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

[precisionworks]

I tried to reply to this thread yesterday morning while walking the dogs, but Android speech-to-text conversion is still not quite perfected :nana:

Justin Case, excellent job of analyzing the differences in thermal mass :thumbsup: My gut feeling was that the additional mass would help in heat sinking & heat spreading, but it is reassuring to get mathematical confirmation.

Perhaps you can make the cap undersized enough to account for any reasonable amount of machining tolerance. You'll get a gain in additional thermal mass, but basically no additional gain in contact area (which looks unneeded anyway).

+1

Even though SF holds tighter tolerances than any other light that I've seen, there are variations of a few thousandths of an inch among identical models. I'll try to find a maximum reflector dimension & undersize the tower enough to allow it to fit any reflector in that model.

I work at the CNC shop tomorrow morning, and will try to get the owner to work his magic and develop a baseline program that will cut a working model. What might be called Generation1, Version 1 🙂

 

[notrefined]

very excited to see this moving along! would it make sense to offer a copper plug to accomodate those who only want to use a single driver (to fill the part of the cavity not used by the second board)? I don't know if that would make thermal management better or worse, but might simplify spring lengths.

 

[Justin Case]

For statistics of two samples:

BBNG driver #1, 0.546"-0.550" diam (apparently, not quite round) and 0.183" tall (bottom face of the driver board to top of inductor)

BBNG driver #2, 0.548"-0.562" (max diam at a pair of diametrically opposed ridges that are probably where the driver boards were attached to each other -- it is easy to file down the board) and 0.177" tall

Unfortunately, I don't have a spare SOB buck driver on-hand to measure.

Measured stem height for XM-L focus, based on one of my Netkidz towers, is 0.831" from top face of the base to bottom face of the XM-L glued to the tower. For this particular tower, the XM-L is reflowed onto a datiLED 8mm diam board and the board is glued to a copper shim. Thus, the 0.831" measurement involves the stock Netkidz stem height, a glue layer, the copper shim, and a second glue layer. A bare datiLED board is 0.0343" thick. Add the thickness of the reflowed solder layer and the XM-L die height above its small ceramic substrate to get the total die height above the tower's base.

 

[notrefined]

if sample drivers are needed for prototyping, give me an idea of what and where to look for and I can order a couple

 

[precisionworks]

if sample drivers are needed for prototyping, give me an idea of what and where to look for and I can order a couple

I'm pretty certain that they are, as well as the top of the stem components.

 

[Justin Case]

14mm driver candidates all would probably be the ones from The Sandwich Shoppe: BBNG boost, GD boost/buck, and SOB buck. For a single driver, I've used sense resistors to give an SOB1500. If you want more than that, you'll have to go to a multi-stack, parallel board sandwich.

17mm candidates are probably the usual 7135-based linear regulator boards (say up to 2.8A for a single board, and more for either sandwich boards or for single boards with more 7135 chips piggybacked), and the AX2002-based buck boards (e.g., DX3256, DX26110, KD1640 -- and assuming that DX or KD hasn't slipstreamed any radical changes to these boards).

 

[precisionworks]

14mm driver candidates all would probably be the ones from The Sandwich Shoppe: BBNG boost, GD boost/buck, and SOB buck. For a single driver, I've used sense resistors to give an SOB1500. If you want more than that, you'll have to go to a multi-stack, parallel board sandwich.

17mm candidates are probably the usual 7135-based linear regulator boards (say up to 2.8A for a single board, and more for either sandwich boards or for single boards with more 7135 chips piggybacked), and the AX2002-based buck boards (e.g., DX3256, DX26110, KD1640 -- and assuming that DX or KD hasn't slipstreamed any radical changes to these boards).

Any chance you would have a 2 d drawing with dimensions called out?

 

[Justin Case]

I don't have any drawings of any of these drivers.

 

[notrefined]

is this the website you are referring to?

http://theledguy.chainreactionweb.com/index.php?cPath=48_49_61&&page=2

I'm willing to send one of each of the 14mm boards, if samples of all three are necessary. Which variants would be most useful, or does it matter?

Justin Case, would you be willing to assemble a couple of towers for me when we get the parts made (appropriately compensated, of course)? I'm thinking high CRI XP-G driven at 1-1.5A. To my limited understanding, it looks like a single SOB1000 or SOB1500 would be my best bet, unless two SOB750 would have some advantage.

 

[Justin Case]

It's hard to say exactly as to what is really necessary for the prototype development. In theory, all of the 14mm boards should be using the same MCPCB (different layout of course, but the physical board dimensions should all be the same). And unless the driver cavity depth is really tight, even slight variances in inductor height in theory should not cause any fit issues (certainly not with a single driver -- this would only be a possible problem if you go with a parallel sandwich). What would give you different inductor heights is the use of different inductors for the various boards. I don't know what Wayne uses. Maybe the BBNG, GD, and SOB all use the exact same inductor part.

As for which drivers are the most useful, the SOB is a buck driver and my measurements suggest that you need to give an SOB1500 about 4.2V to reach full regulation. I've tested it out to 15V (16V is the advertised max). So such a tower would run with 2x123A, 3x123A, 4x123A, 2xLi-ion, and 3xLi-ion (or the voltage equivalent in whatever battery chemistry you plan to use).

A BBNG is a boost driver. I've only tested up to a BBNG1000. My guess is that you need at least 2xNiMH and up to 1x123A. I think that with a low Vf emitter like an XM-L, 1xLi-ion will give you a Vbatt significantly higher than Vf and you risk frying the driver. Maybe you could get away with it using something like a small 16340. But something like an AW18650, IMR18650, IMR26500, etc (even 3xNiMH might be too much, especially if they are something like D Accupowers) -- I think the cell will hold its voltage under load for long enough to drive the XM-L at something like at least 3.9V and risk cooking the boost driver.

A GD is a boost-buck driver. Battery configs can include 2xNiMH, 1x123A, and 1xLi-ion.

So if you have a multi Li-ion flashlight, then the SOB is the driver to use. If you have one of FiveMega's 2x26500 SureFire C head bodies that you can load with 2xNiMH C Accupower, then perhaps the BBNG is the right driver. If you want to use 1xLi-ion (e.g., 1x18650, 1xIMR18650, 1xIMR26500), then it looks like the GD is the driver.

I probably could find time to do some assembly work.

 

[notrefined]

I intend to use mine with otherwise stock M6's, and I guess I just assumed anyone who wanted one of these modules would be running 2-4x lithium primaries or 2-3x lithium rechargeables. Maybe I'm off base in that. Barry, do you want samples of all three?

for a given drive current, does using two boards in parallel have any advantage over using one (i.e. SOB1500 vs 2x SOB750)?

 

[tx101]

The whole idea of using 2 x driver boards is so that you can run a XML emitter at 2 - 3A

 

[Justin Case]

Well, one can use these towers with 6P/G2 bodies (stock or bored for 18mm for the 6P) and KT1/2 THs, for example, and thus run them with 1x17670, 1x18650, 2x123A or 2x16340. Or old 9Ps, 12ZMs, M3s, M4s, M500s, and M900s. Or use an FM 1x26500 or 2x26500 body and use IMR26500s or C Accupower NiMH (for the two-cell case). I have a nice 2x26500 FM host running a low power-draw tower, using 2xC Accupowers. I get very long run time with that setup with zero heat issues.

2xSOB750 will cost twice that of an SOB1500. Plus, the effort/cost to wire up the sandwich in parallel. Supposedly, there is a potential "high Z" stability issue for SOB drive currents above 1200mA. The way I understand it, this "stability" issue concerns a lack of full regulation (i.e., you don't get the full 1500mA drive). But according to Wayne, if the driver works, you are good to go. It is not clear to me how much drive current is lost to this stability issue. Even with SOB1000s, I've typically measured about drive currents that are about 4% or 5% lower than nominal, i.e., about 950mA or 960mA instead of 1000mA. For an SOB1500, that might mean about 1425mA to 1440mA. And in fact, in a small sample of about 4-5 SOBs set up for nominal drive currents of 1227mA or higher, that's what I've seen -- about 4% or 5%. I haven't encountered anything that I would have thought was a stability issue. But if this is a concern, you can avoid it by going to a sandwich for higher drive currents. As for efficiency, I'm not sure if there will be a difference between 2xSOB750 vs 1xSOB1500. Could be slightly in favor of the sandwich. I'd have to build one and test it.

 

[dcaprilia]

I'm interested in one if these towers become available.

 

[precisionworks]

I intend to use mine with otherwise stock M6's, and I guess I just assumed anyone who wanted one of these modules would be running 2-4x lithium primaries or 2-3x lithium rechargeables. Maybe I'm off base in that. Barry, do you want samples of all three?

for a given drive current, does using two boards in parallel have any advantage over using one (i.e. SOB1500 vs 2x SOB750)?

I believe samples of all three are necessary, unless Justin Case can come up with a one size fits all solution ... which doesn't sound likely.

My interest in this project is that the M4 seems to be the perfect host for a power LED - combining SF ruggedness with solid state reliability and lower power draw.

 

[Justin Case]

It might be worthwhile to ping Wayne on the Shoppe forum and ask him about the dimensions of the BBNG, GD, and SOB. I would assume that they all use the same MCPCB, just with different layouts.

As I mentioned previously, what definitely varies is location of some of the SMD components on the boards. For example, some of the boards have the inductor closer to the edge than other boards. This can be rectified by de-soldering any inductor right at the edge and moving it in-board. And as I also mentioned, I don't know if Wayne uses different inductors for the different boards, and thus the overall heights might differ (I would think this would be a very slight difference, if any, but it would be good to know).

 

[notrefined]

Barry, one each of BBNG1000, GD1000 and SOB1000 should be on their way to you 😀

 

[precisionworks]

Barry, one each of BBNG1000, GD1000 and SOB1000 should be on their way to you 😀

TYVM !!! 🙂

 

[notrefined]

If (and I'm assuming it will) this project goes well, how do you feel about taking on an E-series tower, since there's no one satisfying that niche anymore? I'm thinking it should be relatively easy, and I can supply samples for exterior dimensions. The sticking point will be that an XP-G just barely fits, so no room for a PCB...assembly will require more advanced skills, as a result.

 

[precisionworks]

how do you feel about taking on an E-series tower, since there's no one satisfying that niche anymore?

I'm open to machining anything (unless it's illegal, immoral or fattening 🙂) as long as there's a market for it.

I can supply samples for exterior dimensions.

Now you're talking.

an XP-G just barely fits, so no room for a PCB...assembly will require more advanced skills, as a result.

Can the reflector opening be enlarged slightly, or is there not enough reflector to work with?

On a side note, if the E-series tower makes it into production, it can be run at whatever drive level is appropriate:

There's a full writeup in my thread on boring services, but essentially this involves boring the tube for almost a size-on-size fit. This particular battery measured .7250" at the largest diameter point (they are close to round, but always have a bump or two that makes the OD larger). I kept enlarging the tube diameter until the battery was a snug slip fit, and ended up at .7265", meaning that the bore is .0015" larger than the battery. This leaves a wall thickness of .013", exactly the same as a commercial 6P tube that I have (.750" bore & .776" OD at the O-ring). I've been carrying & using the light every night, and have shown it to a number of my friends ... so far it is holding up fine.