

[nbjly]

Will Quiles built a tower with a larger diam base that allowed a standard P60 spring to fit over it. He also built an XR-E tower. IMO, an XR-E is more hassle than it's worth. You have to do a lot of fine grinding/filing of the ceramic substrate to get the XR-E to fit into an SF TH opening. That leaves very little of the solder pads left for the hookup wires. The soldering isn't as hard as with something like a bare XP-G, but you also won't have the generous sized solder pads that you'll have with an 8mm board for the XP-G or XM-L.

And for all that effort, Will found that the XR-E hot spot brightness was less than that for a Lux V.

Well that saves a lot of time and effort, thanks - I would like to read about it though (have a link?), was he using a KT head? I guess it just doesn't work with the XR-E.

In light of this video, I doubt that an XM-L is worth it either -

http://www.candlepowerforums.com/vb/showthread.php?309998-XML-Heat-and-lumen-output

Admittedly, the temps are high. But hanging out there on a post in an almost sealed chamber of polished aluminum and glass, I'll bet they get pretty hot.

So I'll use XP-G's, pay extra for Copper, and I'll buy three. And I want to use the higher quality Shoppe drivers, more mass in the base sounds good. That's my vote -

 

[precisionworks]

I don't know if it's practical to open up the bottom of the reflector to allow the use of a larger diameter post - at some point, the decreased reflectance will become an issue. A larger post would conduct away more heat, and it may be worth looking at a copper base as well. With the two piece design, the base can be made from near net size copper rod (very little waste) and the post can be made the same way.

Copper has at least two issues that limit its use - cost & machinability. By using near net size material, cost is kept as low as possible. By choosing free machining copper (C14500 Tellurium Copper, aka C145) machining time is reduced as much as possible. A copper/copper tower will cost more than an aluminum/copper tower, but I'll have to make a prototype of each to determine final cost of the copper/copper version. I'm sure it would have some thermal transfer advantage over the aluminum base.

 

[notrefined]

a copper/copper two piece design for a 14mm driver sounds like it would optimize cost/performance ratio. And from what I could find, it looks like the 14mm drivers offer a max 1A drive current (at least the ones I've managed to find), which would limit heat production as well.

 

[precisionworks]

a copper/copper two piece design for a 14mm driver sounds like it would optimize cost/performance ratio. And from what I could find, it looks like the 14mm drivers offer a max 1A drive current (at least the ones I've managed to find), which would limit heat production as well.

It is interesting that all the lights that have a visible heat sink use copper instead of aluminum. The biggest cost factor in any machined product is not the cost of material but rather it is the machining time. It will not surprise me if c 145 has a cycle time close to 6061 but I cannot say for sure until we run the parts.

Is it possible to make the outside of the base/heat spreader so that it more closely fits the inside of the head? I have not looked at my own m 4 but hope to have some time with in the next few days to work on the design.

 

[notrefined]

we could go one further- since this is an LED unit (and probably fully potted for most of us), shock compensation is all but irrelevant-- would it be advisable to eliminate the outer spring and use the space to increase mass?

 

[Justin Case]

Well that saves a lot of time and effort, thanks - I would like to read about it though (have a link?), was he using a KT head? I guess it just doesn't work with the XR-E.

In light of this video, I doubt that an XM-L is worth it either -

http://www.candlepowerforums.com/vb/showthread.php?309998-XML-Heat-and-lumen-output

Admittedly, the temps are high. But hanging out there on a post in an almost sealed chamber of polished aluminum and glass, I'll bet they get pretty hot.

So I'll use XP-G's, pay extra for Copper, and I'll buy three. And I want to use the higher quality Shoppe drivers, more mass in the base sounds good. That's my vote -

If anything, an XM-L is more efficient than an XP-G, so your conclusion to use the latter is unusual.

And unless you can find a driver to power the XM-L at more than 2A, I doubt a copper stem will give you any gain. I'm unsure if any 14mm drivers from The Shoppe that can do that. Definitely, the SOB is rated only up to 1.5A and I've run them at that level successfully. I've mod'ed the DX3256 for 2A, but efficiency suffers (80% at best).

For some actual thermal data for these sorts of towers, search for my thread on the MC-E KT1 tower..