100W LED Cob bead

[martinw]

Doesn't work. If the PWM pulse is shorter than the switcher's minimum "on" time then you're stuffed. Most switchers run at 500KHz-1MHz so at 200Hz pwm the best you can hope for is 12 bit (4096 steps) and in real life it's even less. One trick is to dynamically alter the drive current of the switcher so once you get down to a low dim level, you switch to say 100mA output and then PWM that, however it is very hard to get a smooth changeover.

 

Most LED fixtures don't work like this. Feeding PWM into the DIM input of the switching controller falls down for the exact reasons you've stated. The way this is nomally done is to short the LED with a FET during the off part of the PWM cycle, which also helps in overcoming any parasitic capacitance in the LED. With the right microcontroller, you can get near to 1KHz at full 16 bit resolution, which combined with some additional software will give some acceptably smooth fades. See the LM3409 datasheet, section 8.3.8 for an example of this.

 

Martin

[timsabre]

Most LED fixtures don't work like this. Feeding PWM into the DIM input of the switching controller falls down for the exact reasons you've stated. The way this is nomally done is to short the LED with a FET during the off part of the PWM cycle, which also helps in overcoming any parasitic capacitance in the LED. With the right microcontroller, you can get near to 1KHz at full 16 bit resolution, which combined with some additional software will give some acceptably smooth fades. See the LM3409 datasheet, section 8.3.8 for an example of this.

 

That's interesting, thanks for the link... doesn't the fet cook if it's taking all the LED drive?

[martinw]

Most LED fixtures don't work like this. Feeding PWM into the DIM input of the switching controller falls down for the exact reasons you've stated. The way this is nomally done is to short the LED with a FET during the off part of the PWM cycle, which also helps in overcoming any parasitic capacitance in the LED. With the right microcontroller, you can get near to 1KHz at full 16 bit resolution, which combined with some additional software will give some acceptably smooth fades. See the LM3409 datasheet, section 8.3.8 for an example of this.

 

That's interesting, thanks for the link... doesn't the fet cook if it's taking all the LED drive?

 

No, because if you have a 350mA constant current driver, and you short the output with a FET, then the voltage drop across the output is close to zero. Because of V*I = power, this means that there is very little power dissipated in the FET. It's rather counter-intuitive, and at first glance, you might expect it to go bang, but it does work very well. For bigger LEDs, the TPS92641 is a good basis around which to build a PSU, and also provides a driver for the shunt FET.

 

Martin

[alistermorton]

Alister, yes that is right. Even a linear constant current regulator like an LM317 does not like being switched faster than about 250KHz.

 

Thanks, Tim.

 

Most LED fixtures don't work like this. Feeding PWM into the DIM input of the switching controller falls down for the exact reasons you've stated. The way this is nomally done is to short the LED with a FET during the off part of the PWM cycle, which also helps in overcoming any parasitic capacitance in the LED. With the right microcontroller, you can get near to 1KHz at full 16 bit resolution, which combined with some additional software will give some acceptably smooth fades. See the LM3409 datasheet, section 8.3.8 for an example of this.

 

Martin

 

Now that is interesting, Thanks.

[Brian]

For anyone thinking of doing there own thing, Artistic Licence published a couple of application notes detailing ideas for alternatives to straight PWM, including at least one that was inherently 'square law'.

[musht]

Originally published in response to the Color Kinetics Patent grab, seome of which will be beginning to expire, now seems to have disappeared from AL`s site ,

 

http://www.techref.info/web/prod/al/led-fm.php

 

https://www.mikrocontroller.net/attachment/207919/BitAngleModulation.pdf

 

Whilst on subject of drivers , how do these work?

 

0-100% on a triac driver, been around for a few years, information scant , Patent Pending for a long time..

 

http://protolight.com/products/rled-retro-500

[timsabre]

Whilst on subject of drivers , how do these work?

0-100% on a triac driver, been around for a few years, information scant , Patent Pending for a long time..

http://protolight.co.../rled-retro-500

 

I have no personal experience of them, but I suspect you would find that they do 0-100%, but excluding the range 0.1-20%

 

...although it does say "smooth dimming"... but without some sort of secondary power supply, it's impossible to operate in that lower range as there just isn't enough power in the dimmed waveform. Maybe they do 0-100% smoothly, but start at 25% dim

[musht]

 

...although it does say "smooth dimming"... but without some sort of secondary power supply, it's impossible to operate in that lower range as there just isn't enough power in the dimmed waveform. Maybe they do 0-100% smoothly, but start at 25% dim

 

http://www.cantousa.com/rled/ from the PDFs

 

• Dim all the way down with no flicker or bumps

• Dims completely out without flicker, pulse or

stepping

 

I`m just intrigued, its a heck of a tin for a 17W driver, wonder of they use some sort of super capacitor as quiesecnt source or live of the dribble though the snubber networks.

[richardash1981]

Maybe they do 0-100% smoothly, but start at 25% dim

That would be plausible. Depending on how much electronics you put in, could probably be less than 25% (but with some sort of dead band so that the electronics power up before the light comes on at all). The fan doesn't make me enthusiastic however.

[alistermorton]

Dim out is mentioned but what about back in after being off for an hour?

 

Maybe it has a battery which is constantly being recharged when power is applied.

[musht]

Fan is the worst possible idea for an install unit in a ceiling void, that`s for sure.

 

It`s claim is the Holy Grail of retrofit houselights, dim to and from 0%, this would be good if true, but they have been around since at least 2014, Patent Pending for 4 years...

 

Have a couple of Philips dom3t1c lampas that do to probably under 10% start and stop with very little backlash , but it ain`t starting on fresh air 0%.

[Gerry]

Would one of these:-

https://www.hiline-lighting.co.uk/gb/rgbrgbw-led-controllers/281-DMX-rgbw-led-controller-5060440711961.html

be suitable for dimming a 100W LED Cob bead?

The cob needs 24V to switch on and 30-32 V for full brightness.

Slightly concerned that the spec for the controller says "Constant Voltage type"

Cheers

Gerry

[timsabre]

No it would not, that controller is for LED tape. You need, as stated earlier, a constant current driver.

 

Check out this Meanwell ELG-150-54B

https://uk.rs-online...rivers/1034372/

 

That takes a PWM control signal for dimming, you can use an RGB tape controller such as your suggestion above to make this PWM signal from a DMX controller. Just need to put in a 1K resistor in place of the LED tape, you will then get a nice PWM signal at the LED output terminal.

 

Or something cheap on ebay like this

https://www.ebay.co....tm/182698728197

[Gerry]

Thanks for the conformation.

The person I had an online chat with said it was suitable, I had explained what I needed it for.

 

The Meanwell ELG-150-54B isn't suitable.

 

I need to be able to control it via DMX and the DMX drivers I have found , so far, only go up to 24V which is the turn on voltage for the LED Cob bead.

 

 

the Ebay item only handles up to 24V

 

Cheers

Gerry

[timsabre]

No you misunderstand me.

 

Forget the voltage of the LED, you need a constant current driver - that will automatically adjust its output voltage to drive the LED correctly.

 

So far as I know you cannot buy a DMX controlled constant current driver of that power - doesn't exist.

 

So you need to get a dimmable CC driver, such as the Meanwell one I suggested, and then make a DMX control system to dim it using its PWM control input.

A simple way to do this is to use an RGB tape controller like the ebay one I suggested - that's just making a control input for the CC driver, it isn't driving the LED.