I wanted to make a wide-spectrum D4. Richard didn’t quite have the tints I wanted in stock though, so I made it a medium-spectrum light instead. I was going for 7A, 5A, 3A, and 1A… ish. Maybe a little variation at each step depending on what I could find. But instead I went with 5D, 3D, 3D, and 2B. And a frosted optic to help blend them.
The result is about 4700K and pinkish, very similar to a Nichia 219B emitter. The tint variation is certainly not as wide as I’d like, but the result looks pretty nice. I may have even over-done the pink a bit, but it’s not very noticeable unless it’s next to a D4-219c with its rather not-pink 5000K tint. Perhaps one of the 3D emitters should be 3C. Or perhaps a wider tint spread would look nicer. But it’s pretty close.
Of course, while it was open I also updated the firmware to the latest Anduril. Compared to stock, it adds a bunch of modes and UI tweaks and makes moon use a fraction as much power.
It’d be nice if they could be purchased like this.
I flashed Anduril to my D4 as well, the green XP-E4 emitters really heat up at full Turbo so it’s nice to be able to limit ceiling on ramping. The green color washes out quite a lot at full power.
I had the D4 and D1 side by side on the shelf, tail standing, doing a green and white lightning display. lol Do you think I could start em simultaneously and have a matching lightning show? The dual random thing got a bit much in the green/white display… hahaha
Emisar D4 Beacon/Indicator/Extreme-Moonlight-Mod
After adding switch LED support to my Q8 momentary firmware I wanted something similar for my Emisar D4. Instead of using real switch LEDs I decided to use the ordinary front LEDs. PWMing them would prevent sleep mode of mcu, so I soldered two 10k SMD resistors in parallel (=5K) and wired them between LED- and PIN 3 of mcu (don’t have 4.7k currently). With low level at PIN 3 the front LEDs are powered with a current of about 360 uA (Nichia 219C).
I made the indicator function configurable with a specific click combination and have 3 options now:
1.) No indicator function at all,
2.) When off LEDs are glowing, when locked out LEDs are off,
3.) When off LEDs are glowing, when locked out there is a short flash each 4 seconds.
Power consumption in lockout mode with flashing LEDs: less than 60 uAh, the light could run for several years in this mode.
Lockout mode may also serve as a long lasting beacon, and off mode as a really low moonlight.
You find the 2x10k resistor stack at the left of the 7135:
Resistor stack fixed and isolated with Arctic Alumina:
I had previously done some work on my Klarus G35 right before I had surgery, removed the flat at the bottom of the reflector and made a Delrin cup to act as centering ring. Yesterday I removed the HD XHP-35 and put an HI E2 3D in it. The overall reflector height was leaving a bit of a shadow in the middle of the hot spot so today I flamed the head until I could get the glue broken loose and unscrewed it from the pill section enough to seat the bezel completely and now the hot spot is clean. I lost a few lumens in the trade but took candela from 223.5Kcd to 281.5Kcd, from 945.52M throw to 1061.13M. All in all it was worth the effort to fine tune it.
It took about an 1/8” backing the head off to allow the bezel to seat properly and attain focus, you can see here that the flat area at the base of the head is now well clear of the switch platform, where it had overlapped slightly in stock set-up…
And the HI emitter, from enough distance to show the reflector getting the square domeless emitter…
Strange you lost lumens, probably a wrong binning like D2 or C4
There are E2-3C which bumped the lumens from the stock D4-1A Emitters
I got those and customers report a bump in throw
The HI emitter is substantially less than an HD emitter, loss of the dome and all that. So one step up in bin is still slightly less output. (D4 1A to E2 3D) Also, a 3D will always make less lumens than a 1A in the same output bin, regardless of the output bin number. The darker phosphor I guess, but it’s always held true in every light I’ve built.
That a cheap luxmeter shows for same real lumen less for neutral white is typically that they are more sensitive for blue than for red
The only high binned XHP35 in E2 were 3C ones that I can confirm with a Cree label seen
worst case
if the D4-1A were on the top of the bin and the E2-3C were on the bottom the lux reading might be a little less, but a drop over 5% is unlikely
That certainly beats the power draw I got from mine, using only firmware mods. I’m getting about 1.7 mA for moon mode, which puts out about 0.11 lm. It should last about 73 days on a 3000 mAh cell, ish. But that still has the MCU running with PWM. Not nearly as nice as 0.36 mA, but still a big improvement over the stock D4’s ~6.4 mA moon mode.
If future Emisar lights still have an unused pin, I should see if Hank might be willing to wire it up for a dedicated moon channel. It might even be able to out-do Zebralight at their own game.
Yeah the better throw was through your modification
The lumen loss is due to the LED are likely a D4 bin with lower blue light reducing the luxmeter reading a bit,
but definitely not a E4, which should have in worst case 7% more in best case 20%
The G35 was touted as having an HI D4 emitter to begin with. I honestly don’t remember doing it, but I must have put an HD emitter in it before I had surgery. The numbers I posted as stock were indeed as it came out of the box, I posted that the day I got it.
Without high end equipment, the vast majority of us really have no way of knowing just what emitter was sent to us by anyone we order from. Most can’t even get a reading. Those of us that can, know full well that our numbers are for comparison to our previous numbers alone, even if our light boxes were built by the same people. It’s never a rock solid carved in stone absolute, but a way to compare before and after when we mod a light.
Finally got around to rebuilding my S42 with a TA/NarsilM driver from Lexel.
Soldered a 10kOhm resistor to the driver for the switch LED (stuck it down with superglue after soldering to stop it being ripped off by the wire moving).
I wanted thermal stepdown, and as this has awful thermal mass, I made a copper heat sink from an 18mm copper bar:
Thermal epoxied to the underside of the shelf:
I then put a generous amount of thermal epoxy on top of the μC, closed everything up and held the driver in place with the battery tube and left it to cure.
After some very fiddly soldering I managed to get the switch board and cover back into place (I didn’t leave a lot of space behind the switch for any excess wire, so had to cut them pretty short):
I know, the icon is upside down, I may correct that at some point, but I’m just happy it works