PWM LED driver for 1w 280mA LEDs?

I'm working on a project which I have been using variable buck converters to drive series groups of 1W LEDs. The LEDs are wired in series in groups of x6 and the main supply is 24VDC. So far the project has gone well but I would like to expand it in that I would like to use PWM to dim the LEDs.

I have a single variable PWM Buck which set up does the job, but it seems an expensive option. I am looking to do this simply and cheaply, I have achieved the same effect with a resistor inline before now but that's not a very efficient way about it.

Can anyone point me in the direction of a simpler or lower cost alternative? The noon PWM Buck converters I bought are around £1.50 each but the PWM version is £5.69. I have found other buck converters with PWM but they all seem to have a fixed output at 350mA and I need 280mA or they just get too hot.

This is the PWM buck I have and want to replace as I need several of them:

eBay PWM Buck

Thanks

Since those are just constant voltage buck converter presumably you have
LED strings with a current limiting resistor? You should be able to just
add a MOSFET low-side driver to do PWM if lucky. The buck converter's output
capacitor might need a bit of help if it struggles (more output capacitance).

Alternatively have you looked directly for constant current buck converters with
a PWM input? Then you can lose the series resistor.

Thanks MarkT, the one I have put a link for is constant current & constant voltage, although I just adjusted it to limit the current down to 280mA using the pot. But I want to do away with it altogether and use a different component that is slightly cheaper since I am probably going to be using quite a lot of them.

Is it feasible to use a constant current driver which has a fixed output of 350mA and use PWM to reduce the load? to bring it down to 280mA consumption.

I have looked directly for constant current buck converters with PWM but they are all (the fixed ones) higher than 280mA.

OK, you will notice some specific points in the design of the driver you cite - it is designed for PWM but the PWM frequency is specified as 100 to 300 Hz.

In fact, it will undoubtedly operate with lower PWM frequency and the lower limit is merely to avoid visible flicker. The upper frequency relates to the stability of the 300 kHz switchmode converter.

You cannot use PWM with a switchmode converter unless it is designed into the device. This is because the PWM switching will destabilise the feedback arrangements of the converter. The 1000 to 1 ratio of frequencies specified here permits the feedback circuit to compensate adequately for each on/ off cycle.

You are correct in observing that you need a separate current driver for each chain. If you are going for economy, you may have to consider falling back to simple linear circuits.

{Sense resistor 2.7 ohms, control resistor 180 ohms, drive transistor will dissipate about 2W, needs heatsink. If you wish to control more than one of these per Arduino pin you need to buffer the control input as it necessarily draws 20 mA though you may get away with significantly less current.}


spondootre:
Is it feasible to use a constant current driver which has a fixed output of 350mA and use PWM to reduce the load? to bring it down to 280mA consumption.

You cannot actually reduce current by switching it on and off, however it is true that at a reasonable frequency, you can use a higher current at a reduced duty cycle (if the converter is PWM capable); the LEDs probably will tolerate this OK as long as you ensure that any circuit failure or software "hang" can not result in the LEDs continuously switched on.

Those "fixed current" PWM drivers have their current set by resistors. If you don't mind having to buy a few low ohm SMD resistors (less than 1 ohm) and a little soldering then you can adjust their current setting as you like.

As an example, this driver based on the MBI6651 has its current set with three resistors (R1, R2, and R3 are in parallel, which gives you more flexibility in resistor choice) or this PT4115-based driver has its current set with R1 and R2 in parallel. The datasheets for either chip are readily available and you'll find that the current calculation is pretty straightforward where Iout = .1 / R or such.

Thank you paul__B for the comprehensive answer that sheds some light on what to do next....

Chagrin, I was just looking at the fixed current ones and had the same thought, then came here to ask the question and find your answer! Thanks!! I have ordered x2 of these to play with and figure out the circuit + hopefully modify it....

"fixed" constant current driver

Would such a device work properly with a potentiometer instead of a fixed resistors?

Hmm, yes this could be even handier since the different LEDs in strings are rated differently per colour.

It would work with a potentiometer (rheostat) but you'll have a hard time finding one in a 1 ohm range. Best chance is a 10 ohm that is 20 turns.

I'm a little lost with that, why would it matter how many turns? Or are you just meaning package size?

I have a 4.7Ω pot I could use once it arrives, although tbh i'm thinking of just using a fixed value once I work out the value needed for each colour.

I mean "where the heck did you find a 4.7ohm, single turn pot?"

Ok, I did some testing.
I had this thing lying around, so I set the current limit to 100 mA and slowly cranked up the voltage. The 10W LED lights up and I can dim it by fiddling with the current. The LED will not shine very well at low levels (it flickers), but once it gets going you do get some regulation.
I haven't tried getting it up to full 10W, though. I got up to about 4W, since I didn't put much of a heatsink on the LED and wasn't comfortable with the heat.

However, that thing too bulky to be installed all over the place.
This thing, on the other hand, is smaller and cheaper.

I cannot see a PWM input on either of those things. :astonished:

Chagrin - it's this one, it is a panel mount though and I've had it a while....looking up the prices they are over the top expensive anyway so not much use......

Panel mount pot data sheet

I'm going to wait for the boards to arrive and just try a small form factor board mount pot, regardless of the number of turns since it only needs setting once and I think you may have confused what I am trying to achieve which is not to dim the LEDs with a pot just to set the constant current output from the driver so that the dimming can be achieved directly from the arduino code and PWM. Will update when I've experimented, thank for all the help people. :slightly_frowning_face:

Paul__B:
I cannot see a PWM input on either of those things. :astonished:

There is none to be seen. I forgot this is the topic about PWMing the LEDs. I got lost in my own little thoughts.
Anyway, replacing the trimmers with digital pots would offer a dimming capability, right?

Shpaget:
Anyway, replacing the trimmers with digital pots would offer a dimming capability, right?

No.

Really? Why is that?

Thought you might ask!

"Digital Pots" are generally intended for use for such functions as volume controls and have relatively high resistances of the order of 1k at minimum. That much does not seem to be a problem in the switchmode converters cited which use values of 10k and 100k for the potentiometers but you clearly cannot use them for values in the low ohms range.

I seem to recall the OP is looking at 24V or so and if the current sense pots are referenced to the output voltage (the voltage control pot most certainly is) then I think you will find this beyond the ambit of the "digital pots".

Note that current limiting is not native to the LM2596 used in these cheap power converters, so those that offer current control use an additional LM358 op-amp to sense this.

Thanks for explanation.