Dimmable LED stripes

On a sunny day (Tue, 16 Feb 2021 03:53:40 -0800 (PST)) it happened jmariano wrote in :

I designed a 12V version controlled via UDP (via the LAN from anywhere else) It is on 24/7, has build in on / off timers:

and I wrote a GUI program to control it (Linux program):

A day to night cross-fade software should not be hard to write. This circuit would work just as well on 24 V with these MOSFETS I think. One important thing I learned is that if you use PWM to control the brightness than all 3 colors in a RGB strip should run on the same clock. If not then you get the strangest flicker effects, even though your eyes _should_ not be able to see the high frequency, so I had to change the hardware so each of the PWM generating PIC micros ran on the same clock. Maybe this project can give you some ideas... Years after I build that thing I bought some 12V RGB LED strips from China that came with a remote and a controller smaller then the size of a chocolate bar..... It had all sorts of effects you could select too... I guess you could send such a controller IR signals to do anything you want, but those things are 12V. But maybe some of all this is in some way helpful...

Why wouldn't you use horticultural LEDs which omit the green LEDs since green light bounces off plants and have peak emission matched to the photosynthesis and dormancy breaking of seeds.

It looks hideous magenta but the plants will grow quite happily and you save around 30% on the electric bill.

(other brands are available)

At those low power levels it is straightforward and at 500Hz almost anything will do it if suitably driven. Having too much gain available at high frequencies can easily be a disadvantage for this.

--
Regards, 
Martin Brown

This should be very simple, as long as you have a PWM logic signal with enough voltage to turn on your selected mosfet. There are low-gate-threshold mosfets if all you have is, say, 3.3 volt PWM from some little cpu.

Don't worry beyond that about gate drivers or overshoot or anything, into a simple load like this and low PWM rates.

You could also use a solid-state relay instead of a mosfet.

Sketch a circuit and post it here and we'll review it for you.

(I once worked in a language lab and fixed tape decks and copied tapes and things, so listened to most of the languages of the world. Portugese was the most beautiful to listen to.)

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

What make and model of 24V supply are you using? Does it have remote sense terminals on it's hookup diagram or schematic?

It might be possible to modulate the output voltage predictably with extertal DC from a filtered PWM signal, employing these terminals. You'd have to calibrate this with emission levels of the string.

If you're trying to simulate biological sunlight, I would not suggest PWMing the strings directly. Plants are not eyeballs - your susceptibility to flicker is irrelevent. The effects of pwm light source on plant development are undocumented and an unneccessary source of error.

It would be simpler if you just used a pre-set intensity and turned the lights off to simulate night.

RL

w algea indoors using LED light. The light needs to be dimmable so to repro duce the sun light day cycle. The system was a controller build around an a rduino that measures a few parameters, sends data to the supervisory PC and should also controls the LED illumination. Naturally, I would like to use the arduino's PWM outputs (490Hz). One of the requirements, imposed by the "customers", is that 24 V LED strip of the home lighting kind should be use d. These are usually made of parallel associations of 6 LED in series + one current limiting resistors blocks. I will use one white and one rgb (30 W each aprox, 1.5 A each aprox). I already have a 24 V DC power supply on the system. Since there is a resistor in series with the LED and there is this parallel association, I guess I have to use a constant voltage driver, con stant current is not possible

It may be useful to put a light-measuring sensor onto a section of the stri p, and regulate the sensed light by changing the PWM setting. That would compens ate for aging of the LEDs, and any nonlinearity of the resistors and diodes. The Arduino would just need a daily cycle lookup table for the light-reading target value.

My preference would be to NOT let the light flicker, rather would drive the LEDs through an inductor with catch diode. That way, the inductor will hum, but not th e LEDs and their wiring. I'm not sure why 490 Hz is selected, but there will be some audibl e effect at such a low frequency. The light sensor would have to be low-pass filtered if the light is modulated.

Dear All,

Thank you very much for your input. My main question was enlightened, that is, I can use mosfets to drive the LED strip. I already order a few differ ent type to experiment. Replying to your individual comments:

Jan, thanks for the schematics, I'll take it in to consideration. Thanks fo r pointing out the flicker and aliasing effect.

Martin, thanks for mentioning the horticultural LEDs, I will certainly conv ey you suggestion to the biologists. To be honest with you, i'm just implem enting the electronics was they asked, without giving to much thought to th is matters. They are supposed to be the specialist on algea culture with ar tificial light. As for your last sentence, "Having too much gain available at high frequencies can easily be a disadvantage for this", I completely mi ss his meaning.....

John, thanks for the compliment! But maybe you were listening to Brazilian Portuguese, which is softer and more sung. Usually people find European Por tuguese to be more hard to the ear.

RL, I'm using one of these open frame switch PS to supply the all system, n ot only the LEDs, so it can't be modulated. As for the flicker, again you a re totally right, but the biology people have been doing this for a long ti me so I guess that the effect of flicker is integrated on the overall respo nse of the system.

whit3rd, the light sensor is a option under consideration. And, after all t his comments about the biologic effects of the pulsing light, I will defini tely try your solution for filtering the pwm pulses.

Again, thank you very much for your valuable time.

Regards

Mariano

Ming the strings directly. Plants are not eyeballs - your susceptibility t o flicker is irrelevant.

Photon capture in the eyeball is different from photon capture in plants.

I don't know the the 0.09 seconds comes from. I once got to listen to short tutorial on the subject from a professor of inorganic chemistry that I kn ew from our primary school days, but the only message that stuck was that t he process was remarkably complicated.

d an unneccessary source of error.

Putting enough filtering on the current going through the LED to make the l ight output more or less constant might not be cheap, but it might avoid so me perverse results.

Non-linearity shows up all over the place, usually precisely where it is mo st inconvenient.

--
Bill Sloman, Sydney

The light output efficiency from a LED drops when the current grows towards Imax. While some of this is temperature dependent, but this is also observable at Tj=25 C (i.e. during initial startup).

Thus, running at say 50 % duty cycle unfiltered PWM will produce less light than running the LED at Imax/2 DC current. Some PWM filtering will prevent Imax current peaks during the ON period, if the duty cycle is below 100 %.

With very small LED chips and too slow PWM rate, the chip may reach a higher temperature towards the end of the ON period. Again, some filtering will help.

If the greenhouse lights are going to be used for several hours below

100 % duty cycle. filtering the current waveform at least to some degree will produce more light with less mains energy consumption.

On a sunny day (Sun, 21 Feb 2021 09:03:54 +0200) it happened snipped-for-privacy@downunder.com wrote in :

These LED strips have resistors in series with the LEDs. So 100% PWM means LED in series with resistor on DC. Most heat in a 12 V (and even more in a 24 V I'd think) setup goes into the resistors anyways.

---------------------------------

** Question:

What makes you assume LED strips put out enough photons to grow anything? Sunlight is about 20- 50k lux, LED strip are about 50 lux

Big difference.

Make doing theatrical style fade outs seem idiotic.

...... Phil

indeed one or two LED strips are needed when running at 20 mA for each square meter.

a grid with 12 cm between LEDs is needed. Not that hard to arrange.

Thanks for the references.

RL

d here, but I was unable to find an suitable answer.

w algea indoors using LED light. The light needs to be dimmable so to repro duce the sun light day cycle. The system was a controller build around an a rduino that measures a few parameters, sends data to the supervisory PC and should also controls the LED illumination. Naturally, I would like to use the arduino's PWM outputs (490Hz). One of the requirements, imposed by the "customers", is that 24 V LED strip of the home lighting kind should be use d. These are usually made of parallel associations of 6 LED in series + one current limiting resistors blocks. I will use one white and one rgb (30 W each aprox, 1.5 A each aprox). I already have a 24 V DC power supply on the system. Since there is a resistor in series with the LED and there is this parallel association, I guess I have to use a constant voltage driver, con stant current is not possible

eed a more sophisticated solution?

tually using a special IC, preventing overshoot, etc, as well as quality of light, geometry, flicker, etc, but i will address all those on a next stag e.

eciate.

You can buy a universal dimmer:

drivers.html You don't need to worry about the various safety and hazard issues using on e of these. Link is just an example, not a recommendation. I'm sure there a re others out there.