What did you mod today?

Yes sir. I don’t like the idea of minus green filters. With a better lens it would be exactly on and below BBL after mod. It’s a neat light but not sure it will replace a modded D4 or FW3A, we’ll see though.

also not a fan of Lee
can dedoming, which seems easier, work the same as slicing?

I expected to see a bit lower duv, given your leds CCT spread

does your fw have aux lights?
is that stuff fun?

I have not been to a disco in years.

It could have been lower with 2700K instead of 3500K at the cost of overall lower CCT also. I wanted to see how the mixing and beam would do with domed and sliced together.

It has a slightly more intense center that fades into a more typical floody TIR beam. The more intense center is very subtle and doesn’t create any rings or noticeable defined throw. More like a floody beam with bias towards the middle I guess?

And the most important thing. Is this boost? Buck ? Quad config voltage?

Not sure. LEDs are parallel and turbo varies some with battery resistance.

I’ve been curious, but I really need a better up-close shot at the switch-mode controller (not that square attiny44) to tell for sure. It’s that chip between the U3 label and the LED- wire.

I think it’s a FET-driven turbo with all other modes being buck (or buck-boost) driven for constant current. I’m positive about the buck (or buck-boost) bit, I’m just not 100% sure on how Turbo is driven. If you check out the runtime graphs in my review at 1lumen.com you’ll see that in the first 30 seconds or so, output kinda sags like what seems typical of FET-driven behavior. It then goes into temperature-controlled ramp down, then in very stable regulated output.

Not bad , similar to Lume1 :+1:

I think you are right, similar to lume1 driver I guess. Unfortunately, I already re-assembled it with gorilla glue since there is no retaining ring :person_facepalming:

In any case if you compare my picture with Kawi’s picture of the Manker MC13 it looks like they re-used the same driver but where that component is that’s missing on his picture is present in my picture, but hidden. I actually considered taking another angle pic but I (wrongly) assumed a similar picture already existed. Given what we know about the lights (MC13, no direct drive mode) I think some type of FET is a good guess for that component.

Yup, good eye. That “missing” component is a FET that would be used in direct on the E14 and not in the MC13.

But the component I was referring to is the rectangle below that. It appears to not be the TPS63020 that’s used by the Lume1. Likely similar, I’m just curious what it is.

M21C-U with sliced 3000k 90+ cri 70.2, OP reflector and glow tail.

Received a GT90 conversion kit from TA.
My GT is now a GT90.

And… My MF04 is MF04GT.

(Click on the image to zoom)

dedomed 219b 3000k 9080 into a Maratac

compared to dome on:

Yesterday I modded D4Ti V1 (I hope that I can still post it in “what did you mod today?” thread) :wink:

It has:

  • Anduril 2,
  • 2A KR4 constant current driver,
  • Nichia Optisolis 3000K.

First step was to limit default 5A current. So I modified ramping table from cfg-noctigon-kr4-nofet.h.
According to test by djozz, I set the max to 410 (410/1024*5000 = 2A). At the ramping level 130 it should make 400 lumens, 150 (turbo) is 500 lumens.

max_pwm = 410 was set in source code of level_calc.py. Command used to generate the table was:

Here is the table:

To make soldering to VR21P4 easier I used little helpers:

Tint / beam comparisions (5000K white balance):

Those pictures are at max regulated levels:

Album with all pictures

Notes:

  • leds don’t light up at all until ramping level 3,
  • D4 Ti V1 can’t accommodate E21A quad MCPCB from intl-outdoor (there are bigger than standard Noctigon 4XP boards),
  • D4 Ti V1 can’t accommodate AUX led board from intl-outdoor,
  • biggest wires that fit through hole in VR21P4 are 24 AWG. 22 AWG from intl-outdoor (and probably other places) don’t fit.

Great mod g_damian. I have 4000k Optisolis on a quad PCB and I wanted to put them in D4 too, but the driver blocked me. Or rather, no current limiting driver.

I have a few questions.
1.Why did you use cfg-noctigon-kr4-nofet.h instead of cfg-emisar-d4v2-nofet.h?
2.Why you have max pwm level 410 while 255 is max?
3.Why do you need these wires for? Why it making soldering easier?
4.If the leds don’t light up below 3, why don’t you delete what’s before? (0,0,1,1,1,2,2)
5.What would the table look like if I wanted to limit the current to 3A?
6.Why use level_calc.py when values in this are for 7135 and fet?

Go for it! :wink:

Because constant current 5 Amps driver for D4 is very similar to KR4 driver. It also uses the same MCU - ATtiny1634. Please check the source code of cfg-noctigon-kr4-nofet.h :

// Noctigon KR4 (fetless) config options for Anduril
// (and Noctigon KR1)
// (and Emisar D4v2 E21A, a.k.a. “D4v2.5”)

255 is a max for ATtiny85. For ATtiny1634 it is 1023. I believe it has something to do with greater resolution of analog to digital converters.

Yes, I used those small wires to lift MCPCB up during soldering to break contact with flashlight head to limit heat transfer.

It is working for SST-20 in my KR4. But does not work for Optisolis. I probably will remove it during next firmware update.

3A means 750mA per each led, so 150 lumens, 600 lumens for 4 leds, max level 3A/5A*1024=614, so

I guess 7135 just represents constant current regulator.

Its' interesting in some of these new drivers I've looked at - they design a limited or regulated output channel to match the 0.35 amps a 7135 provides, just to be compatible with our firmware. It's hardware designed for firmware, not the usual other way around.

Maratac aaa mod to 660nm Red
pic is a link to info about using it to stop migraines

fwiw, the mod results in all 3 modes being the same, 7 lumens on my meter. This is not a problem for use as a Red Light Therapy tool.

atm you can get a Red LED Sofirn C01R with 3 modes. I recommend them for red light therapy if you dont mod your own lights. They have 3 working modes, and are inexpensive.

I know it’s semantics but I wonder if it’s actually even going to the other modes. When I modded a thrunite aaa with the deep red I was sometimes stuck in the same mode but occasionally it would go to other modes too. Makes me think maybe a bleeder resistor or pencil mod might be able to make the mode issues more reliable. However, this is all above my electrical expertise so it’s really just a guess.

I would welcome education on how to make the modes on my AAA lights work better, when modding to Red

previously modded a drop.com aaa LMH Copper Tool,
it now has 3 modes, 5, 10, 12 lumens

this time I modded an MLH Maratac aaa to 660nm Red
it now has 3 identical modes of 7 lumens w AAA eneloop

both of those aaa lumintop drivers are giving unanticipated results.

I have also modded a Thrunite T10T to 660nm. It retained mode spacing and kept 4 distinct modes, albeit at lower level than the stock XP-L

thanks for sharing that the thrunite aaa also does not retain mode spacing, I was wondering about that, and now I dont have to do a mod myself… I have a thrunite aaa titanium. I modded it to XP-E2 530nm Green LED, and it retained 3 distinct modes of 0.1, 8, and 65 lumens on aaa eneloop.

I dont know why green works fine to retain modes, and red does not.

on my meter the green, on second mode reads 8.4 lumens and the red reads 6.5 lumens (with the same partly depleted aaa eneloop @1.26v)

it would be great if we can learn how to make the AAA drivers give 3 modes with a Red LED…

the Sofirn C01R has 3 well spaced modes on Red… they obviously know something I dont (yet)

The big picture explanation is this (XP-E2 datasheet):

Red and Green have very different forward voltages (in fact they are further apart than any other color)

The C01R was designed around the red’s very low voltage requirement.

How each individual driver reacts to this is what is unknown.

Also, there is an element you are mentioning “spacing” which you are using a lux meter that can barely “see” very deep red and I don’t know how well it correlates to your eyes but surely it’s good enough to know if the modes are even changing at all.