Need some modding advice for a DIY multiple XHP70.2 flashlight

Skylight · 2018-12-12T17:11:18.000Z

Some time ago I discovered this really interesting flashlight host at Taobao.

https://item.taobao.com/item.htm?spm=a230r.1.14.196.33bd2789HVQ4Sg&id=573078531014&ns=1&abbucket=19#detail

Normally lights of this size are not offered as hosts. Since they added an extension tube for 8x 18650 I have been thinking about what you could build with such a flashlight. Originally I wanted to use a single XHP70.2 led. But now this seems like a waste of potential to me. A light of this size can handle some more power. :smiling_imp: Basically, there are three options I thought about.
I could put in a 74mm OD 3 XHP70 reflector and use three hard driven XHP70.2 for around 24.000 lumens. Secondly, there is a 67mm OD 4 XHP70 reflector. For this I would need to measure the flashlight head first but with the original reflector being 80mm there would be quite a bit of free space left.
As multiple led reflectors of this sizes are rare, for more leds I need to put something together with single reflectors. I found out that in a circle of 80mm diameter there would fit a 32mm reflector in the center with seven smaller 25mm reflectors around it. I suppose it would look like a DIY DX80 in 4000K. :heart_eyes: With all of those reflectors (all SMO) some sort of spacer needs to be placed below, I might cut a copper rod with the right diameter.

Now my questions are mostly concerning the 8 XHP70 plan as this is a special configuration where you need some expert knowledge if you don’t want to fry or blow up something. :smiling_imp: Can I expect some throw like 800m with the DX80 if I would build a 8 XHP70 setup? Anyone can predict what the beam would look like? Assuming the empty host weighs 1kg and I would add some copper (300 - 500g) inside the head will it be capable of handling the huge amount of head (Imalent DX80: 1204g with battery!)? I welcome all modders and people with theoretical knowledge :student: to leave a comment on this and I’m thankful for all constructive criticism and improvement suggestions.

Lexel · 2018-12-12T17:44:42.000Z

to fill this deep reflector with a copper spacer would likely add like 500g or so

Why not get one of my Buck drivers and run a single XHP-70 on 12V 7A
you outthrow any multiple XHP70 setup by far, having about 7000 Lumens

you cant expect from 3 XHP70.2s 24000 Lumens, maybe with freezing cold light for the first second or 2, with batteries hot out of the charger

AdrianWA · 2018-12-12T18:11:44.000Z

Lexel:

to fill this deep reflector with a copper spacer would likely add like 500g or so Why not get one of my Buck drivers and run a single XHP-70 on 12V 7A you outthrow any multiple XHP70 setup by far, having about 7000 Lumens you cant expect from 3 XHP70.2s 24000 Lumens, maybe with freezing cold light for the first second or 2, with batteries hot out of the charger

How about the run times?

EasyB · 2018-12-12T18:28:29.000Z

With 8 cells and 8 XHP70.2 you can drive them pretty hard if you minimize resistance in the circuit. I would estimate you could get 10A or more to each 6V emitter. This is a huge amount of current and power; multiple FETs are probably necessary to carry the current.

It sounds like you are interested in not just output but also beam intensity, so I would recommend slicing the domes to increase the luminance. At 10A (6V) the effective luminance of a sliced XHP70.2 is about 55 cd/mm^2. If the effective front diameter is 80mm, that’s 5026 mm^2. To account for wasted area between reflectors multiply by a typical fill factor of 0.75 to get 3770 mm^2. Which leads to an estimated throw of 207kcd.

Jerommel · 2018-12-12T18:36:15.000Z

It would be a waste of a nice big reflector i.m.h.o.
And of the head that’s designed around that reflector…
However, with a multi emitter set up you have room for fan cooling under the LED board, which would be wise i.m.o.

Skylight · 2018-12-19T13:31:39.000Z

You are right, Lexel. But I already have got a plan to build a single XHP70 flashlight with a GT reflector for much more throw. So for this flashlight host I thought about lots of lumens with 700 - 1000m of throw. What got me to think about 8 XHP70 instead of three was a little efficiency calculation. With 3 XHP70 you can run them stable at 16A each, 48A in total and 24000 lumens. If you have 8 XHP70 with 48A you get 6A per led and maybe 40000 lumens. That is quite a big difference, maybe I could get 6000 - 8000 lumens continiously after step down.

https://www.mouser.it/ProductDetail/771-PSMN2R0-30YLE115
EasyB, I came across this FET with 272W power dissipation and up to 100A. Maybe I could connect it thermally to the flashlight body and it will handle the high current. Yes, with 10A per led I could get more than 50000 lumens, but I would be satisfied with 40000 already. According to your calculation I could get around 900m throw, right?
Jerommel, I don’t plan to waste this nice big reflector. I’m sure I can find some other use for it. A cooling fan would be a good idea but I would have to drill some holes in the head and use a more complicated setup. An estimated 1500g of thermal mass should probably be enough for passive cooling?!

EasyB · 2018-12-19T16:31:42.000Z

Skylight:

https://www.mouser.it/ProductDetail/771-PSMN2R0-30YLE115
EasyB, I came across this FET with 272W power dissipation and up to 100A. Maybe I could connect it thermally to the flashlight body and it will handle the high current. Yes, with 10A per led I could get more than 50000 lumens, but I would be satisfied with 40000 already. According to your calculation I could get around 900m throw, right?

Yes, if you adequately heatsinked the FET it could handle the current. But that would require some custom engineering solutions because on our common FET drivers the FET is attached to the driver PCB which is not able to transfer a lot of heat to the flashlight.

Yes, I estimated 207 kcd or 910m with 10A to each emitter, with the domes sliced close to the phosphor.

Skylight · 2018-12-29T18:48:53.000Z

Does anyone have some experience with active cooling? Would it make a significant difference in run time with estimated 1500g of mass? Can you give me some recommendations on how it could be build in and if it is complicated? It does not necessarily need to be waterproof.