BLF 348 Output Mod

In my typical usage of this light I found that the standard output is sometimes a little brighter than I would like. I haven’t studied boost drivers much lately so I don’t how to mod them to adjust the output. I DO know what a 7135 is though. With that knowledge I started hunting for a lower output constant current driver. What I found is this . It comes in different outputs. I purchased 15mA, 20mA and 30mA samples. It’s a two-pin chip that is connected in series with the LED. It can go either before, or after, the LED in the circuit.

There’s not much room in the driver cavity but it seems to fit OK. My apologies for the crappy photos. Best I can do with a cellphone. The chip is soldered in so that it partially sits just above the other components on the board. One end is soldered to the board and the other end is connected to the wire going to the LED. If this works out I’ll add epoxy to better hold it in place.

Here’s the beam shot using the 15mA version. It’s a bit too dim for my use. I have no way to measure it but it’s dim enough that I can look into it without discomfort.

Stock 348 on the left. 15mA chip on the right.

Hopefully tomorrow I can mod another with one of the higher output chips and see if they actually working the way I hope they do.

Well I’m going to call this a successful failure so far. The output was definitely reduced, however, there doesn’t seem to have been a corresponding increase in runtime. I turned the light on when I went to bed and 5 hours later when I woke up it was dead. By my napkin math it should have lasted 20 hours or so. Either the chip doesn’t work like I thought it does or there’s a horrible inefficiency in the way I plugged it into the circuit. My naive thinking was that the boost circuit would work to provide just enough voltage for the AL5809 to run the LED at 15mA. Looks like more R&D is in order!

Well, the IC can resist to up 60V to limit to 15mA. Seem that you can bypass the boost driver and use 10440 straight with that current limiter.

Looking at the datasheet, those 2-terminal ICs have a drop-out voltage of 1.75 V.
So that is a no-go for direct connection to a single Li-ion cell.

Just speculating, but you may be putting two current sources in series:
If the boost converter is doing current control (and it should be to some extent), then it would try to boost the voltage to the emitter until the emitter current reaches the designed current. But the 2-terminal IC will try to limit the current to 15 mA, causing the boost converter to go full open and saturate at some high output voltage, lets say 12 V. The LED thus drops ~3 V (45 mW), and the 2-terminal IC drops 9 V (135 mW). See where the power is going?

If possible, try to measure the voltage over the 2-terminal IC to get an idea of how much power it is dissipating.

Any better pics of this driver and the components? I’ve got some AK-007 drivers on order from kd that have the pic mcu on a stacked pcb but it also has a relatively large toroid instead of the smaller inductor of this one and Fastech is out of stock on their 13mm boost drivers. It would be nice to be able to get an Attiny down to AAA size.

Here you go Rbd

post #1041 if that link doesn’t work correctly

Yes, it would :slight_smile: An ATtiny10 should easily fit with a 7135 or a small FET but it’s fairly limited (no eeprom, so no mode memory). I think an ATtiny13a might fit with a dual sided board. I unfortunately have zero board layout skills to do such a thing.

PD68’s 7135 driver
RMM’s FET driver
theres also the BLF tiny10 and wight has a 10mm driver as well

But I think here they/we are trying to have a controllable boost driver.

We already have 10 mm 13A 7135 or fet boards so it’s more a matter of seeing if stacking one on one of these smaller boost drivers works like the ak-007 driver at kd. I asked about the Fastech 13 mm board as it’s a single board design. I’m thinking that with an mmu instead it might be possible to get that layout down to ~10 mm. Just a few things floating around.

It’s a bit of a necro post but I really like how attiny9/10 are available in a 2x2 mm package nowadays. But I don’t understand the remark above.

The MCUs don’t come with EEPROM but as far as I understand they have a single byte of flash for configuration which could be used instead. But maybe I’m wrong about that?…

I took a closer look at the datasheet. The chip does not allow application to write that configuration byte. Still, it should work fine with OTSM but as stated above - it wouldn’t have any memory.