Dear readers,

I have an idea of a design which uses a LiPo battery which has to power a 12V led strip.
Right now, I have this boost converter to 12V:
https://www.olimex.com/Products/Breadboarding/BB-PWR-3608/resources/MT3608.pdf
12V is going to a mosfet which is controlled with PWM by a microcontroller to dim the LEDstrip.
This works well, I can controll the brightness of the ledstrip by varying the duty cycle as expected.

But I would like to improve the efficiency and reduce part count, because using a boost converter AND a mosfet seems a bit redundant.
I need the microcontroller anyway for other stuff...
I can't use lower voltage LEDs nor put more LiPo in series due to some constraints.

So I was thinking to control the boost converter enable pin with the microcontroller to directly switch the 12V output on/off with the PWM.
I think the boost converter should use a much higher frequency (1.2 MHz) than the PWM (2-10 kHZ) to have a stable 12V output and the PWM freq should be low enough so that the filter capacitor at the output of the boost converter does not smooth the PWM so much as to getting a voltage somewhere between 0-12V which would result in a difference in color because of the change in current and not being able to dim low enough do to the forward voltage of the LEDs.

I think this all sounds very logical and I think that i will be able to design this.
Does any of you see something i am missing?
Either a mistake in my thinking or a more simple solution like an off the shelf product that can do this?

Thank you in advance for any hints.

Best regards,
Zaishens

 

For the price of a MOSFET, is it worth it to go to a processor controlled converter?
Anyway, if the LED strip has the resistors included, you do not want to alter the voltage of the supply.
PWM s easy and works well. To switch the power supply on and off fast enough is probably beyond the design of the power supplies, but it may be possible to toggle the enable on one fast enough.


https://www.ti.com/lit/ds/symlink/lm2596.pdf
You could try the enable on one of these.

 

For the price of a MOSFET, is it worth it to go to a processor controlled converter?
Anyway, if the LED strip has the resistors included, you do not want to alter the voltage of the supply.
PWM s easy and works well. To switch the power supply on and off fast enough is probably beyond the design of the power supplies, but it may be possible to toggle the enable on one fast enough.

View attachment 178042
https://www.ti.com/lit/ds/symlink/lm2596.pdf
You could try the enable on one of these.

Dear dendad,

Thank you for the reply.

This is what i suggested myself, I edited my orginial post to clarify.

I am using a step-up converter to 12V, not step-down like you posted.
Right now I am using a MT3608:
https://www.olimex.com/Products/Breadboarding/BB-PWR-3608/resources/MT3608.pdf

I could use the PWM from my microcontroller that goes to the MOSFET to control the enable pin instead, I'm just wondering if that is a good idea before i tear my current working solution apart.

 

I think I'd drive a small FET from the micro to switch the EN pin in case the internal pullup in the regulator has more that 5V on the EN pin.
You could measure the EN open circuit voltage and see.

 

I think I'd drive a small FET from the micro to switch the EN pin in case the internal pullup in the regulator has more that 5V on the EN pin.
You could measure the EN open circuit voltage and see.

Thanks for the tip, I dont think this will be the case since the datasheet states: "When not used, connect EN to the input supply for automatic startup." but I will check to be sure.

Do you have any further tips relating the differnent frequencies and filtering, to have a steady 12V output (1.2MHz) but the PWM (2kHz) output is not going to act like a variable voltage. If I use the suggested 22uH inductor and 22uF in- and output capacitors, I think this will not work as intented as LEDs have 150 ohms in series with 3 leds (3x3.2=9.6 V) in series and there are 180 of these strings. giving a RC time of 150/180*22 uF = 18 us. Which is way too long since the period for 2 kHz singal = 1/2kHz = 500 us and dimming down to 1% duty gives 5us.

So i guess i should use a lower output cap like 2.2 uF

 

All I can think off is ti try it and see how it goes.
the data sheet says "The MT3608 has internal soft start ..." so that may cause you problems.