Need help with making (SF) Tower Module

Hi friends! I'm looking to give new life to my current Surefire light with a KT-4 "Conversion" Head.
These heads were designed for 250/400lumen MN20/MN21 bulbs from Surefire.
There were Tower Modules for sale, but there were almost always a "limited run" and sales went FAST.
At this moment, these are harder to find than a needle in a haystack ! sheeeshhh haha

So I asked a friend if he could help me out and he agreed to make it for me, if I had the right measurements.
I can figure out how much the "can"-part of the module must be, and the widths to fit the light etc,
but can someone help me out with the optimum hight for the emitter-die, measured from the bottom of the tower ?
The tower will slide into the reflector. I just need to know how far into the reflector the led emitter-die
has to sit, for optimal throw.

I've made a sketch (see pic below).

If someone has measurements (CAD-drawings) of such a Tower Module, or can measure these from an actual part,
I would REALLY appreciate this too !!

I thank your for reading and helping out. :thumbsup:

You've posted variations of this question on about 5 other related threads. I answered your question that you posted in DaFABRICATA's thread on his P7 tower.

The tower height depends on what LED you plan to use on top. The whole point of the tower is to get the LED die to the focus point of the reflector. What LED do you plan to use?

I can tell you from personal build experience that this will entail some room for minor changes. Like Justin stated much depends on what LED you intend to use. If it was me I would go big and forget smaller sized LED's. Something like a XML sized. A U3 would be plenty OTF. I built a E2e LED drop in and it was quite the learning lesson. I wanted to use a XPG2 but after dedoming a couple due to heat I ended up having to use an XML. The XPG2 was too small of mass to take the soldering of the 22 ga wires I was using. By the time I could get enough heat on the ceramic back side the dome would just melt off.


With the XML sized LED there was enough mass to solder the wires before the dome melted off. For me the worse part of the build was because of the small size module, the heat build up was horrible. And I used a small regulator. 900 mah at the tail cap! Installed in the light it would get very toasty in less than 10 minutes. Wish I had just gotten one of Tana's conversions. On the other hand I did learn quite a bit from the build. For the tiny relector the E2e has LED placement was extremely critical. Took a bunch of time getting that one part right. Your larger KT4 reflector wont be so finicky.

You might also want to look at this idea since the SF KT4 reflector is not really shaped well for using an LED.

http://www.candlepowerforums.com/vb/showthread.php?204958-Surefire-KT4-head

m4a1usr said:

You might also want to look at this idea since the SF KT4 reflector is not really shaped well for using an LED. http://www.candlepowerforums.com/vb/showthread.php?204958-Surefire-KT4-head

Click to expand...

The SF THs work fine with the right LEDs mounted on a tower module with the correct focus height. I've gotten excellent results using Seoul P4, Cree MC-E, Cree XP-G, and Cree XM-L.

Justin Case said:

The SF THs work fine with the right LEDs mounted on a tower module with the correct focus height. I've gotten excellent results using Seoul P4, Cree MC-E, Cree XP-G, and Cree XM-L.

Click to expand...

Good to know. I dont have any experience with the KT4 myself and was just parroting what I had read. I do know that the stock M3 reflector is lacking when it comes to using a LED drop in. I built a MN10 tower to experiment with where I could move the LED in and out. It was alright. But a poor thower. On the other hand a deeper reflector like Nailbender sells or the 36mm reflector that FastTech offers are much better at getting decent throw. But than again I like minimal spill. Sorry to hijack your thread modder.

m4a1usr said:

I can tell you from personal build experience that this will entail some room for minor changes. Like Justin stated much depends on what LED you intend to use. If it was me I would go big and forget smaller sized LED's. Something like a XML sized. A U3 would be plenty OTF. I built a E2e LED drop in and it was quite the learning lesson. I wanted to use a XPG2 but after dedoming a couple due to heat I ended up having to use an XML. The XPG2 was too small of mass to take the soldering of the 22 ga wires I was using. By the time I could get enough heat on the ceramic back side the dome would just melt off. With the XML sized LED there was enough mass to solder the wires before the dome melted off. For me the worse part of the build was because of the small size module, the heat build up was horrible. And I used a small regulator. 900 mah at the tail cap! Installed in the light it would get very toasty in less than 10 minutes. Wish I had just gotten one of Tana's conversions. On the other hand I did learn quite a bit from the build. For the tiny relector the E2e has LED placement was extremely critical. Took a bunch of time getting that one part right. Your larger KT4 reflector wont be so finicky.

Click to expand...

There is no need for any change to the stock TH if you make an LED tower module correctly and choose a small enough LED. If you go with something like a P7, then you'll have to bore out the TH reflector opening, which IMO is a mistake. The largest LED I've been able to fit into an SF TH is a Cree MC-E, which required carefully grinding off the corners of the LED case.

Since this is a TH upgrade, I would think that throw would be important. I would go with an XP-G2 LED, which will give you just about the same lumens as an XM-L but with a tighter hot spot and thus better throw. Reflow the XP-G2 onto a datiLED MCPCB. Then use thermal epoxy to attach the board to the tower. IIRC, my estimate for the filament height in a stock SF incan tower is about 0.890"-0.900" (call it 0.895"). The uncertainty is due to parallax when I try to use my micrometer to measure the filament height.

So, in theory, build your tower so that the final height of the LED die is about 0.895". For an XP-G2, it looks like the die height from the bottom of the case is 0.65mm to 0.83mm. A datiLED MCPCB is about 0.86mm thick. So that totals to 1.51mm to 1.69mm (0.059" to 0.067"). So that means you need a tower height of 0.828" to 0.836" (call it 0.832")

Interestingly, I've been using old towers originally optimized for the Seoul P4. I estimate that the XP-G LED die height is about 0.910" to 0.915", which is about 0.015" to 0.02" higher than the estimated optimum of 0.895". All I can say is this taller height seems to work. If I slide the tower in and out of the SF TH, the hot spot seems to have tightest focus when the tower is fully inserted. So, you may actually need a tower height closer to ~0.85" (or maybe slightly more). When you build your tower, you can start with the "theoretical" optimum tower height of 0.832". Machine the flange so you have some extra thickness that you can use to adjust the stem height. Start machining off an increment at a time of thickness from the flange to increase the tower (stem) height. Keep going until you find the optimum. Presumably, it will be between ~0.83" and ~0.86".

You could make the tower and base 2 separate components.
The tower would screw into the base. The threading will give you some degree of adjustment with regards to achieving optimum focus

Or just get a Malkoff MD60

That's an intriguing approach. Not sure, however, if there is enough room for a threaded tube within a threaded tube, and space to run the hookup wires from the driver cavity up the tower stem, and to the LED. Also, you don't want the base to start to turn and thread back into the stem when you screw on the TH to the battery tube. Might be avoided by always attaching the TH to an empty battery tube, then dropping in the batts, then screwing on the tailcap. Or making the stem undersized enough that it doesn't contact the TH reflector opening (may give up some heat transfer capability, though).

I'd probably build a tower with a long stem and use a series of known-thickness shims to adjust the focus height and identify the exact tower height needed. Then make a tower of that specific dimension. The long tower can be used to determine optimum focus heights for other LEDs down the line as well. The downside is that you'd have to pop off whatever LED is glued onto the tower currently, and then mount the new LED of interest. Still, there is a decent chance that you can remove the LED without damage and thus re-use it for something else.

Now, if you can figure out how to build a removable, hollow stem that is mechanically secure and concentric, that could be pretty good. An emitter swap would entail merely desoldering the wires from the LED, removing the stem, putting on a new stem (of the appropriate focus height) with the new emitter, and soldering on the wires. Maybe some sort of miniature bayonet mount?

If this can help, according to your first post drawing, i have found with a LA Surefire N1 (KT1 turbo head) 21.5 mm (middle of the filament), and with MN15 (KT4 Surefire) i have found 22mm.

is not the dome of the led creating a virtual focal point higher than the led die?

el_Pablo said:

If this can help, according to your first post drawing, i have found with a LA Surefire N1 (KT1 turbo head) 21.5 mm (middle of the filament), and with MN15 (KT4 Surefire) i have found 22mm.

is not the dome of the led creating a virtual focal point higher than the led die?

Click to expand...

el_Pablo,

Interesting that you measure 21.5mm for an N1. When I set my calipers to 21.5mm, they are clearly short of the filament position. There is no mistaking that the filament is higher than 21.5mm. The difference is greater than any possible error from parallax.

I get about 0.886", or 22.5mm. To minimize parallax error, what I do is look at the top of the black aluminum tower and adjust my eye position until the back edge of the aluminum merges with the front edge. That tells me that I have the tower level with my eye, at least at the position of the top of the tower. Since the filament is very close to the top of the tower, parallax should be small.

I think the main function of the dome is to improve photon extraction efficiency. I don't think it is changing the focal point for the die.