My experimental LED grow light strips don't seem to like PWM.

The current setup is this:

30A Meanwell (clone) PSU 24-27v

eDMX Artnet to DMX


30 Channel DMX->PWM Splitter


A series of High Speed 30A LED Signal Repeater/Amplifiers


And directly into the COB strips connected 3 in parallel (each strip is rated at max 27v 4A, but I run at 24v ~3A)

From a logical flow perspective we have: ArtNet-> DMX -> PWM -> Signal Amplifier -> LED COB strip (~3A @ 24v)

Am sure the high speed LED amplifiers are stressing them - but all my control circuits assume PWM (i.e. Artnet->DMX->PWM etc.) so am looking at a circuit that can smooth the output before hitting the lights.

Appreciate this can be done with an LC (Inductor+ Capacitor), but each channel will be drawing around 9Amps at 24v so it's somewhat more than electronic levels.

Alternatively, I throw out the control circuits, and start over - in which case I'd want a series of low cost BUCK converters that can each handle around 4A @ 24v and be able to be controlled say with a 0-5v signal line or with the 24v PWM Signal line.

Apologies in advance - I'm an embedded software guy with a knowledge of electronics that makes me probably more dangerous than useful.

Welcome thoughts on a circuit/components.

  • 1
    \$\begingroup\$ Do you have a link for the lights you are using? Are you using PWM to modulate the brightness of the lights? I don't know much about grow lights, but if you are using LED-based lights then you are pretty stuck with PWM. LEDs requires a certain forward voltage to operate. Filtering the PWM would likely reduce the LED input voltage to below the forward voltage of the LEDs and then they won't work at all. Probably need more info about the lights you are using and what you are trying to accomplish here... \$\endgroup\$ Commented Mar 25, 2022 at 21:52
  • 1
    \$\begingroup\$ Most likely the LEDs use analog CC control so some analog CV control voltage is required. It is far easier to filter a high impedance CV signal from PWM than 4A \$\endgroup\$ Commented Mar 25, 2022 at 22:20
  • \$\begingroup\$ I am using raw 24v COB and SMD strips - no controlling. \$\endgroup\$ Commented Mar 26, 2022 at 6:38
  • \$\begingroup\$ The lights will work comfortably in a range of say 12-27v \$\endgroup\$ Commented Mar 26, 2022 at 6:39
  • \$\begingroup\$ I am getting a bit done with PWM. The LED Amplifiers make the lights scream (in pain lol). Where as a dedicated buck will gently increase and decrease the available AMPs with no aggressive switching. And I see that a good Buck is comparable with PWM in terms of efficiency. So ideally there is a buck converter that I can use with the existing PWM signal somehow for control. Or alternatively, smooth the PWM output somehow without costing/loosing too much. \$\endgroup\$ Commented Mar 26, 2022 at 9:18

1 Answer 1


If you're using constant voltage LEDs, and they light up at 12V (meaning 4 diodes in series) but you're driving them at 24V with resistors, your setup is horrifically inefficient since only ~50% of your power is even going to the LEDs. This is the reason horticultural lighting is virtually never constant voltage but rather constant current which does not require heating up resistors. Unless you are not paying for electrical power or only planning to briefly use these, it is probably cheaper to start over with high efficiency, constant current lights as you are going to spend more money heating up resistors than it would cost to buy proper grow lights before long.

That said, if you want to solve this problem while keeping your existing lights, most constant current LED drivers (e.g. Meanwell HLG series) will have an option for PWM input, which they use to regulate the output current. In this case you have a constant voltage strip, but nothing stops you from putting your strips in series, getting a ~3A constant current driver, and then controlling the current with PWM. You'll still be wasting a ton of power heating those resistors, but it will be slightly more efficient since you won't be running the LEDs at higher peak currents.

I mean ultmately changing the PWM on a constant voltage is effectively the same as changing the current (except in my case I'd add a small inefficiency with filtering the PWM).

A constant current driver refers to a special type of power supply that measures its own current output and then adjusts voltage to hold that current output constant. They are the most common way to power LEDs, and are used in most light bulbs, lighting fixtures and horticultural lighting because they are the most efficient way to drive LEDs.

If you take a constant voltage source and reduce its average voltage by PWM, the result is still a constant voltage source. Such a source cannot (safely) be used to power a load that requires a constant current source because the required circuitry to measure and regulate current is missing. More information here: LED drivers: constant voltage vs. constant current

Do not attempt to power individual LEDs or those that expect constant current drivers with a constant voltage source, even if the diode does not blow up immediately (the most likely result), you could easily start an electrical fire later on when no one is around to notice.

  • \$\begingroup\$ Thanks for this. Yes, the COBs run pretty hot and need cooling. Quite wasteful. I am planning to move to these Samsung SMD type lights and see that they have both 12 and 24v models - am hoping the 24v will be not be wasteful as you suggest. alibaba.com/product-detail/… \$\endgroup\$ Commented Mar 28, 2022 at 4:04
  • \$\begingroup\$ @AndrewMcClure Don't waste your money on constant voltage strips, you're just paying to heat up resistors. \$\endgroup\$ Commented Mar 28, 2022 at 4:14
  • \$\begingroup\$ To be fair, I have measured the PAR values as a function of wattage, and the COBs I have are not too different from the fancier SMD variety. But thanks for the heads up! \$\endgroup\$ Commented Mar 28, 2022 at 4:39
  • \$\begingroup\$ @AndrewMcClure At 24v I expect them to be about 1/3 to 1/4 as efficient as you should be aiming for, just based on the circuit and from where you bought them. Unless you have very cheap power, aim for a minimum efficiency of 2.5 μmoles/joule, and be careful about trusting the numbers cited by questionable vendors online. \$\endgroup\$ Commented Mar 28, 2022 at 12:36
  • \$\begingroup\$ Thanks guys. I've had a look at the Meanwell PSUs with PWM control. If I change to constant current SMD lights (i.e. say something like the LM562) and drive it with my LC filtered PWM output - am wondering how much less efficient that would be than going with a (relatively expensive!) PWM controlled Meanwell constant current driver. I mean ultmately changing the PWM on a constant voltage is effectively the same as changing the current (except in my case I'd add a small inefficiency with filtering the PWM). Hmmm. May need to experiment. \$\endgroup\$ Commented Mar 29, 2022 at 0:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.