This driver was pretty much designed around USB functionality, and almost all lights I daily have USB, so I don’t think there will be much for non-USB lights from my side. Some 17/22mm Convoy(ish) driver at most, but those will be a different platform altogether.
For now I only plan to make a Buck version of the same driver.
For lights without USB Anduril makes much more sense than my driver + UI, as it’s essentially crippled without a USB port.
An efficient driver that does not remember about turbo, charging with at least 2 amps, with a backlit key was enough for me, nothing more was needed :). I would have liked to have purchased the other colours if I had known Simon would be interested. I hope you succeed more than you want.
I just realized I don’t have a lot of thin silicone cables, so I ordered a couple meters in 30 AWG (red/green/blue/black/white/yellow) and 22 AWG (red/black).
Not strictly needed to test the protos, but will look a whole lot nicer than all black cables (which is all I have home for now ^^).
I think whether Simon will be interested will depend a lot on the final price per unit I can achieve, and community feedback for this driver. I hope he will like them enough to add them to his shop once they are done.
My ma just signed for the package. I’m very excited, but still have a couple of hours left at the office, and workout in the evening. Hope I’ll manage to devote some time to them on Saturday.
I don’t think Simon will be interested in anything like that at the moment - he’s not posting to the BLF and is also not answering questions in BLF at all (it would also be good to know whether he’s still moving his factory or whether that’s finished, and what the timeline for new products is), and will probably have his hands full trying to get this tariff thing over the line without a hitch. He also showed little to no interest in the newly designed centering rings.
Maybe I am wrong now, but I thinks this is pretty realistic.
I’ll be honest - my euphoria about Simons stuff has completely evaporated at the moment. And additionally, he urgently needs some really modern drivers, the UI and the electronics themselves are simply no longer up to date…
Circuitry and software is easy to make comparing to flashlight host. But now there are CNC online services already. We can make personal BRAND. 
I still have a ~80% done CAD model for a 7x B35AM 3x 21700 sodacan around… I just have too much other stuff on my mind so I never finished it.
Would be interesting to quote as an alu CNC from jlcpcb or similar affordable prototyping service.
Yea, just feed those unfinished CADs to JLCNC to see a price.
And 3 weeks later I see it on AliExpress with some Chinese name 
Likely.
Wow great work!
Recently I’ve received S21E boost and it is amazing light despite poor UI as you mentioned. When you try Anduril there is no way back 
and your idea is great!
You have my instant buy for this driver 
The bad news - I messed up the measurements of the USB port somehow and had to file for half an hour on the inside of the host to fit the driver. I think I still have enough clearance on the PCB to fix this without too much hassle in a future revision.
The good news - it charges with up to 15W as planned, and it shows up as a STM32 DFU device on the computer.
Tomorrow the deep dive into the software world will begin
Add hardware support code only and latter you can adopt some already existing UI code .
Thank you for doing all of this! As a current Electrical Engineering student this is very exciting. I love seeing the cool things people make and learning how they did it.
Awesome! Thank you for making this an open source project. You’re bringing open flashlight hardware out of the stone ages with this one.
What license will you go with?
Do the aux leds have LVP?
I still have to check with the specific chip’s reference manual again, but it should have an analog whathdog that can run while the core is in sleep, and wake it up when a specified range is left. What I don’t know yet is how much current it consumes to keep the ADC running, and if it’s not better to shut the ADC down and have the chip wake up every hour or so on a timer interrupt, check the voltage and go back to sleep.
One of the two I want to implement, will see which is better.
I’m planning to use one of the CERN OHL licenses, they are specifically made for electronics hardware projects. Didn’t really think about which of the three, yet.
Currently my issue is figuring out how to reliably set option bits (comparable to fusebits on AVR) from software. The board does not have the space for an ISP header, so USB programming is the only way. After the first flash, the chip will not enter USB bootloader again unless I set the option bits that enable the BOOT0 pin.
In hindsight I should have bought a devboard with an STLink so I could experiment on a board I can recover… Might use some random other stm32 board I have around for now, and order a NUCLEO-C071RB from mouser.
This is the first time I’m using a chip where I have to access the option bits from code.
EDIT: Just placed an order for the devkit. It should arrive next week. I will read docs and write code until then, and try it once the DK arrives.
Following this with interest, having just built a custom driver of my own. USB flashing/configuration would be really cool.
The board looks great, and it does look very tight. I assume PCBWay didn’t raise any issues with clearances for assembly?
I’ll be interested to hear how you get on with the PWM input. I’ve not tried it, but AIUI the low pass filter makes it hard to fade smoothly between low and high ranges.
One of the things that motivated me was actually getting rid of the half-baked power bank IC in my flashlight, which was intended to power an accessory switch but has a slightly annoying standby current draw. Would be much less of an issue with a 21700 battery compared to 18350, but out of interest, how much current does the power bank IC draw when not in use?
