MeanWell LED Drivers: 3 in 1 Dimming Function.

[BobCajun]

That's not dimming with PWM as we are discussing it here.

You can dim these drivers with a resistor, a PWM signal or a 0-10 voltage. The PWM signal and resistor are actually converted to a voltage inside the driver. So in the end they do exactly the same thing even.

So it's doesn't matter which method you use, these achieve exactly the same thing. Which is that the driver dims its output and the leds are NOT switched on and off at full power.

Nah. Not even. Prove it. It's full on and 10%, no in between. It's possible you're right, I just would need to see some kind of evidence. Actually here's why I think you're wrong. I have used very low frequency duty cycles, where I could actually see the strobe effect. How would that be possible if what you say is true? Well come to think of it, it wasn't a meanwell, it was a Cree. But probably similar circuitry.

 

[TimoTapani]

Optocoupler instead of mosfet. No possible leaks or whatever, because the input and output sides are completely separate and the light that activates the output side can only be on or off. You used a complicated method instead of the easiest method there is. The optos cost like next to nothing too. All you need is arduino, 250 ohm resistor and opto. Simple as hell.

Thanks for the tip. It's second time I've been told so on this thread, but I'm happy to get atleast some advice. Reason why I'm using different parts is that simple that basicly I have to order every single bit from China and try to use what I got in reach in case I don't know a month earlier which parts to order.

Luckily I had a bag of EL817 optocouplers on my bitbox. Most of the datasheet I don't understand and believe that current and voltage handling isn't problem even without looking at the numbers. Hope that'll work (or fresh mosfet). Could you still, to make this idiotproof for me, tell in which position resistor should be placed? I assume between digital I/O and opto positive. How about common ground? Should I wire DIM- to Arduinos GND or not?

(I know this shouldn't be this hard, since I've done some Arduino PWM:ing earlier. I've just forgot some things and modified others, so I'll have to find the magic again. )

 

[TimoTapani]

Nah. Not even. Prove it. It's full on and 10%, no in between. It's possible you're right, I just would need to see some kind of evidence. Actually here's why I think you're wrong. I have used very low frequency duty cycles, where I could actually see the strobe effect. How would that be possible if what you say is true? Well come to think of it, it wasn't a meanwell, it was a Cree. But probably similar circuitry.

Actually most of the link you posted adressed to this question. Only model MeanWell makes that does PWM:ing the actual led-elements is model named "PWM". All others just lower the current according to PWM signal. Check it out.

Maybe it's the same effect I had with leaking pot. Driver could have misunderstood low cycle PWM just as difference in voltage, not PWM.

 

[BobCajun]

Thanks for the tip. It's second time I've been told so on this thread, but I'm happy to get atleast some advice. Reason why I'm using different parts is that simple that basicly I have to order every single bit from China and try to use what I got in reach in case I don't know a month earlier which parts to order.

Luckily I had a bag of EL817 optocouplers on my bitbox. Most of the datasheet I don't understand and believe that current and voltage handling isn't problem even without looking at the numbers. Hope that'll work (or fresh mosfet). Could you still, to make this idiotproof for me, tell in which position resistor should be placed? I assume between digital I/O and opto positive. How about common ground? Should I wire DIM- to Arduinos GND or not?

I hooked resistor to pin D13, on an arduino nano, and the other end to the opto terminal with the dot, so that's the positive. And the negative from the opto was connected to the arduino pin labelled gnd. No other grounding involved. Then I just hooked the two opto outputs to the two dim leads of the driver. Here's one of the codes I used. Note "low" is the on time, as you probably know, because it's reversed by the opto.

void setup()
{
pinMode(13, OUTPUT);
}

void loop()
{
digitalWrite(13, HIGH);
delayMicroseconds(100); // Approximately 84% duty cycle @ 1KHz
digitalWrite(13, LOW);
delayMicroseconds(500);
}

 

[BobCajun]

Actually most of the link you posted adressed to this question. Only model MeanWell makes that does PWM:ing the actual led-elements is model named "PWM". All others just lower the current according to PWM signal. Check it out.

Maybe it's the same effect I had with leaking pot. Driver could have misunderstood low cycle PWM just as difference in voltage, not PWM.

Yeah but if you read the data sheets they say that's for a 10v PWM signal. That's if you feed an external 10v power supply into the terminals, not if you use something like an optocoupler which only opens and closes the 10v current supplied by the driver itself. Two different situations.

 

[TimoTapani]

Yeah but if you read the data sheets they say that's for a 10v PWM signal. That's if you feed an external 10v power supply into the terminals, not if you use something like an optocoupler which only opens and closes the 10v current supplied by the driver itself. Two different situations.

I don't want to argue on this, because I don't know much.

"Most of MEAN WELL’s led drivers achieve the dimming by reducing the amplitude of the output current being fed to the LED lights."

"MEAN WELL 3-in-1 LED drivers are compatible with dimming controllers that generate a PWM control signal, but this does not mean the LED driver has a PWM output."

I feel anyway that these "prove the point". Again I may have misunderstood things, but for me it seems clear that the hlg-driver isn't switching leds on and off to do dimming.

Anyhow. No I feel things are getting even more complicated than before. If I understood right, your point is that it's not efficient to do PWM without 10V external power?

 

[BobCajun]

I don't want to argue on this, because I don't know much.

"Most of MEAN WELL’s led drivers achieve the dimming by reducing the amplitude of the output current being fed to the LED lights."

"MEAN WELL 3-in-1 LED drivers are compatible with dimming controllers that generate a PWM control signal, but this does not mean the LED driver has a PWM output."

I feel anyway that these "prove the point". Again I may have misunderstood things, but for me it seems clear that the hlg-driver isn't switching leds on and off to do dimming.

Anyhow. No I feel things are getting even more complicated than before. If I understood right, your point is that it's not efficient to do PWM without 10V external power?

I just never hear of anyone using a 10v PWM unit to dim their drivers, just using arduinos. Probably works the way you said, with such a unit. With an arduino and optocoupler though, I'm pretty sure it's just flashing at full power and down to 10%, just like if you manualy touched the dim leads together and pulled them apart rapidly. Could be wrong. Only way to tell for sure would be with a current meter and a light meter, and see if efficiency increases with the dimming or not.

 

[wietefras]

Nah. Not even. Prove it. It's full on and 10%, no in between. It's possible you're right, I just would need to see some kind of evidence. Actually here's why I think you're wrong. I have used very low frequency duty cycles, where I could actually see the strobe effect. How would that be possible if what you say is true? Well come to think of it, it wasn't a meanwell, it was a Cree. But probably similar circuitry.

That's a big difference between you and me then. I KNOW for sure that I'm right. Otherwise I wouldn't say it. I wouldn't scream about something that I have no clue about.

Even the article you linked to explains the difference:

WHAT IS THE DIFFERENCE BETWEEN A LED DRIVER WITH PWM DIMMING AND A LED DRIVER WITH PWM OUTPUT?
To answer this question we explain the difference using two different scenarios:

Scenario 1:
LED driver has a PWM output to maintain even dimming, regardless of load.

Scenario 2:
A PWM control signal is used to adjust the output of a dimmable LED driver (with or without PWM output).

It's 100% clear we are talking about a control signal to the driver and not PWM of the output.

Or perhaps a good idea would be to read the Mean Well driver datasheet? See how each three dimming types has the same chart for output current?

When you manually could touch the dim wires together that's not the same as PWM. Unless you can do this at least hundred times per second.

If you actually want to understand how this works then Google the word snubber. That's how they convert that 10V PWM signal into a 0V to 10V signal for the driver.

Same with the resistor. That also ends up as a voltage on the driver input as it's basically a voltage divider based on Ohm's law.

You're welcome.

 

[TimoTapani]

I just never hear of anyone using a 10v PWM unit to dim their drivers, just using arduinos. Probably works the way you said, with such a unit. With an arduino and optocoupler though, I'm pretty sure it's just flashing at full power and down to 10%, just like if you manualy touched the dim leads together and pulled them apart rapidly. Could be wrong. Only way to tell for sure would be with a current meter and a light meter, and see if efficiency increases with the dimming or not.

Thak you, sweet baby Jesus! Just did some tests and optocoupler really does work, even though the logarithmicity is still there near min and max. Maybe I just tackle that with some software function. (That was my problem in the first place )

I didn't mean a 10V PWM unit (atleast not factory build), but rather simple selfmade one, which has been described on this thread few times. It has only 12V to 10V regulator and a transistor/mosfet for control. Not much more complex and with rough estimate works quite exactly same way as the optocoupler.

Btw. Tested same parts as yesterday: 2n7000 Mosfet with simple led attached to protoboard: Perfect PWM dimming. Took it back to real led driver and hooked it in. Got same result as yesterday: only 5-10% power. Conclusion: 2n7000 isn't suitable for this job while some transistors could work connected straight to DIMs. Maybe I should add a diode or a resistor somewhere to make it work, but naah - I'll give up on that and go with the photocoupler.

Thank you for your guidance.

 

[BobCajun]

Thak you, sweet baby Jesus! Just did some tests and optocoupler really does work, even though the logarithmicity is still there near min and max. Maybe I just tackle that with some software function. (That was my problem in the first place )

I didn't mean a 10V PWM unit (atleast not factory build), but rather simple selfmade one, which has been described on this thread few times. It has only 12V to 10V regulator and a transistor/mosfet for control. Not much more complex and with rough estimate works quite exactly same way as the optocoupler.

Btw. Tested same parts as yesterday: 2n7000 Mosfet with simple led attached to protoboard: Perfect PWM dimming. Took it back to real led driver and hooked it in. Got same result as yesterday: only 5-10% power. Conclusion: 2n7000 isn't suitable for this job while some transistors could work connected straight to DIMs. Maybe I should add a diode or a resistor somewhere to make it work, but naah - I'll give up on that and go with the photocoupler.

Thank you for your guidance.

Something you could try sometime is hooking the arduino directly to the dim leads. Since it puts out 5v instead of 10v the on part of the duty cycle would be 50% current, theoretically. So you just use twice as many LEDs, like if you were running them soft for higher efficiency. Hooking the arduino up directly would save a lot of screwing around. The arduino only puts out 40 ma so I'm sure that wouldn't hurt the driver. Your mosfet thing probably put out more current than that.

 

[BobCajun]

That's a big difference between you and me then. I KNOW for sure that I'm right. Otherwise I wouldn't say it. I wouldn't scream about something that I have no clue about.

Even the article you linked to explains the difference:


It's 100% clear we are talking about a control signal to the driver and not PWM of the output.

Or perhaps a good idea would be to read the Mean Well driver datasheet? See how each three dimming types has the same chart for output current?

When you manually could touch the dim wires together that's not the same as PWM. Unless you can do this at least hundred times per second.

If you actually want to understand how this works then Google the word snubber. That's how they convert that 10V PWM signal into a 0V to 10V signal for the driver.

Same with the resistor. That also ends up as a voltage on the driver input as it's basically a voltage divider based on Ohm's law.

You're welcome.

Okay so prove it then, or you're just flapping your gums. I already told you how, current meter and light meter. Though actually I don't give a shit if you're right or not. It's apparently way more important to you than it is to me. Here, I'll help you out, you're right and I'm wrong. That's how little I fucking care.

 

[wietefras]

@BobCajun, I proved it already. It's literally explained in the article you linked. Plus it's in the datasheet. It cannot get much more than 100% clear.

That you don't understand the "proof" is not my problem. It does indicate that whatever I say, you will never understand or believe any "proof" I provide. ie I'm wasting my time on you. Again.

 

[wietefras]

Something you could try sometime is hooking the arduino directly to the dim leads. Since it puts out 5v instead of 10v the on part of the duty cycle would be 50% current, theoretically. So you just use twice as many LEDs, like if you were running them soft for higher efficiency. Hooking the arduino up directly would save a lot of screwing around. The arduino only puts out 40 ma so I'm sure that wouldn't hurt the driver. Your mosfet thing probably put out more current than that.

Come on man, what daft advice is that. Double the number of COBs and drivers just because you cannot be bothered to add a optopcoupler/transisitor and resistor to the dimming circuit?

Seriously, can't you first try to understand the matter at hand before you start posting things like this? You seem to understand that you don't really know what you are talking about why then still post things like you are posting the ultimate truth?

 

[BobCajun]

Come on man, what daft advice is that. Double the number of COBs and drivers just because you cannot be bothered to add a optopcoupler/transisitor and resistor to the dimming circuit?

Seriously, can't you first try to understand the matter at hand before you start posting things like this? You seem to understand that you don't really know what you are talking about why then still post things like you are posting the ultimate truth?

Do me a favor? Go fuck yourself. Cuz pretty unlikely I'm going to tailor my advice to your specifications, numbnuts. I draw your attention to this part of post #311;

"Though actually I don't give a shit if you're right or not. It's apparently way more important to you than it is to me. Here, I'll help you out, you're right and I'm wrong. That's how little I fucking care."

You seem to have an aversion to the notion of using more LEDs at reduced current. I suggest seeking treatment for that. Because unless I suggested it to YOU, why the fuck would you care unless you had a bizarre form of psychosis?

 

[lukio]

bobs an idiot, dont listen to bob

 

[wietefras]

Indeed, now trying to stop an idiot spreading misinformation is a psychosis?

Seriously, you seem to know you don't really understand things. It's baffling how you can then in one breath continue to actually pretend that you do know all the answers. It's with every "discussion" you start here. You rant about how everybody should be listening to your ultimate truth cause everybody has been doing it wrong all along and you are finally seeing the light how it should really be done. Even though you're not quite sure how it works, but you must be right (somehow). When people try to explain you are wrong you just keep pretending that you know it all and defy all the logic and proof that is presented to maintain your nonsensical claims.

What psychological condition is that herr doctor? Ostrich syndrome? Cognitive dissonance?


Anyway, all you need to do 10V PWM dimming from an Arduino 5V PWM signal is a transistor and a resistor. Or use 2 transistors if you don't want the signal inverted. Otherwise 0 on the Arduino will be 100%, but you can also send signal=255-outputlevel to the arduino pin to invert the signal back.

The DIM wires are effectively a 100mA 10V constant current source. That's how all this works.

 

[BobCajun]

Indeed, now trying to stop an idiot spreading misinformation is a psychosis?

Seriously, you seem to know you don't really understand things. It's baffling how you can then in one breath continue to actually pretend that you do know all the answers. It's with every "discussion" you start here. You rant about how everybody should be listening to your ultimate truth cause everybody has been doing it wrong all along and you are finally seeing the light how it should really be done. Even though you're not quite sure how it works, but you must be right (somehow). When people try to explain you are wrong you just keep pretending that you know it all and defy all the logic and proof that is presented to maintain your nonsensical claims.

What psychological condition is that herr doctor? Ostrich syndrome? Cognitive dissonance?


Anyway, all you need to do 10V PWM dimming from an Arduino 5V PWM signal is a transistor and a resistor. Or use 2 transistors if you don't want the signal inverted. Otherwise 0 on the Arduino will be 100%, but you can also send signal=255-outputlevel to the arduino pin to invert the signal back.

The DIM wires are effectively a 100mA 10V constant current source. That's how all this works.

I didn't read your speech beyond the first sentence, hope you had fun typing all that for nothing though, but let me reiterate; go fuck yourself.

 

[TimoTapani]

Something you could try sometime is hooking the arduino directly to the dim leads.

Yeah, maybe I try to see if that affects logarithmic response. On the long run it's not useful solution as I'm using Bridgelux chips, so running soft doesn't provide that much benefit. There's also other thing that could affect a bit, but wouldn't expect this much: I have independly tempadjusting fans and they speeding up with watts could raise my consumption, but they're total only 6W. It's wired a way that it's not easy to separate fans from the mains plug.

Are any of you others experiencing this behaviour that like "raising" PWM eg. from 10 to 5 (~5%) raises my wattage from below 140W to 150W (over 10%). It's quite normal below 90%, but those last 10% seem to control almost 25% of usable power region. I already formed a kind of theoretical filterfunction to sort it out, but would prefer to solve it on its roots than bandagefix.

Anyway, all you need to do 10V PWM dimming from an Arduino 5V PWM signal is a transistor and a resistor. Or use 2 transistors if you don't want the signal inverted. Otherwise 0 on the Arduino will be 100%, but you can also send signal=255-outputlevel to the arduino pin to invert the signal back.

The DIM wires are effectively a 100mA 10V constant current source. That's how all this works.

As being noob with electronics, I have to follow advices when I don't understand something (at the moment). While studying electronics I came across this Tom Jennings' rant "Do not TIP!" (sorry, can't post links) which confirmed me not to use transistors, if not neccesary. It told me that with any practical application I could substitute a transistor with a mosfet. Looking through datasheets over and over again, it could been too slow compared to special "signal transistor", but I still feel that there's also some other diffence in application principal, which I do not yet understand. (As it seems that DIM- is leaking somehow towards DIM+ if Mosfet is in off-state, or the Mosfet doesn't go into full off-state with driver connected for some strange reason. With DMM there's no leak from Drain to Source while off).

I did had optocouplers on hand and they worked atleast on satisfactory level, so I'm happy. Doing more tests later on. I actually have some spare motorshields for Arduino and got impression that I could drive the drivers with those also. (Is there some confirmed models or some which surely won't work?) For me it's not so big deal, since making all these mistakes have trained me quite good at soldering, even while saying it myself. But to educate myself even further, I'm doing a tutorial translated to a local forum and would be more than happy to offer a completely solderless solution (with precautions of loose connections ofc). You know "Those who can, they'll do. Those who can't, they'll teach"

 

[wietefras]

A MOSFET is also a transistor. So yes that works too.

But indeed you need to pick a component that fits the requirements. Not every mosfet, transistor or opto-coupler will do.

 

[VegasWinner]

I use a xd pwm motor controller. They use a 10v power supply and work great. I use them for my dmx light fixtures. Google 5v pwm to 0-10v dimmer. They are on ebay used for motor controls.