I bought it from Banggood (using a coupon from Martin) less than two months ago, but sadly they appear to be out of stock now.
It looks like the official Lumintop store on AliExpress has them for $19.19, but will supposedly drop to $17.75 in 5 days. That’s about what I paid for mine.
Really great to see you have solved the problem gchart. I ordered some 1.4 boards and parts.
Life got in the way and I finally attempted a v1.0 just today. Used 130K resister and just one LED channel. Worked great on the bench. Seemed to work fine in a Tank007 with a 14500. With a Nimh, the taillights would stay lit for a few seconds and then turn off. The driver seemed to work fine in both cases. I didn't try any workarounds as your low-current booster solution sounds like the best way to go.
Your lights above look great!
Thanks! I’ve been really pleased with the v1.4 boards and how the lights have turned out.
The only strange thing I’ve encountered lately is that mode switching with the Ultratac A1 has issues on NiMH: if the tap isn’t ultra-fast, the light flashes the higher mode and then kicks back to low. My theory is that LVP is recognizing the boosted voltage as a Li-Ion for an instant, then sees the voltage drop down and kicks it back to low. Just a guess. Works perfectly on Li-Ion though (which kinda defeats the purpose of a boost circuit, of course).
As you’ve found out with the Tank007, it seems like each of these dual-voltage drivers have their own nuances. So far, both the Sofirn SF14 and Tool AA have worked great.
I haven’t seen the behavior you’ve described. When that happens, does the main LED illuminate dimly?
^
Not that I can see. I suspect that, on NiMh, the driver is not leaking enough current to run the tail circuit. I bet a bleeder resistor would fix the issue. I might try that since it looks like the v1.0 might work in this light.
I’ve got a new tailcap design with consistent brightness and LVP with minimal additional components. Brief post and video demo about it here, I’ll do a write-up shortly.
you are genius Gabe.
I finally managed to solder the parts for V1.4 from scratch and it work 
That’s great, Nick! Did you use reflow for the boost side, or an iron? I would imagine it’d be extremely difficult with an iron on that WSON package. Using solder paste and a hotplate (or even a skillet) is pretty do-able.
Low-Voltage Protection, Constant Brightness
This board uses a Voltage Reset Monitor to provide LVP and an LDO to provide a constant voltage (an in effect, constant current). Everything is on a single side for quick & easy reflow soldering. It’s dual-channel, but includes a Bridge pad for making it single-channel.
So altogether, there should be a total quiescent current of around 3.3uA while the tailcap is lit, going down to 0.5uA after LVP trips. In comparison to the current for the LEDs (typically 0.15mA = 150uA or higher, this is absolutely minimal)
Standard (16mm): Oshpark
.
Big Switch (19mm): Oshpark (warning - the 19mm doesn’t have reverse polarity protection!)
.
Testing video:
Such exciting stuff going on here. I have all the stuff to build you latest design (before the above), but big lights have been winning out on my attention for some time now. Looking forward to building some of your awesome tail lights eventually. Thank you for sharing your designs and expertise with us. :)
Wow. We are in the age of Constant Current lighted tail caps with boost circuit and LVP. Who woulda thought we’d ever be here? 
Hey Gabe
i reflow with solder paste
With iron it’s impossible to do . I tried couple of ways but it’s way too small the legs end up short circuit so i gave up.
Now im in love with AA flashlights 
thumbs up to you
BTW, I just noticed that they’re back in stock in the Lumintop AE store for $17.51 with free ePacket
Thanks gchart!
I had a bit o’ fun tonight with a new tailcap: my first “smart” tailcap. It’s definitely not a new idea, PD68 was toying with this 3 years ago!
It’s using an ATtiny416 for the smarts side of things, driving 8 individual channels. To keep the power low, it checks the voltage, sets the LEDs, then goes to sleep. For testing, I have it waking up once per second. In a light, I’ll probably have it wake up once an hour or so. Just because there’s an abundance of flash for this use (4K), I added a basic startup animation: spinning circle, looped 3 times. Even with that, it compiles at only like 700 bytes or so.
I could probably do a little adjusting to the voltage on-off levels, but not bad for my first test. By some miracle, this was the first time I ever flashed the 416, and my first time writing code for it. It all came together pretty well.
Here’s the boards on OshPark. This is a small update from what is shown in the video. I’m using a P-FET for reverse polarity protection. I realized that I had two of the legs swapped on my initial design, so I just bypassed the FET pads for testing. This v1.1 has that corrected. Also not that I have not tested this in a complete light that might detract from the functionality. I’ll update this when I complete that test.- Ok, the first in-flashlight test is complete; see post 95.
Edit: BOM
Wow keeeewwww
Can u post the parts to built
I like. It’s fun to think I have a sort of legacy here.
the question would be if its running on a bleeder or needs some intelligent driver that connects battery+ with a FET when off
bleeder and MCU running might get even with 470Ohm too much drop to get a real voltage reading, or even power up as voltage drops
I had same problem with my comparator tail cap and too high current draws in the >1mA range, rund fine on PSU but with bleeder you got no real voltage reading as too high drop on bleeder
Great to see you on, PD!
Oh, and BOM added to the post
Thats a great question. My plan is to have it running on a bleeder, but I have not tested that yet (so keep that in mind before ordering parts, if someone is so inclined). I might have to make tweaks based on how it acts behind a bleeder.
Nice job gchart!
Especially like the decreasing lighted led’s to battery level indicator. 