Sourcing LED Components for High-Wattage COB Array (1200w - Woah)

[MPTESE]

TL;DR Summary

I want to power (4) 300w LED Cobs simultaneously, probably each with their own LED Driver. I need help choosing the right drivers and also dimming technology for flicker-free camera application.

Hello,

I am looking to build an LED Panel for a film/video application. I have a few specific requirements in mind.

1) Color Accuracy beyond general CRI ratings.

2) Flicker Free to Camera (What specifications in LED drivers should I be looking for to ensure this flicker free application?)

3) Dimming Functionality (Ignorant to which dimming technology will provide the best results)

4) COB LED Array to maximize output and minimize form-factor.

5) High-Wattage Application. I am attempting to power (4) 300W (max) COB LEDs.

Parts:

I am considering this LED Cob array. To power 4 arrays simultaneously, I doubt I will be able to find one 1200w Driver with 120v power input. I was thinking each array can be powered by this driver. Since it's advertised for "LED Stage Lighting", perhaps my chances of it performing flicker free at higher FPS are increased? I doubt the manufacturer tests for such a specific application - Is there a way to ensure flicker free operation at High Camera FPS without testing myself? (ie What specs should I be looking for?)

I am still researching lensing, heatsink, and active cooling systems (I imagine I will need a fan). I'm hoping to have the (4) LEDs within a 2'x2' surface area.

Can anyone provide deeper insight to seeing if not only the driver's can be flicker-free, but also comparing/contrasting dimming controllability for all (4) LEDs simultaneously? Is there a dimming technology that would work better for my application? (PWM vs Potentiometer, etc).

I am thankful for the resources we have over the internet and I don't take anyone's knowledge or help for granted. Your guidance is truly appreciated in helping me learn myself and begin to develop a fixture hat will help my industry!

 

[anorlunda]

Wow, those are impressive lights. I'm sure some of our engineers can help you with the drivers.

That brochure makes it sound like the modules are sold as components to be integrated into an off the shelf product. Are you sure you want to design your own? Might there be finished products already designed using these components? Ask Luminus.

I did notice one thing in the brochure that you need to be very careful with.

Static Electricity Luminus
COBs are electronic devices which can be damaged by electrostatic discharge (ESD). Please use appropriate measures to assure the devices do not experience ESD during their handling and or storage. ESD protection guidelines should be used at all times when working with Luminus COBs

 

[hutchphd]

Summary:: I want to power (4) 300w LED Cobs simultaneously, probably each with their own LED Driver. I need help choosing the right drivers and also dimming technology for flicker-free camera application.

1) Color Accuracy beyond general CRI ratings.

2) Flicker Free to Camera (What specifications in LED drivers should I be looking for to ensure this flicker free application?)

3) Dimming Functionality (Ignorant to which dimming technology will provide the best results)

I do not know what (1) means. The color (spectrum) of these LEDs (for a designated color temperature type) is what it is. It is slightly T dependent.
Using 4 individual drivers for four segments will improve (2) (I do not think they are constant frequency drivers so will not be in step)
What are your dimming modalities and requirements?

You will need to pay attention to designing the heat control at the LED plate and a fan is likely a good idea

 

[MPTESE]

Hi Hutchphd,

My dimming requirements are to achieve smooth dimming without causing any flicker appear on camera. I can compromise a bit on the "smooth", and have it dim in "steps" if that's what needs to happen. But I cannot compromise the LEDs appearing as flickering when recording at High Frames-Per-Seconds on a camera. Please correct me if I'm wrong, but the way to ensure flicker free operation at High FPS is to not use PWM dimming and utilize DC current regulation somehow?

 

[hutchphd]

DC current regulation would be one method but if you use a linear regulator it is equivalent to using a series resistor in terms of heat produced. The spec on the supply you mention is 5% ripple in current at max output which is light power output for LED. I assume you can get that number as small as you want (at the price of $$ and size of unit)
I have not done the high speed photography thing with LEDs but they will track the current up to MHz. What is the expected maximum frame rate you want? How much light variation is too much (flicker-free is not a spec!)? Maybe somebody else can help.
The dimming appears to be continously (analog) adjustable.

 

[MPTESE]

I apologize if I was vague, as I shouldn't have been. Up to 2,000FPS is fine to maintain flicker free operation, at any shutter. Another assumed characteristic of dimming that is required is little change in color-shift. Obviously, also, if one dimming technology is more efficient as far as preserving light output, then that is a favorable characteristic as well. Thank you for taking the time to look at the specs of my proposed driver. What dimming solutions are affected by the MHz rating? From another source, it seems that frequencies in the 25,000Hz range is a safer area to be in for high FPS camera application . Maybe I missed it in the sheet, but should the LED Driver data sheet have a Hz rating for each dimming technology it accepts?

 

[hutchphd]

I was just mentioning the LED spec because, unlike filament bulbs, LEDs will respond to any current fluctuation in microseconds (not a good thing for you). I think those type dimmer circuits run (meaning the 5% ripple) at various frequencies depending upon load (maybe ~1 kHz) and I think you might see it at commensurate frame rates. You might build up one test unit (single supply) to check it out. I do think they will flicker independently which should help.

That's what I would do if I couldn't find someone with direct experience...sorry I don't know.

 

[berkeman]

I've only read a little bit in this thread, and haven't looked at any of the datasheets so far (sorry). But consider using a PLL circuit to synchronize the LED PWM drive circuit (at much higher frequency) to the camera's shutter signal. Does your camera have a shutter trigger output?

 

[Tom.G]

Depending on the particular sensor you use, a simple PWM control could need to be faster than the pixel video rate. This would allow each pixel to average the light intensity at each instant. If the sensor readout is a frame at a time then synchronizing a PWM to the frame rate could work... providing tha AGC (Automatic Gain Control/Auto Brightness) is fast enough to work on individual frames. If the sensor readout is a line at a time, you should look for a different sensor.

A sensible work-around could be to use some electrical filtering between the PWM controller and the LEDs. A quite large capacitor across the controller output is a possibility. Another, maybe more practical, approach is an inductor in series with the LEDs. Either would smooth out the power supplied to the LEDs.

Overall, the best tradeoff seems to be a frame readout sensor with series inductance from a PWM controller.

Sounds interesting. Please let us know how this project progresses!

Cheers,
Tom

 

[hutchphd]

OK I just noticed that they do spec Power Factor Correction and Pulse Width Modulation frequencies (45kHz and 100kHz respectively) on the block diagram. I think these will likely work fine even at 2000 frames per second. Personally, I would still build up one channel (300W) and test it before going whole hog.
It also occurs to me that the internal cooling blocks for desktop microprocessors might work well for the COB array...right size and heat dissipation.

 

[MPTESE]

Yes, I do plan to test one prior to building all.

Thank you for verifying the PFC and PWM setting.

To Tom's point, wouldn't a simple analog potentiometer be a safer option? By using PWM, wouldn't it introduce a higher chance of there being flicker, due to the nature of its design?

Concerning Camera settings, the goal is to avoid flicker regardless of camera, as most professional lights in the industry do. (most cameras fall within a certain spec.. The sensor specs Tom was referencing I cannot speak to).

Thank you also for the suggestion about cooling blocks. My instincts tell me to go with something that requires a fan, but the kick is to get one that is whisper quiet.

 

[hutchphd]

An analog (linear) supply is certainly less likely to produce any flicker. The price you pay is efficiency (more heat generation), size, and cost of the supply.
There are too many camera variables to really know. If the LED works it will be smaller and cheaper. The color spectrum will also not be quite as good as tungsten probably (you should research the two spectra). But testing it is a good exercise and worth the effort. I think this will work. Keep reporting...I will learn stuff, guaranteed.

 

[MPTESE]

I will start with the variable potentiometer with testing and report back. A potentiometer is certainly the most inexpensive option. The color spectrum specs are listed, and are within the tolerable ranges for high quality light. The main things to look out for are R9 values (skintones) and R12 values. Those values are especially hard for an LED manufacturer to produce. The phosphor in them is what helps make it appear as daylight. Tungsten is perfect, literally. HMI lights usually live around 90. Older LEDs had green spikes, that's mostly done away with - You can tell by looking at the spectral curve. CRI is a poor way to ultimately judge the quality, especially by itself. There's TLCI, and a couple TM30 tests to look at. The latter being more in-depth and providing more accurate results.

 

[Tom.G]

To Tom's point, wouldn't a simple analog potentiometer be a safer option?

And simpler. Just be aware that at half power the potentiometer will dissipate as much heat as the LED(s) it is driving. If you are thinking of using a small pot driving some transistors, then the transistors are dissipating the heat.

By using PWM, wouldn't it introduce a higher chance of there being flicker, due to the nature of its design?

Yes. That was the reason for using a series inductor, to smooth out the current to the LED(s). Remember that an inductor resists a change in current thru it. This would effectively average the PWM pulses.

Cheers,
Tom

p.s. I'm not trying to push any particular approach here... just trying to make you aware of some options and trade-offs. If you are building just one, the actual design doesn't matter as much as making it usable!

 

[hutchphd]

And simpler. Just be aware that at half power the potentiometer will dissipate as much heat as the LED(s) it is driving. If you are thinking of using a small pot driving some transistors, then the transistors are dissipating the heat.

This is certainly so. And the disclaimer that follows is important.

I'm a little concerned that everyone needs to be on the same page here. There are two pieces:

1. driving the LEDs with DC power using a controlled current
2. Making that current variable

Both pieces are necessary, and each is capable of introducing flicker. A potentiometer by itself
is not a solution. When applied to a controlled current source it will not perform well. Instead one would need to use it with a DC voltage source of capacity to drive the COB voltage with some overhead. At some point there you are better off using Halogen Filament bulbs on AC.
So we need to consider the entire system, and for me be more exact about what we are discussing...

 

[Tom.G]

At some point there you are better off using Halogen Filament bulbs on AC.

Some high-end studios use DC to drive their incandescent lamps. With many lamps running, there is enough acoustical hum from the filaments that the microphones pick it up.

 

[MPTESE]

Hello all, and I appreciate everyone chiming in.

Some update on my end..

I received a sample heatsink and am comparing my options. One of the heatsinks advertise compatible mounting holes for the MeanWell HBC series (a different driver than what I originally proposed). This is attractive to me due to the smaller form factor, and ability to mount directly on the back of the heatsink. However, the voltage maximum for the HGB is higher than the maximum rated voltage for the cob (58 vs 55). Here's the driver data sheet https://www.meanwell-web.com/content/files/pdfs/productPdfs/MW/HBGC-300/HBGC-300-H-DA.pdf

And the COBs again for reference
https://download.luminus.com/datasheets/Luminus_CXM-32_GEN4_Datasheet.pdf

To ensure I do not overclock the voltage, how do I calculate the correct device to prevent frying the COB? Is it as simple as attaching a 3OHM resistor in line between the driver and COB?

 

[hutchphd]

Perhaps you should explain your design intent and analysis for the 3 ohm resister.

 

[Tom.G]

After inspecting the datasheets for the LED and the MEAN WELL controllers, the only configuration that seems reasonable is to have one controller drive two LEDs. This could the -M- driving two LEDs in series (preferred method), or the -H- series driving two LEDs in parallel (not recommended).

I strongly suggest you contact Luminus and ask them for controller recommendations.

Cheers,
Tom

 

[rbelli1]

Another assumed characteristic of dimming that is required is little change in color-shift.

The pulsed current will give you the best color stability as you can run the LED at one current from full brightness to just above off.

With a linear supply or adding lots of low pass filtering to a PWM dimming supply you will get a color shift as shown on page 6 lower right graph of the datasheet.

You also get a color shift with temperature as shown in the lower left graph.

BoB