LED Zeppelin
Flashlight Enthusiast
I just completed three Quad-Seoul Mag 2 C mods using USW0H P4 emitters, Shark converters, external dimming pots, and some custom touches. I documented the builds along the way and here are the results.
These were not easy mods mainly due to the use of Seoul LEDs with the soft fragile dome. Luxeons have durable domes that can be used to support the reflector in compression against the lens - the reflectors do not need to be self-supporting. Tightening the bezel and imparting that same torque on the reflectors will cause them to shear the domes off the Seouls.
Two (pewter) were commissioned by a local friend of mine, and one (HA black) I built for myself. Mine uses McR20S reflectors, and the others use modded IMS 20s.
I began by disassembling the lights and modding the switch housings for clearance with the dimming pots. The switch assembly is removed by pinching out the boot, and sticking a 5/64" hex key down the center of the switch button to loosen the grounding set screw. It will then slide out the back. Inside the tube there is a circlip retaining the front of the switch assembly. The pot housing will sit inside the gap of the circlip, and removing it or pushing one end aside is necessary. In the following picture you can see the switches in various states of modding, and how the pot and switch will reside:
I used Modamag's PES2 heatsinks which have pockets milled for Luxeons, and as is are not suitable for the Seouls. I sent them out to get refaced flat, and had one machined with stepped pockets for the Seoul. The first step locates and supports the LED surround, and the second is for clearance and isolation of the slug. The one with pockets is for use with the McR20S'.
I made a paper template for locating the LEDs when using the IMS20s. Since each reflector has it's own legs, I planned to epoxy the reflector legs to the face of the heatsink individually. Thus the position is not overly critical since I can fiddle with each reflector. But the McR20S do not have legs, so I designed a reflector holder that held and positioned all the reflectors relative to the heatsink by a center screw. The reflector holder plate is supported by a spacer tube, with a SS button head socket screw through the center. Since the reflectors are all held in unison, the LED positions need to be accurate realtive to each other and the heatsink. I tapped the center of this heatsink to accept the screw, and drilled a wire passage off-center (not yet pictured). The first pic shows the patterns and the stock PES2, the second shows the machined sinks - at top are the IMS sinks, and below are the McR20 sink and the reflector holder:
EDIT: I should add that it is possible to use the PES2 heatsinks without machining. By spining the template I was able to find a location where the emitters were mostly in contact with the flat portion of the heatsink. There was one slug that would have overlapped a milled pocket by roughly 20-25%, but filling it with Artic Alumina would probably have been fine.
The IMS reflectors needed to be modded for proper focus with the Seouls. I used a Dremel bit in my drill press, accurately positioned the table, and slid the reflector around by hand. I took off 0.030" from the back, and also shortened the legs so they would touch the heatsink when the reflector contacted the LED surround. This was not 0.030" (forgot the actual dimension) since the surround height of the Seoul is not the same as a Luxeon. I tested the focus by holding one on top of Lambda's Procharger P4. Here are the pics including the beam test (underexposed). The second pic shows the modded (left) and stock IMS:
Next I prepped the pots by cutting off all mounting tabs. I used 2 different styles of mounts, and found the horizontal to be preferrable for my application. The specs are 20K ohm, logarithmic curve, metal bushing, and 9mm (Digikey part # P3R7203-ND). Other than the mounting tabs, the horizontal style has a shaft where the set screw flat is full length. I had some custom low-profile knobs machined and the set screw was lower than the flat of the vertical style pot and didn't line up. I had to machine those shafts and extend the flats.
Another difficulty I had was finding a 9mm X 0.75 tap for mounting the pots. I had online quotes for $90 each! Of course ebay came to the rescue and I found some for $10. Although there are other ways to mount the pots, I elected to tap the correct hole, and epoxy the threads of the pot into the tapped hole. It is permanent. The only way to remove it is to drill it out; don't ask me how I know.
I also fabricated heatsinks for the Sharks from bent sheet aluminum.
I carefully cut the paper templates, double-stick taped them to the heatsinks, and mounted the LEDs. I used the teenie-Kapton-tape-strips method to isolate the slugs on the flat sinks. The tape acts as a mechanical spacer to prevent contact.
The template cutouts for the LEDs are oversized so that I can look down from above and center them by uniform gap. I coated the heatsinks with a thin layer of excess epoxy to isolate the boards.
The on-board pot of the Shark needs to be removed for use with an external one. After several unsuccessful attempts to finesse it off with the soldering iron, I resorted to more drastic measures. I just grabbed it with some fine pliers, twisted back and forth, and disintegrated it off.
EDIT: The onboard pot has three connections, two on one side, and a center opposing. Viewed with the opposing at 12 o'clock, the right and opposing pads connect to two adjacent soldering points. I didn't see one for the left pad so I surface-soldered the lead. I wired my pot to increase in brightness clockwise. You can see my particular wiring in a following pic.
I prewired the Shark, wired it to the switch, and slid/fished the entire assembly down the tube. I secured the switch set screw, soldered the wires to the pot, temporarily connected the LEDs, installed cells, and tested the light before epoxying the Shark and heatsink. The heads were thermal epoxied to the tubes to prevent rotation. I added thermal compound to the PES2 and inner bulkhead before final assembly.
The epoxying of the IMS20s to the heatsink worked like a charm. They are very difficult to remove, but it is possible (don't ask). It would take quite a fall, or several, to dislodge them unintentionally.
Here are pics of the final stages:
With the refaced heatsinks, the dimensions work out so that in both the McR20 and IMS lights, the reflectors sit below the plane of the lens and do not contact it. I used the thick UCL (52.1 X 3mm) from flashlightlens.com which allows the bezel to fully seat and still snug up the lens without any pressure from the reflectors.
EDIT #2: McR20s (for Luxeons) can be used in a stock head without boring, but since the McR20S' now sit lower in the head with the refaced sink and their shorter length, I had the heads bored out slightly. I don't think this is neccessary with the smaller dia IMS20s, but I had all the heads bored anyway. The head wasn't quite bored out enough and I ended up sanding a bit off the lower part of the first McR20S outer step. Originally the reflectors/holder was a press fit, but I found it beneficial to be able to fiddle with each reflector, so I sanded out the holder and made them slip fits. Final securing was done with dabs of epoxy to prevent shifting and rattle. In addition I sanded portions of the OD of the IMS for additional clearance and fiddle room. Be careful if you do this with a belt sander or Dremel. Do it in small stages and keep the IMS cool, otherwise the heat will result in localized clouding of the reflector surface.
Here are pics of the finished lights:
So, where's the beef?
Here are some beamshots with some other interesting comparison lights. All the shots are taken with the same camera settings with white balance set to sunshine, and lights at full brightness. The first shot is taken with the light at 1m from a white wall, F10, 1/10 sec. The second is underexposed, F10, 1/60 sec. The third (outdoors) is at F3.5, 2 sec, and is slightly dimmer than what my eyes saw.
In the third shots outdoors, the driveway circle is about 50' out, the trees at the bend of the drive about 150', and the white post in the distance at center is about 250' away. It was a particularly hazy night, which cut down throw noticeably.
Also noted with the light specs is a Meterman Lux reading taken at 1m. Lux readings are tempermental, so I included a ball-park average of several readings.
First up is the just-built Quad-Seoul IMS20 Shark running 2 X 18650. Lux reading: 9300.
Second is the just-built Quad-Seoul McR20S Shark running 2 X 18650. Lux reading: 9250.
Note how the beam skirt is has petals due to the larger dia McR20S overlapping the bezel more.
Third up is a Quad-Lux (TV1K) Mag 2C running a Fatman w/ ext. pot, 2 X 18650. Lux reading: 5270.
Fourth light is Lambda's Procharger P4, a single Seoul 2D Mag running a 2 X 18650 parallel pack. The pack is not fully charged, I previously measured this light at about 20,000 Lux @ 1 m. Lux reading: 14,100.
Last and definitely not least is the Surefire M6. Mine is running the 500 lumen HOLA and a rechargeable M6-R 7.2V pack. Surefire rates the M6 at mid-run brightness, so it's probably at least 600 lumens pictured. Lux reading: 19,860.
These are my first builds using the Shark and an external pot. I noticed some major differences vs. the Fatman. Most notable was the ability of the Shark to dim to zero output. This can lead to the light being left "on" when it's off. But what is really cool is the ability to adjust the output to such a low level that the LEDs barely glow. It makes a great ultra-low level light. The Fatman starts at about 300 mA or so, not too bright but far from dim.
This feature also led me to swap out some emitters after the lights were completed. I had measured the Vf of all the emitters with my power supply prior to the build. But on ultra-low, some emitters were glowing before others. Swapping out one or two in a couple of the lights resulted in uniform glowing and matched outputs.
Another striking thing about these mods is that they generate some serious heat. Although they never get uncomfortable to hold, they get toasty warm. Considering the massive heatsink, the near ideal path to the head/tube, and the overall mass of aluminum, the Seouls are pumping out some calories. I would still say heatsinking is more than adequate, and the light does not seem to dim after heating up.
Overall these were much more time consuming than I originally thought. But I am very pleased with the results; they came out just the way I hoped. The modded IMS give excellent results, and the ability to epoxy them to the sink without the need for a holder is key. Other than the cool factor of the McR20s and the aesthetics of the holder assembly, you would be hard pressed to notice any difference in use. The McR20S' have a bright center spot that blends to spill more gradually, while the IMS has more contrast throughout the beam.
Special thanks go out to Wayne and Cindy at the Shoppe for virtually all my modding supplies and the awesome Shark converters, Mirageman for the HA black host, Modamag for the PES2 heatsinks which are a bargain at twice the price, Arcmania for the emitters, Fivemaga for the deep tailcaps, and especially my machinist Matt (WS6MRO).
These were not easy mods mainly due to the use of Seoul LEDs with the soft fragile dome. Luxeons have durable domes that can be used to support the reflector in compression against the lens - the reflectors do not need to be self-supporting. Tightening the bezel and imparting that same torque on the reflectors will cause them to shear the domes off the Seouls.
Two (pewter) were commissioned by a local friend of mine, and one (HA black) I built for myself. Mine uses McR20S reflectors, and the others use modded IMS 20s.
I began by disassembling the lights and modding the switch housings for clearance with the dimming pots. The switch assembly is removed by pinching out the boot, and sticking a 5/64" hex key down the center of the switch button to loosen the grounding set screw. It will then slide out the back. Inside the tube there is a circlip retaining the front of the switch assembly. The pot housing will sit inside the gap of the circlip, and removing it or pushing one end aside is necessary. In the following picture you can see the switches in various states of modding, and how the pot and switch will reside:
I used Modamag's PES2 heatsinks which have pockets milled for Luxeons, and as is are not suitable for the Seouls. I sent them out to get refaced flat, and had one machined with stepped pockets for the Seoul. The first step locates and supports the LED surround, and the second is for clearance and isolation of the slug. The one with pockets is for use with the McR20S'.
I made a paper template for locating the LEDs when using the IMS20s. Since each reflector has it's own legs, I planned to epoxy the reflector legs to the face of the heatsink individually. Thus the position is not overly critical since I can fiddle with each reflector. But the McR20S do not have legs, so I designed a reflector holder that held and positioned all the reflectors relative to the heatsink by a center screw. The reflector holder plate is supported by a spacer tube, with a SS button head socket screw through the center. Since the reflectors are all held in unison, the LED positions need to be accurate realtive to each other and the heatsink. I tapped the center of this heatsink to accept the screw, and drilled a wire passage off-center (not yet pictured). The first pic shows the patterns and the stock PES2, the second shows the machined sinks - at top are the IMS sinks, and below are the McR20 sink and the reflector holder:
EDIT: I should add that it is possible to use the PES2 heatsinks without machining. By spining the template I was able to find a location where the emitters were mostly in contact with the flat portion of the heatsink. There was one slug that would have overlapped a milled pocket by roughly 20-25%, but filling it with Artic Alumina would probably have been fine.
The IMS reflectors needed to be modded for proper focus with the Seouls. I used a Dremel bit in my drill press, accurately positioned the table, and slid the reflector around by hand. I took off 0.030" from the back, and also shortened the legs so they would touch the heatsink when the reflector contacted the LED surround. This was not 0.030" (forgot the actual dimension) since the surround height of the Seoul is not the same as a Luxeon. I tested the focus by holding one on top of Lambda's Procharger P4. Here are the pics including the beam test (underexposed). The second pic shows the modded (left) and stock IMS:
Next I prepped the pots by cutting off all mounting tabs. I used 2 different styles of mounts, and found the horizontal to be preferrable for my application. The specs are 20K ohm, logarithmic curve, metal bushing, and 9mm (Digikey part # P3R7203-ND). Other than the mounting tabs, the horizontal style has a shaft where the set screw flat is full length. I had some custom low-profile knobs machined and the set screw was lower than the flat of the vertical style pot and didn't line up. I had to machine those shafts and extend the flats.
Another difficulty I had was finding a 9mm X 0.75 tap for mounting the pots. I had online quotes for $90 each! Of course ebay came to the rescue and I found some for $10. Although there are other ways to mount the pots, I elected to tap the correct hole, and epoxy the threads of the pot into the tapped hole. It is permanent. The only way to remove it is to drill it out; don't ask me how I know.
I also fabricated heatsinks for the Sharks from bent sheet aluminum.
I carefully cut the paper templates, double-stick taped them to the heatsinks, and mounted the LEDs. I used the teenie-Kapton-tape-strips method to isolate the slugs on the flat sinks. The tape acts as a mechanical spacer to prevent contact.
The template cutouts for the LEDs are oversized so that I can look down from above and center them by uniform gap. I coated the heatsinks with a thin layer of excess epoxy to isolate the boards.
The on-board pot of the Shark needs to be removed for use with an external one. After several unsuccessful attempts to finesse it off with the soldering iron, I resorted to more drastic measures. I just grabbed it with some fine pliers, twisted back and forth, and disintegrated it off.
EDIT: The onboard pot has three connections, two on one side, and a center opposing. Viewed with the opposing at 12 o'clock, the right and opposing pads connect to two adjacent soldering points. I didn't see one for the left pad so I surface-soldered the lead. I wired my pot to increase in brightness clockwise. You can see my particular wiring in a following pic.
I prewired the Shark, wired it to the switch, and slid/fished the entire assembly down the tube. I secured the switch set screw, soldered the wires to the pot, temporarily connected the LEDs, installed cells, and tested the light before epoxying the Shark and heatsink. The heads were thermal epoxied to the tubes to prevent rotation. I added thermal compound to the PES2 and inner bulkhead before final assembly.
The epoxying of the IMS20s to the heatsink worked like a charm. They are very difficult to remove, but it is possible (don't ask). It would take quite a fall, or several, to dislodge them unintentionally.
Here are pics of the final stages:
With the refaced heatsinks, the dimensions work out so that in both the McR20 and IMS lights, the reflectors sit below the plane of the lens and do not contact it. I used the thick UCL (52.1 X 3mm) from flashlightlens.com which allows the bezel to fully seat and still snug up the lens without any pressure from the reflectors.
EDIT #2: McR20s (for Luxeons) can be used in a stock head without boring, but since the McR20S' now sit lower in the head with the refaced sink and their shorter length, I had the heads bored out slightly. I don't think this is neccessary with the smaller dia IMS20s, but I had all the heads bored anyway. The head wasn't quite bored out enough and I ended up sanding a bit off the lower part of the first McR20S outer step. Originally the reflectors/holder was a press fit, but I found it beneficial to be able to fiddle with each reflector, so I sanded out the holder and made them slip fits. Final securing was done with dabs of epoxy to prevent shifting and rattle. In addition I sanded portions of the OD of the IMS for additional clearance and fiddle room. Be careful if you do this with a belt sander or Dremel. Do it in small stages and keep the IMS cool, otherwise the heat will result in localized clouding of the reflector surface.
Here are pics of the finished lights:
So, where's the beef?
Here are some beamshots with some other interesting comparison lights. All the shots are taken with the same camera settings with white balance set to sunshine, and lights at full brightness. The first shot is taken with the light at 1m from a white wall, F10, 1/10 sec. The second is underexposed, F10, 1/60 sec. The third (outdoors) is at F3.5, 2 sec, and is slightly dimmer than what my eyes saw.
In the third shots outdoors, the driveway circle is about 50' out, the trees at the bend of the drive about 150', and the white post in the distance at center is about 250' away. It was a particularly hazy night, which cut down throw noticeably.
Also noted with the light specs is a Meterman Lux reading taken at 1m. Lux readings are tempermental, so I included a ball-park average of several readings.
First up is the just-built Quad-Seoul IMS20 Shark running 2 X 18650. Lux reading: 9300.
Second is the just-built Quad-Seoul McR20S Shark running 2 X 18650. Lux reading: 9250.
Note how the beam skirt is has petals due to the larger dia McR20S overlapping the bezel more.
Third up is a Quad-Lux (TV1K) Mag 2C running a Fatman w/ ext. pot, 2 X 18650. Lux reading: 5270.
Fourth light is Lambda's Procharger P4, a single Seoul 2D Mag running a 2 X 18650 parallel pack. The pack is not fully charged, I previously measured this light at about 20,000 Lux @ 1 m. Lux reading: 14,100.
Last and definitely not least is the Surefire M6. Mine is running the 500 lumen HOLA and a rechargeable M6-R 7.2V pack. Surefire rates the M6 at mid-run brightness, so it's probably at least 600 lumens pictured. Lux reading: 19,860.
These are my first builds using the Shark and an external pot. I noticed some major differences vs. the Fatman. Most notable was the ability of the Shark to dim to zero output. This can lead to the light being left "on" when it's off. But what is really cool is the ability to adjust the output to such a low level that the LEDs barely glow. It makes a great ultra-low level light. The Fatman starts at about 300 mA or so, not too bright but far from dim.
This feature also led me to swap out some emitters after the lights were completed. I had measured the Vf of all the emitters with my power supply prior to the build. But on ultra-low, some emitters were glowing before others. Swapping out one or two in a couple of the lights resulted in uniform glowing and matched outputs.
Another striking thing about these mods is that they generate some serious heat. Although they never get uncomfortable to hold, they get toasty warm. Considering the massive heatsink, the near ideal path to the head/tube, and the overall mass of aluminum, the Seouls are pumping out some calories. I would still say heatsinking is more than adequate, and the light does not seem to dim after heating up.
Overall these were much more time consuming than I originally thought. But I am very pleased with the results; they came out just the way I hoped. The modded IMS give excellent results, and the ability to epoxy them to the sink without the need for a holder is key. Other than the cool factor of the McR20s and the aesthetics of the holder assembly, you would be hard pressed to notice any difference in use. The McR20S' have a bright center spot that blends to spill more gradually, while the IMS has more contrast throughout the beam.
Special thanks go out to Wayne and Cindy at the Shoppe for virtually all my modding supplies and the awesome Shark converters, Mirageman for the HA black host, Modamag for the PES2 heatsinks which are a bargain at twice the price, Arcmania for the emitters, Fivemaga for the deep tailcaps, and especially my machinist Matt (WS6MRO).
Last edited:
.
... :twothumbs ... :wave:
