The 5770 is the 70 CRI model, so just a regular good-brand power LED. The datasheet shows 2700, 3000, 4000, and 5000k 90CRI LEDs, but Mouser only has the 2700 and 3000k models in 90CRI.
Unless this is a recent addition to the product line or there have been any press releases on the subject recently, I’m inclined to agree with you Hannes.
EDIT: oops, and they are. Looks like the LED Engin products require you to drive the dies separately, I’m assuming that just hooking them all up in series or parallel does not give the desired result. So those are out. So just the pricey OSRAM, pricey Luminus, and nicely priced Nichia.
I’m still thinking of it, but it would be a very large initial investment to get the variety and quantity I would need. I need to build up some more disposable income before I jump in, so that worst case scenario I’m not deep financial s**t.
However, I do need to sell off some of my lights—that may provide enough float to run a small beta test and see if it works.
One thing to note too is that the highest quality photo strobes still aren’t all that high in CRI. They tend to be 80-85 CRI. And in movie making the really bright HMI lights used to light up sets are only 80 CRI.
Don’t get hung up on CRI. A candle is 100 CRI and yet if you try lighting your photography with them it’ll still look terrible.
I mean if it’s that important to you to have the best color, led’s still won’t render it all that well because of the nature of their output. Even with high CRI the color spectrum ends up being spiky and not smooth. Then you might be better off with photography fluorescent tubes if maximum color rendition is a priority. But just saying decent led’s will perform as well as photo strobes. The big determining factor in my opinion isn’t the CRI it’s the tint bin. If you can get a consistent color temp and tint bin you’d be pretty well set.
I only partly agree, tint is very important, I also don't care much for 100 cri warm white light except for cosy home-lighting. But 90 cri 5700K light will look stunning. It does not say much that in film and photo 80 cri is good enough -better colour rendering will not be captured/reproduced anyway with present-day techniques-, what matters is the real life experience, a multi-emitter light with these mystery luxeons will look like real daylight!
And that’s the point I’m trying to make. It may look amazing to your eyes, 100 CRI neutral white looks amazing but does it translate equally as well into your photographs? Not by a long shot. You can get by just like Hollywood and the professionals. They don’t harp on the CRI because they realize it’s just a small part of the equation.
If you had $500 to spend on your lighting, you could spend it on a couple very good specialty led’s and have crap for lighting, inadequate amounts but hey the CRI is good. Or you could invest that in a lot more decent lighting and light the photograph the way you want and make a simple photoshop filter to correct the slighty off color rendering. This usually means cutting down on the greens and boosting the deeper reds. Or build your light like some of the professional lights, they use normal neutral whites and then have a couple deep red led’s in there to help boost the reds, and maybe use gels to filter out some green. But the LED’s themselves are just commercial off the shelf inexpensive leds that are in the same tint bin.
The importance being the various lights you have match in tint more than CRI. That’s what people do with conventional xenon flash tubes and it’s worked quite well for the professionals for the past 100 years or so. And they didn’t even have photoshop to be able to correct slight color imbalances.
I lost the question in the OP a bit, sorry Hannes , I am personally interested in those 5790 leds for flashlight use. I am modding a headlamp right now for a friend who is going to do botanical excursions at night (weird hobby, but he's a good friend ) and this would be a nice led for that (it is going to be one of those 219b 92cri ones now)
I don’t mind spending 500 USD on a good light but it should not be a 80CRI 1000 lumen light. I am sure I can build one that is closer to 4000 lumen and with the Nichia LED I should be able to get to 90CRI.
18 Nichia NVSL219B R85 (90 CRI) 5015k will cost $100
I see there is a 18 x led t6 flashlight on eBay running from 5 26650. If I am lucky the driver will come available on KD which should not cost more than $30 max. If not I can still buy the flashlight for $150
I plan to use 5 Panasonic 18650b $70 max
I have a Imax b6 charger so charging them would not be a problem, Cost $40
Hosting should cost no more than $100 USD if I splash out on optics etc
Total $340
That is a far cry from what is on the market and should out preform what is available.
Only problem is that it would need some mod on the driver to prevent the flicker from effecting the cameras. But that should be easy to do
ehhhhh. For better or worse, I’m not sure you’re fully considering Mr_Krabs point. Or maybe you simply think his point is invalid.
Anyway for drivers have you considered the TaskLED Hyperboost That seems like a good option for photo hotlights since you can control it with a trimpot. I’d much rather have one of their drivers than an expensive one from KD.
I don’t know what level you intend to drive these lights at. Since the Hyperboost handles up to 3.2A you are looking at maybe 190 watts of output power at if you run the thing wide open. That’s a LOT of heat. The hyperboost might be like 80-85% efficient doing that, so like 30 watts of heat there, plus half the wattage going to the LED comes out as heat I think, so over 100w of heat. You’d want a couple of fans built into your stick. Also you might get like 10 minutes of runtime at full blast with the batteries you listed.
You are looking at boosting to approximately 2x the input voltage no matter what power level you intend to drive the LEDs at, if you could get them closer together that would probably be a good thing. Have you considered splitting this up into multiple sticks that fit together? That might increase your flexibility in terms of using the thing and you wouldn’t have to use such exotic drivers.
EDIT: went back and realized that you are talking about 4000 lumens, so maybe more like 60w of power to the LEDs. Same high voltage needed (nearly 60v) to drive them, so the Hyperboost is still a great choice I think. Maybe 88-89% efficient in that situation, so still a little heat involved from the driver and the LEDs. So maybe a 45min runtime or so, not counting the smallish current of a fan or two?
Wow thanks a million for that driver. I would say it is perfect for what I plan. I hear what Mr_Krabs is saying and yes he is correct for normal photo studio work. But I have a need for mobile window lighting which leds are better suited for.
Sure thing. I have no personal experience with the TaskLED drivers, mostly due to their size (big for a flashlight) and also partly due to their price. I have seen nothing but positive opinions on them however, so I have plenty of confidence in that recommendation. Based on the amount of test info posted on the TaskLED site and elsewhere I really feel like George knows his stuff.
Please keep us posted on your project, it would be great to hear and see about it!
Well I just bought 2. And 40 Cutter-NVSL20-NW90 leds.
Net challenge will be to design the light stick to have a magnetic pot so that I can vary the intensity using one hand.
Here is a rough sketch of what I plan to do. When I am back at my office next week I will make a 3d module. The light stick shine it's ligh to the side.
At first I was planning on putting the 18650 behind the leds. But I realized that heat will be a issue running the leds at 1A. 20 LED will consume 60W of power. Of which most will become heat.
So I'm going to go for 32650. 5000mAh cells. They will fit my 40mm diameter aluminum tube well.
With them I should be able to maintain max output for 50min. And based on the data sheets the leds should produce about 8000 lumen. Yes yes I know the real output should be about 6000 lumen. That is still a lot more than what is commercially available. 8x to be exact.
EDIT: I see that in your sketch you have the driver far away from the magnetic pot, that is good. Another important note here is that there is an absolute maximum of 7A on the input side. The tech doc recommends 6A or less. If you plan to output 60W+ w/ a driver that is 90% or less efficient, you should really plan for having over 3 cells. Again, I’m not actually sure that quality 32650’s are available currently? Maybe consider some good 26650 cells.