I have quite a few Fet+1 drivers with the BLF A6 UI. I like the ability to go forward or backwards through the levels. However, on some of my lights the Fet is a little overkill. E.g., 3v XHP50.1 in an S2+ @19A.
What would be ideal is a single sided 4x7135 for my Convoy S2’s and S2+’s. I can always piggyback 7135 for more output.
Plenty, my favorite is the HQ10D but the moonlight special is the same thing (just 7+1 instead of 9+1).
HQ10D looks like a DIY build, or does someone sell them with reasonable UK delivery?
Dude, if you can stack 7135’s you can populate a board!
Having said that, lexel might build you one, he ships from Germany.
No problem building the boards. Programming the ATTiny (I presume) may require buying in a clip. The post about this says its based on the Nanjg, so could I buy them and re-flash them? Or does someone already sell a version re-flashed?
You can’t just reflash nanjg’s, they’re only single channel (all 7135’s driven in parallel, there is no “+1”)
OK, so I took the plunge and ordered some HQ10D boards from OSH Park.
I built a test board with the following
MCU: Attiny13a
C1: 10uF
OTC: 1uF
2x 7135 Regulators (350mA versions), just for testing, one connected to pin 5, the other to pin 6 of the Attiny.
R1: 19.1k
R2: 4.7k
D1: BAT54JFILM Schottky Diode
What I’ve learnt today:
You can’t program the Attiny on the board. On the USBAVR there should be a link for slow MCU, which the Attiny13 needed (had to solder one to my USBAVR). There is more than one driver for the USBAVR. Eventually, after much Googling, downloading Zadig and selecting libusbK worked. The wiring of my programming clip WAS correct the first time, but I measured it end to end several times and rewired it twice. I need glasses.
The Attiny13a appeared to program fine, with no errors in AVRDude. I downloaded the blf-a6.hex from https://bazaar.launchpad.net/~toykeeper/flashlight-firmware/trunk/files/head:/ToyKeeper/blf-a6/. However, it isn’t working as expected. The LED does not light every time I connect power, sometimes I get different brightnesses or strobe. Wondering now if I downloaded the correct hex file. I’ve tested every connection on the PCB and there are no shorts or breaks.
Why you don’t try to test in on breadboard. Much easier to tweak or undestand how the driver works. Yesterday I make the test bench for Narsilm driver with Attiny85.
Which exact 7135’s? Crappy sailboat or raptor claw? Sounds like maybe the times it doesn’t light up its actually in a mode who’s pwm_lvl is to low to light up the not good kind of 7135’s.
I read there was problems with the 380mA versions, which is why I bough 350mA from a UK seller off eBay. The listing had this image:
If it was the regulator that didn’t like the PWM, then there still would be some logic to the modes, which there isn’t. My thoughts are either a bad OTC or the hex file was for a ATTiny25. The OTC appears to be soldered correctly and I’ve traced the connections, which are fine. Is there a difference between the ATTiny13 and ATTiny13A? Do I need to recompile the code for a ATTiny13A?
I’ll try compiling the code later, flash another ATTiny13A and see what happens.
Built up another board, pretty confident both boards are fine. Out with the oscilloscope, it shows the PWM on channel 2 is working, but no drain from the 7135.
I have two other batches from different suppliers, I’ll try them this evening.
7135 doesn’t like pwm at all especially on on high frequency. I don’t have any 7135 based driver to test it. But from that old video you can see AMC7135 PWM based 1W LED driving - YouTube 7135 is awful in step response especially at lower duty cycle. Can you post any waveform from oscilloscope to the output and vcc pin at same time if you have 2 channel oscilloscope.
Just replaced the eBay AMC7135 for a L7135 (I think I bought from FastTech) and it now works. My soldering iron at home isn’t as good for SMD components as my work one, so I’ll replace the other one on Monday and add the rest of the 7135. At least I can use the AMC7135 on the first channel. I’ve got a hot air soldering station at work, so once happy I may get a little production run going, to replace the drivers in my Convoy S2 (and similar) collection. Just need some springs now.
My oscilloscope is just a single channel cheap pocket one. Its great for tracing UART signals, when instruments aren’t communicating with the PC, but I wouldn’t trust it.
Thanks TrueRMS for your advice and thanks to HarleyQuin and ToyKeeper for your PCB design and firmware.