thefreeman’s HDR Anduril 2 high efficiency drivers - update : FWxA boost driver

Thanks for the compliments :blush:

I installed one on them in a PL09 Quad with sliced 519A and AUX LEDs :

I transformed the MCPCB to 2S2P by cutting the trace between D2 and D3 (the line on the right is just some removed soldermask) and by creating another solder pad for the + wire :

I used a flex PCB I made for the FWAA before because I didn’t want to wait for a proper one made for the PL09, hence there are only 3 LEDs :

Though once the optic is on it’s not obvious where is the missing one :

Min level (70uA), could be lower with some tweaks but it’s low enough IMO :

Color measurements on level 7 (0.75A per LED) and turbo (3A per LED):

I’m very satisfied how it turned out, I really like this host and the beam is like pure neutral white for me.

Very well done, that looks amazing!!

Very nice!

Thanks :slight_smile:

Another kind of magic :open_mouth:
When I see those tiny wires, I know magic is happening :wink:
Well done :+1:

It’s not dark magic like Tterev3 and his enamelled wire wizardry, but yeah I like using those wires, they take so little space that they can be comfortably long.

BTW Haukkeli Assembled two BST21-KRx , please check it out. He’s sort of my beta tester and I believe that was his first time assembling a PCB, demonstrating that it’s definitely doable to assemble them with just previous reflowing experience.

It would likely be a bit expensive to be a beta tester across the pond.
I’ll keep hitting a daily F5 on you Oshpark, then…

What’s delaying publication is more the shortage, I had to send a crucial component to Haukkeli so he could build them, there is an alternative but it’s not my preference (the ramp table would need to be changed and I don’t really want to have to manage a lot of firmware builds for everyone), digikey announces availability for early October, Mouser did too then pushed it back to December (and before that it was July)…

The Attiny1616 is also not available from the usual sellers, though LCSC has 6000 in stock, Digikey should get a good quantity in October.

Oh I also would like to find a name for them, but I’m quite uninspired :person_facepalming:

FreeBoost HDR (by thefreeman)™

Not managing that many firmware builds makes perfect sense. And the chip shortage is what it is - I was supposed to get a Fireflies T1R from an OL contest not last year, but the year before. They still don’t have the drivers so I finally had them ship their newish 21700 Quad instead.

TheFreeman- this is such an exciting build for me to see, for many of us FireFlies I'm sure. I have the new PL09"Q" and I love the host size and battery compatibility, but mine has very Hi PD with the aux on high mode (orange), and they're imperceptible on low. Anyway, more importantly, a buck-boost driver we can make soon!

So I have been feeling stale in my current state of modifications (swapping drivers and LEDs and the usual intermediate-level shuffle of mods), and I'm no wizard and have never flashed a firmware on a light, but I am Very much anticipating building some of these drivers of yours, when that is available.

is there anything you could guide me to for a "parts list", I am sure you already mentioned it but it probably went right over my head lol.

Am I correct in understanding that OshPark will have thee available for printing soon? But not right now?thanks again!

this is excitement I haven't had since my first couple of months in the hobby.

I’m glad to see that you both are motivated to assemble some drivers :smiley:


I’ll share OSHpark and Aisler PCBs, the ones I used here are from Aisler (Germany), with part lists and instructions, a ramp table calculator and Anduril source code of course, with a few compiled builds (.hex), although I would encourage people to setup their machine for compiling Anduril and making their own tweaks.

So stoked to read this. Can’t wait to build up a few drivers!

Very keen to try my hand at building this also

Yup, I’ve seen his mods with those wires and my eyes crossed :stuck_out_tongue:

I still like to acknowledge that these “tiny” mods are something spectacular for me, hence my words :wink:

I’ve seen Haukkeli’s work as well and it was well done. Umfortunately I am not that good with reflowing nor with the “tiny” things like components, etc, so I guess I will only be a spectator :smiley:

BTW, I have a TS10 on the way and will hopefully test your programmer :partying_face:

PD is Parasitic Drain, I think

Meet the most powerful and efficient Nimh/li-ion boost driver in existence :

Wait, WHAT????

Yes this seems very cool


Any chance of releasing a BOM and PCB file?

To detail a bit more, the MP3431/32 can boost down to 0.8Vin, but it needs 2.7V for startup, and the Attiny1616 needs 1.8V, I believe that usually in AA drivers the MCU is powered by the always on LED boost converter that outputs a constant voltage below the LED Vf when the light is off, this means of course using a boost converter with low Vin and low voltage startup.
Here instead the necessary voltage for the MP3432 startup and the MCU is provided by another small boost converter.

The advantage of using the MP3432 over the available low Vin boost ICs is that it has much better characteristics, the best low Vin boost IC available is probably the (very costly) LTC3425 used in the Zebralight SC5c, likely the most powerful and efficient AA light on the market, if we compare the internal FETs resistance it’s 40 and 50mΩ for the LTC3425 (equivalent 1 phase, because it has 4) vs 6 and 10mΩ for the MP3432, about 7~5times lower which help achieve much higher efficiency and power, although the limit here is the current deliverable by the AA cell, I haven’t tested yet with an Eneloop, only with my bench PSU. According to HKJ discharge curves, it should still be possible to pull quite a bit of current from it :

With the PSU there is a bit of funkyness at 1Aout, the efficiency is significantly lower and sometimes it would pulls very high current, very strange especially since the inductor saturation is higher than necessary, but I’ll see with an Eneloop.

The other advantage is that it has higher output voltage capability, low Vin boost IC usually have a maximum output voltage of ~5.5V, the MP3432 can output as high as 16V, the limitation is the maximum duty cycle, which should be around 93%, meaning 1/(1-0.93) = 14.2Vout at 1Vin, so it should be able to power any LED from 3V to 12V.

Because I wanted high efficiency for both chemistry I didn’t implement a linear regulation or FET PWM dimming when Vin>Vout, thus the big drawback is that it’s a boost only driver, if li-ion capability is desired then a 6V or higher LED must be used, here it is configured for 3S LEDs for the FWAA (I need to figure out how to modify the 3P original MCPCB). Options for a single LED light would be XHP50.3 HI, B35AM for 6V and XHP35 HI for 12V, which would maybe be more adequate for a small reflector and the usual 3535 MCPCB in 14500/AA lights.

The efficiency would probably break the 90% barrier at 1.2Vin 6Vout, and even better at 3Vout, also I had to use a 2mm tall inductor (XAL6020) for the FWAA shallow cavity, a 3mm inductor would be appreciably better, a XGL6020 would also have been better but they are only sold from Coilcraft direct (high shipping and duty fees), thankfully someone from the US will be able to reship some XGL inductors for me to test.