Need help finding the perfect host for my Seoul P7 emitter...

[Gryloc]

Hello all. I recently purchased a Seoul P7 emitter from a sales thread. I thought that it would be a good opportunity to get my hands on some new technology. I actually did not like the Seoul P7 emitter at first because the LED package was so large, and the four dies were pretty spread out. However, when I saw actual pictures of the emitters from the seller, I noticed that the design changed a bit and the dies were closer together. Actually, I saw that the P7 resembled a Luxeon V a bit when it came to die spacing. I did a few tests with pictures of the Luxeon V emitter, and the Seoul P7 emitter, and I estimated very roughly that there was approximately a 0.17mm gap between each die of the Seoul P7, and approximately 0.12mm gap between each die of the Luxeon V. I always like the beam of the Luxeon V with different reflectors, and I still have a good respect for this monster from the past, despite the older technology. With this goofy looking Seoul P7, I hope to get a similar beam pattern when used with a large reflector, but I expect the brightness to be amazing!


So, I need help finding the ultimate host for this flashlight. I am on a budget, and I really cannot afford premium flashlights or one-off custom designs. I was hoping for something that looks nice (respectable), and not too cheap looking. I hope that it has a big, smooth or light orange peel reflector, or, it can easily handle a McR-38 reflector (which I own). I want to stay away from primary CR123 cells, but the use of 1 or 2 larger li-ion cells (16, 17, 18 millimeter by 50 to 67 millimeters long) would be great. I do not mind if it can handle C or D cells, since I can get cheap li-ion versions of them, or I can make a 6AA pack to fit in a 2D light.

Any previous flashlight that worked as a host for the 'ol Luxeon V should work for this flashlight. I do not mind modifying the reflector to let the ~9mm dome stick through the opening enough to focus properly. I found that the beam of a Lux V and my lonely McR-38 reflector looks nice. I do not mind removing material from this nice reflector. A C/D sized Mag reflector may work, but because the emitter has to stick up pretty far in to the reflector, I am unsure if the big Seoul P7 emitter will fit into the hole of the plastic reflector. Then the power leads have to be considered in to all of this. A host that uses a reflector with a shallow/short focal point will be ideal!

I am pretty open. Just remember that it will have to handle over 10W of heat. The LED has four dies in parallel, so it has a Vf similar to most single die LEDs out there, but it is rated to handle 2.8A of current. I may DD the emitter with one larger li-ion cell, depending on the Vf of the LED, to get over 2A of current. Some Li-ion cells may not be happy with that current, so I may have to buck the voltage of 2+ cells. I do have these circuits from DX that can be hooked in parallel on the output to increase current. I had two in parallel, running from 2 C sized Li-ion cells, that output 2A with no trouble. I may add another to see if I can get 3A.

So, I am naturally thinking of a 2D or 2C as a host (I do like the C-sized form), but I am not sure which other lights are out there that look nice that will meet my needs. I was searching though DX, but those lights look so bulky and awkward, and the reflectors seem cheap. I would like a light that is practical for light use, and something that nice enough so if I would choose to sell this one of a kind flashlight, I would get some interested buyers.

What do you have for me? I want something that can be used to parade around this Seoul P7, but I cannot afford the "Cadillac convertibles" that many others use. Thank you!

Here are some visual aids to get an idea of the emitter's size. The two pictures were taken by HgRyu, and the diagram was from the official datasheet...
(click for larger image)







-Tony

 

[TigerhawkT3]

First: good stuff.

As far as DD off Li-Ion, an ordinary (but good quality, like AW) 18650 should be able to handle 2-3A easily. With Li-Ions, you'll generally want to stay under 2C, so with a 2.2Ah cell, you're looking at 4.4A, and you don't need nearly that much. AW's 18650 protection circuits are short-protected at around 5A. For more on this, read about double-tapping in the Welcome Mat.

Maybe the McR-38 that works well with a LuxV will also work with a P7. I mean, in theory, it's basically just four P4s stuck together, so a Lux/P4-compatible reflector should work, and, given enough size, look pretty reasonable. LuxV reflectors would probably be a good, if uncommon, place to start.

The easiest host I can think of would be a Mag, cut down for a single Li-Ion (18650, C, or D). I can't see any reason why a DHS heatsink from H22A wouldn't work for this. (Sorry for the low coolness factor here, BTW. )

I think that's all I've got for now. Good luck!

 

[Drewfus2101]

I would agree that the best host would be a D size Mag. Its pretty hard to find more size and versatility in a flashlight these days. Kai has some D size protected Li-Ions in stock now. He has them in singles and doubles. You could run a double (7.4V pack) in a 2D Mag with the P7 on any good heatsink and should be just fine.

The fun part will be finding the right driver board. You might have to talk to the guys at TaskLED or Cutter (same guys?) and see if you can work something out. Otherwise I know some people have had success in running multiple drivers wired in parallel to up the output current.

Regardless, this should be interesting.

 

[yellow]

an idea for a smaller light, that still offers enough space to put electronics, single 18650, good sink but only a 27 mm reflector in (wich is the better one imho, because it is deeper and thus grabs more light)
http://www.dealextreme.com/details.dx/sku.9634
will get hot with a real full power running 4-die, but hey: its this small
(reducing current a bit to get more runtime and less heat wont hurt at all)

If that "reduce current"-option is a NO, then imho one of these 2 18650 tri-quad Cree lights from DX/Kadomain are nicer than the Mag-Option. Smaller while offering good space for real sinks and for the 38 reflector


if that Led is 4-die parallel: isnt there that chance to get different Vf-ones, so that one runs on way too much power and thus dies quickly, then the 2nd one, the 3rd, ...
:thinking:
Sorry for that question, I'm not into this Seoul till now

 

[evan9162]

if it were me, I'd use a 3D Mag, some NiMH or NiCD cells, and replace the switch or wire up a low resistance MOSFET in place of the switch. With 3D cells, the voltage difference is just too small to justify a regulator of some kind.

 

[Drewfus2101]

if it were me, I'd use a 3D Mag, some NiMH or NiCD cells, and replace the switch or wire up a low resistance MOSFET in place of the switch. With 3D cells, the voltage difference is just too small to justify a regulator of some kind.

Good point. 3 D size NiMHs fresh off the charger would be real close to 4V with a standing voltage of 3.6V. Thats pretty perfect. And they could handle the current draw no problem.

 

[NetKidz]

I'm planning to put it into a Mag1D with 4x eneloop driving by 8x AMC7135. Maybe also with a D2DIM to do multi-level control.

Eneloop should be able to stay above 1.1v under the load and 8x AMC7135 should be able to deliver about 2.5A.

I think the Aluminum MOP Mag reflector should be fine, but the opening must be enlarged a lot.

 

[LukeA]

Good point. 3 D size NiMHs fresh off the charger would be real close to 4V with a standing voltage of 3.6V. Thats pretty perfect. And they could handle the current draw no problem.

I'd fill a 3D mag with 9 or 12 Eneloops in series running 2 1000mA buckpucks in parallel.

 

[Gryloc]

Thanks for posting your ideas! I want something that is pretty efficient, so I am unsure about using 8 linear AMC drivers. I do like the idea of using large C or D NiMH cells, but I am not sure about one thing. Considering power density and voltage sag under load, which would allow for longer battery life: a C or D NiMH at 2.8A of current draw, or a C or D li-ion cell at ~1.4A of current draw? I would need three NiMH cells, and I always preferred a 2 cell C and D Mag form. The main downfall of using li-ion cells is that if I would sell the light, I would be giving someone 2 big, unprotected li-ion cells. This is so unless I use a D sized li-ion battery pack thats protected, as Drewfus mentioned. Kai also sells the charger for that pack. With 3 lower voltage cells, someone could use Alkaline batteries to some extent if they cannot find a charger for the big NiMH cells.

I need some more guidance in the battery topic due to the reasons of practicality. With 2 Li-ions, I can easily use a buck circuit to get the current that the emitter needed, no matter how the li-ion cell voltage drops. Here is the driver I can use. It is the one that I hooked two in parallel with success and got 2A on the output when using two li-ion cells as the power source:
http://www.dealextreme.com/details.dx/sku.3256

With the NiMH cells, as the voltage drops from use (not sure how long it will take), the emitter will dim quite a bit due to the reduced current (while still being bright). Maybe I should wait for the emitter before choosing which batteries to use. I will hook it up to my CC power supply and check the true Vf with my multimeter. If the Vf is really low (like 3.3V @ 2.8A), then there is no way to keep the current down if the NiMH cells do not experience any voltage sag. On the contrary, if the Vf is really high (4.2V @ 2.8A), then I doubt that the emitter will ever see the desired 2.8A for long, so the cool, one-of-a-kind emitter wont be shown off properly.

Also with the NiMH cells, if the Vf works out perfectly (very slightly overdriving the emitter with fresh cells -like 3.2A- and still powering the emitter above 2A after about an our), then I think that the D2DIM would be perfect (thanks -I forgot about that driver). I would get variable brightness, and everything would be simple and clean in the Mag.

Evan9162, I understand the basics of electronics, but I would not know where to start on a basic mosfet driver circuit. It would be great to reduce the resistance in the switch, but the resistance, especially in the Mag C lights, should not be that high should it? The D2DIM would be very similar if I would wire the Mag switch to be momentary. I have modded a 2C mag to work with George's MaxFlex very nicely (I tweaked it so the button press was buttery smooth and easy to press, but not easy to bump accidentally). If the Vf of the P7 is really low, maybe a bunch of AMC drivers may work (if the driver does not consume too much of the voltage to power itself and contribute to voltage drop).

Yellow, did you just add that link? I was confused by your statement a few days ago because you described a nice sounding host, but you did not give a name or link. I was to busy to ask more. Thank you for that link. I just opened it after typing that above paragraph. That looks like a nice, solid, and simple host. I may order that to experiment with; to see the quality and see how it feels in the hand. I may use that light in the future for other mods, or if I haven't gotten the P7 into anything yet, it may go into that host. The reflector would have to go (maybe), but a McR-27 or McR-27L may be a good match up with the P7. If this host wont work, I will find a good use for it. We will see if someone can machine really quickly a heatsink for 4 Rebel 0100s and 4 McR-12R reflectors (I like the beams with those for their size). Maye someone can shorten the body part (bezel or head with just the reflector?) so those short reflectors will mate up with the Rebels. :naughty:

Oops, I got off track. I wonder how the LOP Mag reflector from Kai would work (if the hole is widened). Does anyone remember how hot the maglite got with an overdriven LuxV on a single emitter heatsink (like the o-sink, hotlips, and H22A's DHS)? Will heat be dissipated as well as a multi-emitter light (using the PTS or PQS heatsink)? I hope heat injected into the Mag battery tube will be transfered to the head well. Thanks all for the help. Great comments and suggestions!

Tony.

 

[yellow]

I think I added the link with editing the post a few mins after posting, but ... :thinking:

unfortunately the machining of this host is not perfect:
the treads are good,
but no knurling on the body and the fins feel "sharp".
imho a perfect light for some 1st mods (cheap and enough room to put whatever in), but for sure not offering that smile a modded 6P or Inova puts into one's face

 

[Drewfus2101]

If you went with a 3D running direct drive it would pull as much as 2.8 amps. The best NiMH batteries you can buy have 12000 mAh capacity. So that works out to around (meaning not accounting for losses) 4.3 hours runtime.

Running Lithium ion batteries it gets more complicated. If you went with 2 D size lithium ions, like the protected ones Kai sells, with a resting voltage of 7.4V and a capacity of 5000 mAh you would also have to use a few of those boards from DX that you linked to to buck the voltage down and (using 2 possibly 3 in parallel, which I've done with those same boards) get the current up. Thats where the math gets REAL tricky.

If we assume those boards are around 80% efficient and do the power in equals power out we can crunch the numbers.

Pin * 80% = Pout
Iin * Vin * 0.8 = Iout * Vout
Iin * 7.4V * 0.8 = 2.8A * 3.6V
Iin = 1.98A
5000 mAh / 1980 mA = 2.52 hours runtime

So you would actually get better runtime on the NiMH batteries than the lithium ion batteries.

*Disclaimer: this was done pretty quickly and could be done wrong. Feel free to correct my math.

 

[Gryloc]

Your math is not wrong. However, the main reason that the NiMH batteries will win is because there will be three of them. Considering just two D sized cells of each type, the Li-ion would win due to their greater energy density (even the cheaper KD cell). However, considering a 80% efficiency of the converter, then the numbers start to get closer between the two. Of course, DD a LED from 2 NiMH cells would lead to very low current going to the emitter (if any) due to the Vf of the emitter.

So, I guess it comes down to using a 2 cell (with li-ion) or 3 cell (with NiMH) Maglite. Can anyone make any suggestions about driver circuitry that can buck 7-9V down to 3-4.5V at currents up to 3A, even when multiple ones are hooked in parrallel? The only one I can think of is the one I mentioned above (DX sku#3256). I am not sure about that driver's efficiency, but it could be found quickly by testing samples, I suppose. I have two drivers, so I will try to test them out later on.

Should I wait to see the Vf of the Seoul P7 before making a decision on the exact driver and setup to use? Well, once it shows up in the mail, I will begin testing it. I will present some form of "Vf vs I" curve here.

Finally, does anyone have any beamshot comparisons of a LuxV behind a Maglite reflector versus some other well known light, or maybe in comparison with a Seoul P4 behind a Maglite reflector? Are KD's aluminum reflectors any better in beam quality, or in efficiency, compared to the plastic stock Mag reflector? What about those premium redesigned ones I heard about sold here on the CPF? Again, will a single emitter heatsink, like the hotlips and DHS transfer 10W of heat effectively? Thanks!

-Tony

 

[Drewfus2101]

The Lux and P4 should have pretty much the same beam shot. They have always been referred to as "direct replacements" for each other, and I have done numerous upgrades of Luxs to SSCs with no change in beam patterns. So that shouldn't be a worry.

I would probably go with an aluminum reflector just because of the heat (along with glass lense). I've used Kai's stippled reflectors with P4s, and I didn't think that they imprvoed the quality of the beam any. I like the stock Mag reflector with P4s. I think you can get a aluminum reflector from Kai without the stippling. My heatsinks come from H22A and seem pretty beefy, but you might be better off to take one to a friend at a machine shop and let him copy it but make it deeper, as in sticks farther down into the Mag body (more aluminum = better heatsinking).

You can use the same equations I used, but solving for the effeciency, to find the effeciency if you can measure the other factors if you have a driver board to test, power supply with voltage and current readout and 2 multimeters. I have that exact setup built but I'm out of those boards. I do have some on the way so if you havn't done that test by the time they get here I'll do it.

 

[Gryloc]

Drewfus,

BTW, interesting name (play with words). Anyways, I was actually hoping for screen shots because the 'ol Luxeon V was an unique emitter. It had 4 dies placed close to each other in series-parallel, and it had a pretty broad beam that sometimes had a donut hole in the center with some reflectors due to the slightly darker area in the center where the 4 die corners meet. It was a unique, but very nice beam. I never seen a LuxV behind a smooth Maglite reflector, and that was why I asked. I wanted to see what I should expect with the quad die Seoul P7. The Seoul P4 is closer to the Luxeon I, Luxeon III, and Luxeon K2 emitters. Due to the single die placed in the focal point, the beam is very sharp and narrow with the Maglite reflector (and with other reflectors, in comparison). Even the Rebel behind the Maglite reflector has the same beam as the other Lux's, K2s, and P4s.

I wanted to see what the aluminum reflectors are like because I thought the LOP would remove some artifacts or maybe clean up the donut hole in the beam, or maybe the aluminum reflectors were slightly optimized to become more efficient. Maybe the aluminized coating may be of better quality, meaning more light reflected possibly. If there are no difference, then I wont bother. Don't forget that the LEDs do not emit heat out the front (like the incandescent does in IR light). If you feel some heat, that is actually because the pure light energy (which is steadily increasing due to the greater LED efficiencies) is being absorbed by your skin, which causes your skin to heat up some. A glass lens would be wise just because the light transmittance (I think that is the term) is much better than that of the polycarbonate lens. I worry that the cheap glass from DX/KD isn't that good (since I can see and detect a slight green tint), and how it compares to more expensive crystal watch lenses and AR coated glass.

So, anyone have beamshots, or know where to send me? That would be helpful. Thanks. It seems like too few still have the old beast in their torches. I always thought that they were a good part of LED history. I always admired them. I need to pick up a more premium binned emitter and just make another LuxV light for old times sake. I always got emitters from Future, so the Luxeon Lottery always provided me with V-binned LuxVs. :green:

-Tony

 

[gswitter]

Your math is not wrong. However, the main reason that the NiMH batteries will win is because there will be three of them. Considering just two D sized cells of each type, the Li-ion would win due to their greater energy density (even the cheaper KD cell). However, considering a 80% efficiency of the converter, then the numbers start to get closer between the two. Of course, DD a LED from 2 NiMH cells would lead to very low current going to the emitter (if any) due to the Vf of the emitter.

So, I guess it comes down to using a 2 cell (with li-ion) or 3 cell (with NiMH) Maglite. Can anyone make any suggestions about driver circuitry that can buck 7-9V down to 3-4.5V at currents up to 3A, even when multiple ones are hooked in parrallel? The only one I can think of is the one I mentioned above (DX sku#3256). I am not sure about that driver's efficiency, but it could be found quickly by testing samples, I suppose. I have two drivers, so I will try to test them out later on.

What about direct driving it with 2x or 3x Li-ion in parallel? Maybe cut down a Mag 2D and dual-bore for 2x 18650 or tri-bore for 3x 17670.

 

[cmacclel]

As far as I know now one knows if any reflectors will work at all with the P7. I have 2 coming in and suggest before you buy anything waite for my input before wasting any of your money.

Mac

 

[d1dd1]

:thumbsup: great to have a professional tester :twothumbs

I will wait for your reviews

Thanks

 

[Nitroz]

As far as I know now one knows if any reflectors will work at all with the P7. I have 2 coming in and suggest before you buy anything waite for my input before wasting any of your money.

Mac

Do you plan to test this in a Mag with stock reflector?

This is the light I would like to build for my father in law to use for work.

 

[Rommul]

I think the possibility of an aspheric lens should be considered.

 

[cmacclel]

Do you plan to test this in a Mag with stock reflector?

This is the light I would like to build for my father in law to use for work.

Yes.

Mac