Any thoughts on fitting a Cree MC-E on an AW TH tower?

Justin Case

Flashlight Enthusiast
Joined
Mar 19, 2008
Messages
3,797
Looks like it might fit, with some judicious filing of the corners of the plastic case. But how would the beam look in a TRTH, SRTH, or KT TH, assuming the tower height was adjusted for proper focus? What converter board, if any, might work, and what wiring scheme that would fit in the TH without having to open up the hole in the reflector?

It seems like the 0.55" boards like a Down Boy or SOB could work, except that the power draw might be excessive (looks like a bit over 5W for a 400ma board, assuming Vf = 3.25V for each die).

Also, wiring the bare emitter looks messy. 4P might be the best in terms of fitting the physical space and avoiding a tangle of spaghetti. Maybe 2S2P.

I have a spare tower. If feasibility looks reasonable, I might give it a shot, but I'd like some guidance from the pros on what components/setup might have the best probability of working well.

Thanks.
 

Justin Case

Flashlight Enthusiast
Joined
Mar 19, 2008
Messages
3,797
For these AW/Arcmania towers for SureFire TurboHeads, there isn't a lot of space for extra wires. So a 4S configuration looks to be difficult.

If 4S could fit, however, does it make sense to use something like a Badboy 400 stepup converter (416ma nominal output)? Typical acceptable battery configurations look like 1xIMR16340, 1x17670, 1x18650, 2x16340, 2x17670, 2x17650, 3x16340, 1x123A, 2x123A, 3x123A, or 4x123A.

At 416ma per die (ignoring any Vf differences), output for a K-bin could be ~425-500 emitter lumens.

A 2S2P configuration looks possible, although two wires look like they will have to cross over each other, which increases the amount of space required. You don't want to have wires rubbing against the reflector opening every time you remove and re-insert the tower, although perhaps that wouldn't be done very often.

Does a Badboy 750 look right for 2S2P? That could allow typical battery configurations of 1xIMR16340, 1x17670, 1x18650, 1x123A, and 2x123A. How about using a BB 1000? That would seem to mean that the BB delivers 6.6W to the MC-E. At about 85% efficiency for the board and 1.5A maximum input current, you'd need at least 5.2V input, but less than ~6.6V. So, the only feasible battery configuration looks like 2x123A.

With a BB 750 and essentially 375ma per die, a K-bin could give ~400-470 emitter lumens. With a BB 1000, maybe ~500-580 emitter lumens.

A 4P configuration looks best in terms of wiring, though maybe not in terms of output and thermal runaway issues. I suppose I could go direct drive, but I'd prefer the flexibility of being able to select from different battery configurations. So, it looks like an SOB 1000 board could work here. I could use 1xLi-ion (would go direct drive at some point), 2xLi-ion, 3xLi-ion, 2x123A, 3x123A, or 4x123A. The downside seems to be that each die would effectively get only 250ma. Is that correct? The result would be about ~280-320 emitter lumens, not much different than for a SSC P4 U2-bin which is a lot easier to mount to one of these towers.

Still unknown is what the MC-E beam might look like in a SureFire TH.
 

Justin Case

Flashlight Enthusiast
Joined
Mar 19, 2008
Messages
3,797
I believe I've found a simple workaround to allow and MC-E wired in 4S, 2S2P, or 4P and mounted on one of AW's turbo towers.

The tower uses a center channel to run the LED+ and LED- wires from the driver board to the emitter. The wires go up this center channel and out a horizontal feed hole near the top of the tower. All you have to do is enlarge this feed hole to accept multiple wires. In fact, I think the hole is already wide enough to take two 24 gauge hookup wires, which is all you need for 2S2P.

So, the connection options can be:

4S: -1 to +2, -2 to +3, -3 to +4; LED+ from the driver to +1; LED- from the driver to -4

2S2P: (+1 and +2) to (-3 and -4). LED+ from the driver to (+3 and +4). LED- from the driver to (-1 and -2). This requires only one "crossover" wire that connects the + and - sides of the MC-E, and you'd run this wire through the horizontal feed hole in the tower.

This is a better wiring configuration, IMO, than the equivalent -1 to +2, -3 to +4, +1 to +3, and -2 to -4; LED+ from the driver to (+1/+3); LED- from the driver to (-2/-4). This has two crossover wires.

4P: Connect +1, +2, +3, and +4; -1, -2, -3, and -4. LED+ from the driver to any of the + terminals. LED- from the driver to any of the - terminals.

For 0.55" drivers, the following limitations seem to apply. Max input current is about 1.5A. These small driver boards should be limited to about 4W maximum power output.

For 4S, you have an effective voltage drop of about 13V (based on a Vf~3.25V @ 416ma). 416ma corresponds, for example, to a Badboy 400. Each die in the MC-E should get 416ma, assuming identical electrical behavior for each die. That gives an estimated output of 425-495 emitter lumens. The power is 5.4W, which looks high, so I don't know if the BB will work well for continuous operation or over the long term.

The board is about 85% efficient. Can we use a single Li-ion to run the flashlight? Let's assume Vbatt=3.7V. Then we get 0.85*3.7*1.5=4.7W. If we assume Vbatt=4.2V, then the max power is 5.36W, which is close. But it is doubtful that a single Li-ion is going to operate at 4.2V for very long before voltage sag under the 1.5A load occurs. So, it doesn't look like a single Li-ion is a good choice for power source.

How about two Li-ions? Ibatt = 5.4/(0.85*7.40)=0.86A. This looks good. You can use any of the 3.7V Li-ion chemistries and be well within the 2C discharge rate limit.

You probably could also use battery configurations like 3xLi-ion, 2x123A, 3x123A, and 4x123A.

For 2S2P, you could use a step-up (e.g., Badboy) or step-down (e.g., SOB) converter. The voltage drop is about 6.4V-6.5V for 2S2P. The drawback with using a step-up converter seems to be that you can't use 2xLi-ion since the source voltage exceeds the voltage drop. Can you use one Li-ion? A Badboy 750 would deliver 375ma per die, assuming each die is identical and assuming standard current divider behavior. That equates to about 405-475 emitter lumens. The power is 4.8W, again perhaps borderline for these 0.55" drivers. A single Li-ion would have to deliver Iin = 4.8/(0.85*3.7) = 1.53A. So it looks feasible, although a bit close. At ~1.5A, you'd have to use an IMR16340 if you want to use that sized Li-ion. Otherwise, a 17670, 18650, or larger cell should work in terms of discharge rate.

How about the voltage at which the driver falls out of regulation?

Vreg=(6.4*0.750)/(0.85*1.5)=3.8V, so it looks like this is also a bit close.

You could also use 2x123A. Let's assume 5V for that battery combo. The 123A cells would have to deliver about 1.1A, which is feasible.

Now, let's look at an SOB step-down converter for 2S2P. I don't think you can use an SOB 1000. That would mean 6.8W of power, and I don't know if the SOB can handle that. An SOB 750 would mean 4.8W, like the BB 750. Same voltage drop and same emitter lumens as well. The difference is that the acceptable battery configurations look better. The SOB max voltage is 16V, so you can run the light using batteries that range from about 6.4V to 16V. That means typical battery configurations could be 2xLi-ion, 3xLi-ion, 3x123A, and 4x123A.

Finally, let's look at 4P. Voltage drop = Vf now. An SOB 1000 means about 3.1W, which looks good. Unfortunately, each MC-E die gets only 250ma, or about 280-320 emitter lumens. May as well just stick with a single-die LED like a Seoul P4. How about an SOB 1416? That's basically 350ma per die, or about 370-430 emitter lumens. To build an SOB 1416, we need R1=0.05 and R2=0.12. Power would be about 4.5W. Board efficiency is 75% to 85%. Input current using one Li-ion would be about 4.5/(0.75*3.7)=1.6A. So trying to use a single Li-ion looks marginal.

Two Li-ions should be fine; basically, input current would be 0.8A. You could run battery configuration from about 4V up to 16V.

The main question now seems to be whether a Badboy 400 works for 4S, a BB 750 or SOB 750 works for 2S2P, or an SOB 1416 works with 4P. Does anyone know the answer to this, empiricially or otherwise?
 
Last edited:

Ctechlite

Enlightened
Joined
Dec 7, 2004
Messages
750
Location
Philadelphia, PA
IMO the SOB 1400 (rounded) would be the best bet with 2S2P configuration. It would just be a question of if the SOB could handle it.

Best bet is to ask Wayne about it. Let us know what he says, if he says anything.

I'd really like to build/try this, but am hesitant without knowing if the SOB can handle the power that it will be passing thru and dissipating.

I think the hardest part with be finding the correct focus. I've tried to look for die height figures but I give up after 15 minutes of searching. If you could find this and compare to the die height for the led that your TM is set for you could work from there. If not it will be trial and error. Thermal paste the MC-E to the tower, solder on some wires for a 4P connection and connect the wires direct drive after inserting in the turbo-head to check focus.
 

nailbender

Flashlight Enthusiast
Joined
Nov 9, 2008
Messages
3,682
Location
North Carolina
Wayne will build you a 1500ma sob there is a place to specify resistors that you want. They will take up to 16 volts of power I believe but no more so three rechargables not four.
The hard part will be wiring the tower module 2s2p but that is just me. I think your direct drive option is a good one.

Good luck on your project I hope all goes well.

Dave
 

Ctechlite

Enlightened
Joined
Dec 7, 2004
Messages
750
Location
Philadelphia, PA
No doubt he will build it. But will it survive for any use?

Direct drive is never an option for me (not anymore anyway!)

My projected configuration is 2s or 3s Li-ion.
 

koala

Flashlight Enthusiast
Joined
Mar 5, 2003
Messages
2,295
Location
Melbourne, Australia
Works very well in a KT1 and M6 reflector head. :) Tower height is different. No donut, just a nice hotspot. :grin2:;):laughing::sssh:
Heats up VERY fast at 2.4A 4P config. :devil: Need to file corners, Very tight fit.
 
Last edited:

Justin Case

Flashlight Enthusiast
Joined
Mar 19, 2008
Messages
3,797
I've built two MC-E towers now. One with an SOB 1000 and the other with an SOB 1200 (actually a 1227). Wired the MC-E in 2S2P.

The trick is to thread the "crossover" wire for 2S2P through the horizontal feed hole near the top of the tower. The hole is just large enough to fit the crossover wire and the wires from the SOB's LED+ and LED-.

I used one of Wayne's 0.030" (0.76mm) thick copper shims to raise up the MC-E for better focus on my Seoul focus tower. You need about 0.94mm total height, and I achieved the remaining height via the Arctic Alumina thermal epoxy layers between the tower/shim and shim/MC-E.

As expected, the fit is very tight. You have to file off the corners of the MC-E's case, as well as fold down the the MC-E's 1+, 1-, 4+, and 4- legs. Soldering has to be tidy (or you can also file off any excess), or else the extra solder exceeds the acceptable lateral size again.

It throws out a lot of light, the beam quality is pretty good, but I'm undecided as to whether it is really "better" than a standard SSC P4 U2-bin tower. The P4 is a better thrower and puts out a decent wall of light in my SF TH.
 

Justin Case

Flashlight Enthusiast
Joined
Mar 19, 2008
Messages
3,797
No doubt he will build it. But will it survive for any use?

Direct drive is never an option for me (not anymore anyway!)

My projected configuration is 2s or 3s Li-ion.

This is what Wayne posted in another thread on the SOB:

"For the SOB. The components can handle currents of 2A.

The SOB max duty cycle is not 100%, but, IIRC 85%??? That means it will not hold regulation down to Vin = Vout. In order to remain in regulation you would need Vin to be higher by 15% or more to be fully regulated."

And:

"The IC is rated for 2A. The IC packaging and as mounted on the SOB combined with a possible issue with high value feedback resistor network has made the current SOB rating a safe 1.2A. Between 1.2A output and 1.5A there is a 50/50 chance it won't work due the high value resistor network interference with proper operation. I have had plans to investigate this and use lower value resistors for the network. This same IC is used in the SPY light and can easily do 1.5A with no issue. The SOB should do 1.5A, but, due to some high Z resistors may cause stability issues above 1.2A."

That's why I went with an SOB 1227 and also potted the driver using Arctic Alumina thermal compound. The TH and the tower itself only get moderately warm after my 15 minute continuous thermal test at room temp.

Also, the MC-E seems perfectly fine using AA epoxy for the tower/shim and shim/MC-E interfaces. I also experimented with AS epoxy for the tower/shim interface, but I don't have any equipment to determine if there was any tangible difference in thermal performance. Certainly, neither tower showed any tint shift. I didn't go with AS epoxy for the shim/MC-E interface because the MC-E terminals are very close to the tower and I was concerned that any excess AS epoxy that got squeezed out would cause a short.

Note Wayne's comment about Vreg being about 15% above Vdrop. Assuming that the MC-E Vdrop is about 6.6V-6.7V at 500ma-600ma drive current per die, Vreg is about 7.6V-7.7V. Thus, it looks like 2xLi-ion is cutting it close. In fact, if I use my OpticsHQ multifunction tailcap on my 12ZM running 2xAW17670 and either SOB1000 or SOB1227/MC-E tower, output is less than if I use the standard SF 12ZM tailcap. Also, the OpticsHQ tailcap's strobe function doesn't work right -- strobe rate is slower than normal. If I switch to 4xSF123A, then the OpticsHQ tailcap functions properly. Seems pretty clear that this electronic tailcap adds just enough resistance to drop the SOB/MC-E tower out of regulation when using 2xLi-ion. When going to the SF123As, there is enough voltage headroom to withstand small added resistances.

So for SF 6P-sized lights, I might tend toward an SOB1000/SSC P4 tower. For 3 cell and greater lights, maybe the SOB(1000,1200)/MC-E is the better choice.
 

Justin Case

Flashlight Enthusiast
Joined
Mar 19, 2008
Messages
3,797
Sorry for the delay in getting photos posted. My camera died and I was waiting for a new one to arrive.

MC-E tower. Note the layer of regular epoxy on the side of the tower. I added that for weaponlight use to help make sure the emitter stays put under recoil. Under the MC-E, you can see some excess AA epoxy that got squeezed out and also the copper shim. I wired the MC-E in 2S2P: 1+/2+ connected to 3-/4- with a "crossover" wire that goes through the feed hole near the top of the tower, 3+/4+ connected to the SOB driver's LED+, and 1-/2- connected to the SOB's LED-. All wire is 24 gauge, stranded Teflon wire. I folded down the MC-E terminals, inserted the stripped ends of the wires in-between the appropriate pairs of MC-E terminals, and soldered the wires to the terminals. If you solder the wires on top of the terminals, you will never get the tower to fit into the TH. As it stands, it is a very tight fit. The white mark in the lower left corner of the MC-E is where the cut corner of the case used to be, before I filed it away. I used White-Out to remind me where 1+ is on the MC-E.

With the MC-E glued to the tower, but not wired up yet, I tested fit into my TH by incrementally filing the corners of the MC-E case. I also slightly bent the terminals downward. When I could see clearance for the corners, I inserted the tower into the TH. The reflector hole automatically "sized" the MC-E terminals for me, bending them down the right amount for the tower to fit into the reflector. I did my preliminary terminal bending so that the final bending didn't require a lot of force. I didn't want to shear off the terminals or have them start to bend sideways instead of downward.
AWtowerMC-Eemitter.jpg


Tower inserted into a T-62 TH:
AWtowerMC-EemitterinT-62TH.jpg


Beam shots are the best I could do with the pocket camera I have. Couldn't capture the minor dark spots in the hot spot (maybe the third beam shot has something there). Inside of about 3 feet, you can see the classic Cree MC-E cross shadow. But beyond 3 feet, the shadow mostly disappears, and just a few very minor dark blobs are slightly noticeable on a white wall.

I measured about 8500 lux at 1 meter for the hot spot and about 350 lux at the halfway point to the edge of the beam for the spill. In comparison, my old SureFire L4's hot spot is only about 500 lux. So the MC-E's spill is more than half as bright as the L4's beam center. And the spill from the T-62 TH is huge.

One of my AW towers with an SSC P4 U2SW0H-bin LED measures about 14,000 lux for the hot spot, but about 100 lux for the spill.

MC-E beam shot 1, against a white painted brick wall, about 6 feet:
MC-Ebeamshot1.jpg


Same shot, EV set to -2:
MC-Ebeamshot2.jpg


Same EV -2 setting, background is a smooth, white wall, again about 6 feet:
MC-Ebeamshotsmoothwall.jpg


About 1 foot from a sheet of paper:
MC-Ebeamshot1foot.jpg
 
Last edited:

Ctechlite

Enlightened
Joined
Dec 7, 2004
Messages
750
Location
Philadelphia, PA
:takeit: Where do I send payment? :naughty:

Thanks for helping me feel better about my parts purchases that I'm about to make (now I know I won't be flushing $100 in parts away).
 

Justin Case

Flashlight Enthusiast
Joined
Mar 19, 2008
Messages
3,797
The main thing is to take your time fitting the MC-E to the TH reflector hole. That means go slow and steady when filing off the corners of the MC-E case, and be very careful when bending the terminals. Whatever angle the 1+, 1-, 4+, and 4- terminals acquire from the "sizing" trick, also angle the other terminals to approximately the same angle.

For your soldering step to connect the wires to the MC-E, 24 gauge fits in-between the MC-E terminals just about exactly, filling the gap. Lightly tin the terminals and the wires. Push a wire into the gap between the appropriate terminal pairs (I'm assuming your going to wire in 2S2P), hold the wire in place, and touch your iron to the work. The solder should flow almost immediately and form a solid connection. If the solder connection is a little "fat", you can file off whatever is necessary for the tower to fit into the TH again. The MC-E fit in a TH is very, very tight, so try to use the minimum amount of solder required for a good joint.

I haven't tried 4P, if that MC-E configuration is of interest to you. If you go that route, you'll probably have to wire a parallel stack of two 14mm driver boards, such as 2xSOB1200 or 2xSOB1400, to get some decent drive current to the MC-E dies. IIRC, the driver cavity for the AW tower can barely accommodate the height of a stack of two SOBs, if the SOBs are touching (top of the inductor for SOB1 pressed against the bottom of the non-component side of SOB2). Also, the driver cavity has a very slight recessed "stop" designed for just one 14mm driver board. Thus, you may have to remove this stop -- Dremel, sanding, machining -- so that the driver stack can be inserted into the driver cavity. Fortunately, the stop is only about an 0.01" wide ledge, so you don't have to take off very much material. And there also seems to be some manufacturing variance in the diameter of the SOB boards, so they might fit without any driver cavity modification.

Good luck.
 

wquiles

Flashaholic
Joined
Jan 10, 2005
Messages
8,459
Location
Texas, USA, Earth
The main thing is to take your time fitting the MC-E to the TH reflector hole. That means go slow and steady when filing off the corners of the MC-E case, and be very careful when bending the terminals. Whatever angle the 1+, 1-, 4+, and 4- terminals acquire from the "sizing" trick, also angle the other terminals to approximately the same angle.

For your soldering step to connect the wires to the MC-E, 24 gauge fits in-between the MC-E terminals just about exactly, filling the gap. Lightly tin the terminals and the wires. Push a wire into the gap between the appropriate terminal pairs (I'm assuming your going to wire in 2S2P), hold the wire in place, and touch your iron to the work. The solder should flow almost immediately and form a solid connection. If the solder connection is a little "fat", you can file off whatever is necessary for the tower to fit into the TH again. The MC-E fit in a TH is very, very tight, so try to use the minimum amount of solder required for a good joint.

I haven't tried 4P, if that MC-E configuration is of interest to you. If you go that route, you'll probably have to wire a parallel stack of two 14mm driver boards, such as 2xSOB1200 or 2xSOB1400, to get some decent drive current to the MC-E dies. IIRC, the driver cavity for the AW tower can barely accommodate the height of a stack of two SOBs, if the SOBs are touching (top of the inductor for SOB1 pressed against the bottom of the non-component side of SOB2). Also, the driver cavity has a very slight recessed "stop" designed for just one 14mm driver board. Thus, you may have to remove this stop -- Dremel, sanding, machining -- so that the driver stack can be inserted into the driver cavity. Fortunately, the stop is only about an 0.01" wide ledge, so you don't have to take off very much material. And there also seems to be some manufacturing variance in the diameter of the SOB boards, so they might fit without any driver cavity modification.

Good luck.

If you go higher current, keep in mind that there is little/poor thermal path when using these modules - the main contact is to the reflector, which in the weapon heads is literally floating with a piece of foam, so the heat never fully reaches the head/body. So basically depending on the head, you might have slightly better thermal path. For higher power you need "solid" contact of the LED module with the head and body, which is something that these KT4 LED Module don't do today :mecry:

Will
 
Top