[ GXB172 - 50W Single Cell 17mm Boost Driver! ]

Schoki has been working on a version of this driver that would use a ATtiny85 MCU and thus allow normal BLF firmwares to be used on it for some time now. He has been super busy recently though so not much progress has been made.

Sooner or later someone will get it to work with BLF firmware. The only real question is how small will it be?

How sweet it be to have this driver running Narsil or anduril, that’s like putting icing on the cake.
Thanks for the info TA. :+1:

the loneoeans firmware is on github on the gxb172 website, so I am thinking if anyone here has skills to modify it to work with narsil or anduril firmware? I will also like to try it with my build to see how we can make it work!

The issue is his driver uses a different MCU then we normally use. So narsil / anduril / bistro would have to be converted to work on a different MCU.

This has been discussed for years but nothing has come of it yet and when it does happen it will most likely be for a more modern MCU option.

Actually TK is working on porting Andruil to Attiny841 and Attiny1634. :+1:

Actively? I heard she was wanting to work on it a long time ago but never saw anything come of it.

If it is happening that is great to hear. They allow for some interesting changed to driver design due to the smaller footprint and extra pins.

Attiny1634 is already working though the code is not published yet. :slight_smile:

Very interesting, I will be curious to see how that progresses.

Is it possible to get a boost driver like this to have an output ~9V to run a triple in series (something like XP-L2)?

It’s my understanding that boost drivers are typically more efficient than linear especially at lower outputs? But I haven’t seen this sort of setup before so I’m guessing there’s a good reason why….can anyone enlighten me on this?

Boost drivers tend to burn up when too much current is passed through it. 50 watts as about the most. So maybe 4 amp total at 9 volts (triple 3 volt leds in series). This is quite low for a triple xp-l2.

You could get much more power wiring the leds in parallel then using a FET driver. That’s what most folks do.

Actually it’s been done. F.w. lux-rc boosts their triples. It’s not any harder than 12V. It just isn’t mainstream.

I think the main reasons are:

  • it costs way more than FET
  • it is hard to do any compact+powerful boost driver, this is not a solved problem yet (and GXB172 is a big step forward)
    • and compact boost driver is still much larger than a compact FET one
    • most good triple hosts don’t have enough space for a bulky boost driver
  • powerful boost drivers are not very efficient at low currents (but still better than linear)

While we are discussing boost drivers, can I digress a little ?
I have two spare (yeah, I know) Olight M2R. They are XHP35 lights, so the driver boosts the input voltage of a single 18650 (4,2V and below) to an output of 12V and regulates the current to allow different modes…
Now if I were to make a triple with one of this, and put 3*3V emitters (say, SST20) in series for a total of 9V required, what would the driver do ?
WIll it produces more current ? Is there a risk to burn it ? Or will the current remain the same and the voltage will burn the LEDs ? or no ?

Without knowing what IC it has and what the setup is, it is impossible to know what the effects will be.

Many will work with the a mildly different voltage setup with minor issues that you may or may not notice (such as higher ripple current then would be ideal).

Some will burn out and some will burn the LED.

Only way to find out is to try it.

Generally if you research the IC you can change a few components to make it work with a 9V setup but that is a lot of work.

On post 84 loneoceans say he configure this driver for 12V operation and this driver can be configure for also 9V use:
[ GXB172 - 50W Single Cell 17mm Boost Driver! ] - #84 by loneoceans It look like the maximum voltage possible is 16V from this boost chip.

Thanks for the replies, makes sense.

The reason I asked is because I’m looking into making my own headlamp design where I’m specifically needing maximum run times at lower outputs ~200-500lumen range to be able to run all night. Hence why I’m interested in the boost drivers efficiency.

On that note, do we have any freelance driver gurus here that would design a custom pcb for a fee?

Im guessing that it may remove some limitations of present flashlight boost drivers by being able to allow for a larger pcb footprint and also being able to mount flush straight onto the alu heatsink/body.

The issue with custom boost drivers is that we do not have a working “BLF Boost reference design”. All the drivers that have been attempted so far have had issues.

Buck driver is also hard on a compact setup using BLF designs, the only working buck driver setup we have is the GT but it takes up a lot of space.

Until we have a working base design to use, it is doubtful that a custom driver could be made.

I am still a huge fan of boost drivers myself, they are far more useful then buck drivers for single cell lights.

Are you really wanting SST-20 or just thought of those as an example? I’m wondering why not just a triple XHP-35 in parallel, since that is the voltage the driver has been optimized for.

the used boost chip needs proper reesistor values for a working voltage window
9V LED on 12V setup wont work

You should be able to get over 200 lumen all night without a boost driver, but I guess that depends on cell choice. Going with a single 18650 cell?

I had issues too so I designed a MP3431GL-P testing board to test some options but I haven’t gotten around to building it yet, too many ongoing projects.

SST20 would have be my first choice…
I can’t stand CREE tints anymore, they can’t even offer a good tint in their “high” CRI variant :frowning:
A quad would be an option but the optic would be partially hidden by the bezel so it’s not a viable solution :frowning: