Mixing LED colors circuitry.

[videoman]

I have a total of 14 leds ( 8 cool whites and 6 red ones) that I wish to put in a lamp fixture, adequately heat sinked. I wish to vary the color temperature to swing between 3000K to 6000K with a pot to blend the blues with the reds to achieve the desired tint. Also,once the desired tint is achieved, I wish to have dimming capabilities for the resultant color output. I presently have buck drivers with dimming functionality. I wish to use XPG's as well as XPE's if need be. Is there an alternative method, different tint bins, number of leds, etc. that can be used.It is really a led light with dialed in color temp with dimming.

 

[snarfer]

You might be able to do it with op-amps, but I think it would be pretty tricky.

It's probably a lot easier with a microprocessor.

Thermal feedback would be a very good idea, and optical feedback would help keep the color constant during dimming.

You might look into LEDs like the Uno-Plus Dome White HiCRI from ENFIS.

They will sell you the driver and everything.

Be ready to pay through the nose.

 

[uk_caver]

Presumably, if dimming was via a common PWM drive (so the actual operating current for white and red strings was constant, just with a varying duty cycle), the colour balance should stay pretty consistent while dimming?

 

[snarfer]

Presumably, if dimming was via a common PWM drive (so the actual operating current for white and red strings was constant, just with a varying duty cycle), the colour balance should stay pretty consistent while dimming?

That would be nice. But actually red LEDs shift in color mostly in terms of temperature, not momentary current. So driving them with a PWM signal would not only be less efficient than CC, but also result in increased color shift. Furthermore the changes in color and output of the red LEDs would be unlikely to match the changes in color and output of the white LEDs so the proportion of emitted white light to emitted red light would vary over the dimming range.

Also, the application is video lighting, and the PWM signal could cause flicker visible only to the camera.

 

[Archie Cruz]

Nice to see more and more folks using our concept to dial in targeted CRI/Color Temp! What you are constructing will be 90% of the ideal video light. Let me know if you need patent help:twothumbs or the other 10%:twothumbs BTW. There is an RGB die out there, forgot who makes it. should make the whole blending process easier.

I have a total of 14 leds ( 8 cool whites and 6 red ones) that I wish to put in a lamp fixture, adequately heat sinked. I wish to vary the color temperature to swing between 3000K to 6000K with a pot to blend the blues with the reds to achieve the desired tint. Also,once the desired tint is achieved, I wish to have dimming capabilities for the resultant color output. I presently have buck drivers with dimming functionality. I wish to use XPG's as well as XPE's if need be. Is there an alternative method, different tint bins, number of leds, etc. that can be used.It is really a led light with dialed in color temp with dimming.

 

[blasterman]

Another problem here is that cool white + red results in a pinkish looking light.

I've tried this in an attempt to get warmer / higher CRI colors in mixing boxes - you actually need some amber in there along with the the red / cool whites to fill in the gaps.

 

[uk_caver]

Also, the application is video lighting, and the PWM signal could cause flicker visible only to the camera.

Ah - I'd read the post, but not the username, so I'd missed that.

How high a frequency would be needed to avoid flicker effects in video work, if artificial light tended to mean that actual/electronic shutter speeds were relatively long?
(At relatively low light levels, do video cameras tend to expose each frame for something like the maximum possible time?)

That would be nice. But actually red LEDs shift in color mostly in terms of temperature, not momentary current. So driving them with a PWM signal would not only be less efficient than CC, but also result in increased color shift. Furthermore the changes in color and output of the red LEDs would be unlikely to match the changes in color and output of the white LEDs so the proportion of emitted white light to emitted red light would vary over the dimming range.

Bearing in mind that PWM may well not be useful anyway, I was also probably a bit brief/unclear in what I suggested.

I was thinking more that if the red/white balance at full power was set to a desired level, then dimming from full power via PWM would keep the instantaneous currents in the various LEDs constant, so differences between the red and white LEDs (or even between different brands of white or red LEDS) in the shape of their instantaneous power/output curves wouldn't cause the red/white colour balance to shift with changes in average power.