[KC0PET]

don't mind me, i'm kinda cranky tonight. :P this whole conversation has gotten way more pedantic than it needs to be. a rough ballpark estimate of the needed led wattage is all that's needed to prove photonicz's claims are false. their prototype would have to be literally hundreds of times larger to get anywhere near what they're claiming, especially air cooled. with that heatsink they're not gonna get much more than 100w averaged output, so even if they had a tiny magic led that could do 700w/s pulses at 1/250s, you'd still have to wait several seconds between shots just to avoid frying it. closest i've gotten is a design i need to get around to prototyping that will run at around 2000w continuously with a 3" 5-sided led cube, but even that's going to need to be tethered to a pretty massive water cooler.

No problem and I am a little on the cranky side myself, so I'm sure I am not helping the situation. Pedantic can describe some of my ramblings pretty well, just ask my Wife (but she will not use that word ). Anyway it sounds like we may be on somewhat parallel paths. Just curious, have you homed in on LED emitter devices? I bought some Cree "chip on board" emitters a while back to play around with, but I won't be anywhere near 2000w even with all 10 that I bought.

[victormeldrew]

It is a red herring to think in terms of Watts for lighting. The use of Watts as a measure of the 'power' of a light is a hangover from incandescent bulbs and cannot be used in any way to compare other types of lighting, especially a directional light source. The SI unit for light output is the Candela and even that is less useful than the Lumen which is probably the most useful for direct comparison between various technologies. By far more useful in this case would be Lux, which is a measure of directional luminous intensity over a given area - it is most often used for flashlights where lighting a constrained area rather than lighting a volume.

Watts will only tell you the amount of power required to light the emitter and, given that each technology varies in output efficiency so vastly, it is utterly useless to compare a strobe with an LED.

[maltfalc]

It is a red herring to think in terms of Watts for lighting. The use of Watts as a measure of the 'power' of a light is a hangover from incandescent bulbs and cannot be used in any way to compare other types of lighting, especially a directional light source. The SI unit for light output is the Candela and even that is less useful than the Lumen which is probably the most useful for direct comparison between various technologies. By far more useful in this case would be Lux, which is a measure of directional luminous intensity over a given area - it is most often used for flashlights where lighting a constrained area rather than lighting a volume.

Watts will only tell you the amount of power required to light the emitter and, given that each technology varies in output efficiency so vastly, it is utterly useless to compare a strobe with an LED.

actually watts is the correct term to be using in this case because we're discussing things like how much power you can push through an led the size of photonicz's prototype, how many watts the heatsink can handle, how many watts the battery would need to put out, etc.. and when comparing a strobe to flash leds, you're looking at <50% additional lumen output from flash leds, so they could maybe get away with 500 w/s to match a 700w/s strobe, not a significant difference in this context. and since their prototype seems to use a standard cob led, i'm actually being extremely generous.

---------- Post added 08-31-17 at 09:22 PM ----------

No problem and I am a little on the cranky side myself, so I'm sure I am not helping the situation. Pedantic can describe some of my ramblings pretty well, just ask my Wife (but she will not use that word ). Anyway it sounds like we may be on somewhat parallel paths. Just curious, have you homed in on LED emitter devices? I bought some Cree "chip on board" emitters a while back to play around with, but I won't be anywhere near 2000w even with all 10 that I bought.

my main goals with my design are to make it as small as possible at a ridiculous wattage and for as cheap as possible. i'm using 20 100w cobs (cheap generics for now), 4 per side on 5 sides of a cube with mounting bracket,water cooling and electrical connections on the 6th side. that's 4 separate strings of 5 cobs which run at just over 30v per cob so i can run them directly off of rectified and filtered a/c with no driver, varying the output by switching on 1-4 strings.

[Class A]

my main goals with my design are to make it as small as possible at a ridiculous wattage and for as cheap as possible.

Cool.

Looks like you are going for emulating a "bare bulb" radiation pattern.

What kind of colour fidelity can you achieve with the COB LEDs you are using?
For the best fidelity, would it be necessary to attenuate the typical "blue" spike in the output spectrum, or otherwise modify the spectrum?
Would that be possible easily with filters (with some loss of overall power, of course)?

The combination of A/C and water cooling sounds exciting.

Do you need to resort to water cooling because you are aiming for continuous output?

[maltfalc]

Cool.

Looks like you are going for emulating a "bare bulb" radiation pattern.

What kind of colour fidelity can you achieve with the COB LEDs you are using?
For the best fidelity, would it be necessary to attenuate the typical "blue" spike in the output spectrum, or otherwise modify the spectrum?
Would that be possible easily with filters (with some loss of overall power, of course)?

The combination of A/C and water cooling sounds exciting.

Do you need to resort to water cooling because you are aiming for continuous output?

colour fidelity's not going to be great with $5 cobs. i may end up using a mix of cool and warm white cobs to fill in the spectrum a bit better, or i may stick with all cool white and use the excess blue to try out another project, gel filters using phosphors instead of dyes. instead of throwing away the extra blue light heating up a normal filter i'd be converting it to red and bluish green to improve colour fidelity or orange to shift the colour temp, etc. i'm going with water cooling because at 2000w it's the easiest, cheapest option to custom build and you don't have to mount the whole thing directly to the back of the leds at the top of a light stand, so you can scale it up as big as you want. if i really wanted to be cheap and lazy i could even just attach a couple of garden hoses to the led cube and it would work just fine.