> I know you said 5V, but I suggest using a 19-20V laptop supply instead
This is totally a viable option, but for now I want to play with reusing usb chargers. That's just part of the fun
And this means learning to solve the constraint of 5V psu with ~3V LED.
> You don't have to draw 3 watt.
Using 3V LEDs with 5V PSU, without PWM I will have to waste and dissipate 40% of the power, and that's too much.
> I use mainly TI LED drivers and boosters.
Yes, the final version will probably have a custom PCB with a microcontroller, but I still want to learn about using ready-made modules that could work in my case.
> If you do want to go the PWM route you want a 555 timer.
If we look at
this one, it indeed uses a 555 to generate a PWM waveform, which is fed to the two channels of a 9926A (schematics
here. There are still three problems for me my case :
- it can go up to 100%, which will be too much. There is no easy place to add a serie resistor with the potentiometer. I have not done the math yes but it may be possible to replace D9 with a diode with some higher forward voltage to add some off time to he pwm.
- it is a voltage source, not a current source, meaning I'll still need some (lower and less heat dissipating) serie resistor
- there is no output inductance, so the LEDs even if they have an ok average current will have a too high pick current (so I still need a pretty high serie resistor)
What I'm really trying to learn is to which of those cheap drivers would suit my need. For now my best bet is using a dimmable current source (e.g.
this driver) which has an output inductance and something else to PWM it. The PWM would solve the problem of dissipation too much power, and I'd still have to make sure PWM is maxed to about 5/3. If modifying the PWM motor controller is not possible, I can directly use an STM32 or RP2040 board to generate and a potentiometer with some code, but the lamp would look terrible with two boards five wires ^^