So I could take a 105c and with a bit of creative soldering connect a 7135 to one of the stars and reprogram the driver to PWM that line for moonlight? I might have to get the programing tools.
Yes, it would definitely need to be reflashed with suitable firmware. None of the drivers comes with any way to control the other (potential) output pins, they only use pin #6. The others are either not used or used as inputs (from the stars). It would need custom firmware to change those input pins to outputs.
I sort of, theoretically at least, worked out a way to do this with the attiny13 grafted onto the gen1 DQG tiny 26650's boost driver. It uses no PWM on any of the outputs, each output pin is either on or off and each pin has its own small FET and resistor. Outputs are used only one at a time, and the level for each mode is set by the resistor value on that 'channel'. Wight wrote a pre-alpha firmware for it but I have yet to assemble & test it. It's definitely possible to do with that particular mix of hardware, though.
The now-discontinued Flex Asgard driver also used individually addressable 7135s, though it uses a different MCU.
ATtiny 25 or a PIC would be better suited for that, alot more suitable, l as larger flash space, I think the PIC does internal volt check, so no resistor bridge needed (more driver space for 7135’s)
Wasn't Tido the first to use that kind of setup? One pin with one AMC and PWM, then 1*AMC, 2*AMC,4*AMC on other pins without PWM; a level of 20% could then be achieved by 1*AMC full and the PWM-AMC at 60%, gives 12.5%+60%*12.5%=20%.
I used some creative cutting and soldering to have 4 independent channels of 2*AMC each with software-PWM on a NANJG105C for my first RGBW driver (for my newer ones I designed a PCB to avoid that).
I mentioned in passing a long while ago that I wanted to do this multiple pin thing, and did eventually get around to building a driver for my D4 tube light that worked well and I liked a lot.
My theory was(is) that LED’s are more efficient at lower currents, and also better regulation is achieved using constant current rather than pwm’ing down from a high current.
So I modified Tido’s firmware to PWM three 7135’s, giving me moonlight, low, medium and a normal high at 1amp, three fast clicks turned on pin 3 which were wired to seven more 7135’s giving a hidden turbo mode.
Unfortunately after dozens of drops on to concrete floors the wire to the emitter pulled off the PCB and blew the driver, to get it going I chucked in a single mode 1Amp driver but I miss the original.
Thanks for all the responses. I may want to try this using a Nanjg 105c (when time allows it). And try to code for it. Seems like a lot of work :weary: .
I found a data sheet for 7135’s that are available in as low as 120 mA output does anyone know of a source of 7135’s other than the 350 mA and 380 mA ones?
Actually, it is just binary, so you can do better than I said. The PWM is the part after the binary point. For the integer part, the number of 7135s should double at each pin.
Assuming no other considerations!
This is the maximum control with limited control data, if all 7135s are equal.
Personally I think there’s little point in controlling individual 7135’s it’s easier to have three or four and PWM down to the level required, plus tint shift is avoided.
Here’s a old pick of the mod, pin 6 control three 7135’s on the front and pin 3 controls the one at top right then through a via to six more on the back for turbo.
Another thing to consider is that for the mode spacing to look right it needs to increase by around 3x between levels. Just adding one 7135 at a time would look awful.
I had a flashlight that ramped (the original driver on a Yezl t9 before it broke). It was good for higher levels, but it was hard to select a very low level because they went by too fast. So how about a smooth logarithmic ramp, doubling the current maybe once a second? Of course there needs to be a separate way to access high.
Did anyone mention that forward voltage increases with (instantaneous) current, so reducing the amount of pulse width modulation needed would reduce the heat output of the LED, for a given average current, and cause the driver to stay in regulation longer, as well as increasing the light output?
The tint, on the other hand, may depend on temperature of the phosphor, so tint shift might also be reduced.
Radio frequency interference would also be reduced.
The only disadvantage seems to be the complexity.
