[[ GXB20 Driver – Homemade Constant Current Programmable XHP50 Single-Cell Boost Driver! ]]

The menus only allow for very basic setting changes and they are not that easy to use. First off you are limited to a pre-made set of mode groups. They don’t always have what you want. Then you have to figure out which group it is you want, count out the blinks and hope you get it right.

By flashing it you could set up the mode group exactly like you want, press a button and boom, you know it is exactly what you wanted with no guessing and annoying counting.

Plus as drivers improve we will be able to select many many more options then we do now and having all of those in the flash menu would be silly, not to mention a lot of space used on the menu alone.

For example in the future I see you being able to set the current you want the LED to see in constant current, the PWM that you want to add or subtract from that, the regulation settings, the thermal regulation points and settings, exact mode groups, custom strobe modes and more I have not even thought of yet.

Put simply is the flash programming a 100% must? No.

Is there any reason not to do it? No. You loose nothing by having flash programming except the programming time. I always come at things from a “Why not?” perspective.

I don’t count out blinks my self, I press the button X times to enter digits. If you can count it’s very easy.

Adjusting for a specific current is probably overkill for most people. I adjust the light to how ever bright I want it in a ramping style setup. Seems a bit annoying to have to grab the phone if one of the modes turns out to be a little brighter or dimmer than you actually wanted.

I do see advantages by wireless upgrade of firmware, but then you still have to flash the basic “operating system” into the MCU anyway, so why not flash it with what you want from the start? Do people really need a whole bunch of mode groups to choose between? If you have a few, pressing the button the amount of times to select it really is easy.

I know, everyone is different and has their own preferences. I ask these questions because as a hobbyist driver/firmware developer I’m always interested in developing new features if I can find personal use for them, but as it is now I’ll guess I’ll just say “why not?” like you and see what others come up with.

What is lost by having the option for flash programming?

It costs 1 extra trace and 1 pin (of which we have lots of extras on new MCU’s) plus some programming time. I just don’t see the downside. Not like people have to use it but knowing general people, they are 10x more likely to actually change a setting on a flashlight if it has a nice GUI interface then counting out blinks or clicks.

Heck even I will hardly even try to mess with the firmware in most lights, just not worth the hassle in most cases.

Plus if you will have to look up the manual to figure out what setting is what and thus be able to program the light anyways (even having used bistro hundreds of time I still don’t have the menu memorized and have to look up the manual before changing anything), I would rather just pick the settings I want while on the computer / phone anyways and then push a button to have them sent to the light. Save the middle step of me having to manually enter it.

Thanks for your answers. I can see some people finding it useful.

So yeah, why not. Go for it.

Has anyone had any luck with obtaining a boost driver to run an XHP50 from a single 18650?

Considering there are 5000K XHP-50 with CRI of 90+ (eg: SC600Fd Mk III Plus) which give better runtimes and output than the low voltage Nichias, one would assume that demand for a driver of this nature would be quite high

Has anyone besides loneoceans actually built one of these?

Ahil, Kaidomain's H1-A driver? Related: Buck and Boost Drivers, Testing, Modding, and Discussion (Pic Heavy)

By the way, where are those high-CRI XHP50s for sale?


But the Nichias have a much better tint, they are not so yellowish…

I have loaded the parts list to my cart twice but have not pulled the trigger. It’s a little pricey. If I knew I could get to making it right away I would still do it but not sure when I would get to it.

I have a couple of empty GXB20 first version boards, but I just couldn’t be bothered to build a board based on the 84A and convert Arduino code :confounded: to C. Now I’ve just taken his schematics and, with some help from Schoki, made a 17mm board with the 1634 instead. Boards are ordered at OSH Park, but would be nice to know if someone else built and tested the circuit before ordering all those components.

Whats the 1634? How does it compare to the 1617?

It’s available to me, supported by AVRDude and has same footprint as the 841. The 1617 has the same 4x4mm package but pins are tighter because there are 4 more. I don’t need 4 more, and the 4x4mm 20 pin package is tight enough.

I wrote a little about it here: Mike C drivers: v8 series, ATtiny1634 based. - #52 by Mike_C

Interested ! Easy

Soo, new 1634 “brains” but the same old 7135 steam engine which stalls at high PWM frequencies. Featuring:

  • PWM. :facepalm:
  • Low efficiency.
  • Early loss of regulation.

This is in a driver thread/discussion about GXB20. Where do you see a 7135 on that? :person_facepalming:

In that link you posted, Mike C: https://budgetlightforum.com/t/-/45057/52

Started to read in there and guess I got carried away.

That looks brilliant, just what I’m looking for…!

Is this driver available for sale already? How can I purchase a couple one 17mm and one 20mm…?

Should bump this with the same question.
Is it (GXB20 V2) available for sale anywhere?

Sadly, no.
There’s a similar driver from MTN but the last time I checked they were out of stock:

Thank you for the point. Agree, it’s sad there is no option to get subj.
I found only H1-A as an alternative (need 20 mm diameter) and it is too tall (7.5 mm).