50 Watt UV Light

[LarryDFW]

UV LED technology has advanced a lot in the last year . . .
Several manufacturers have chips which produces 1.15 watts of UV with an input of 700ma @ 3.7 V.
That comes out to 2.59 Watts of input power .
1.15 watts out/2.59 watts in = 41 % UV power.
50 watts in = 20.5 watts UV out (without focusing lens)


This light has been operated for 1 hour with a temp on the aluminum heatsink of only 125 F .

 

[Cyparagon]

Flawed thinking. Efficiency of state-of-the-art small-scale LED does NOT translate to chinese "let's pack 50 random dies on a plate and hope for the best". Even if LED Engin has managed to pack together 20+ of their own top-notch dies, your calculations do not translate to all LEDs. Even most of LED Engin's other UV dies are under 30%.

So unless you've got a datasheet for the actual LED you're using, I'm calling BS on 20W out. It's like saying "Lamborghini Diablo can go from 0-60mph in 2.4 seconds. My ford explorer is also a car. Therefore, my car is capable of similar acceleration."

 

[DrMario]

Also, considering the fact that the LED module you're using has multiple LED dices, there will be dark spots, so the luminosity efficiency drops inversely depending on how closely spaced the LED dices are to each others - a common complaints among the flashlight addicts. It actually affect Lux in whole.

Also, the UV LEDs will remain inefficient compared to special HID bulbs for a while, since the growth of special cut Gallium Nitride for sub-390nm lines is problematic so the manufacturers just sell those UV LEDs as-is, until they have the wriggling room, financial-wise, to be able to improve on them. This is one of few reasons the 365nm or even the 245nm germicidal LEDs are expensive. So expect 20% efficiency from wall to light, or even lower depending on growth methods used.

EDITED: If the LED remains in 125 °F range, that's fine but it obviously will get hotter after probably 1 hour and 30 minutes - are you sure you are using calibrated Infrared thermometer? Also, at 125 °F, some LEDs will shift towards longer wavelength. And optical output may be affected too. So I recommend retest this LED on the computer water cooler block so you get honest output on that LED because many LED manufacturers always test them cold (25 °C) and then hot (50 °C) so they can bin them properly. Lastly, don't deduct the LED output wattage from amperage consumption, there is always losses somewhere to be considered.

 

[DashApple]

Expensive 365nm leds , Been there : P

My 40W 365nm led Engine is specked at 5W optical output typical

 

[DrMario]

5 Watts UV output from 40 Watts of electrical power may sound right while cooled properly (with decent heatsink). Most LED manufacturers usually specify 15% -+ within measured total of optical wattage rating.

However, optical wattage output is obviously affected by the bandgap (voltage drop at the PN junction(s) in a given LED dice) and current consumption (some drivers allow adjustable ampere output). The only most efficient LED ever known is a 808nm laser diode rated beyond 2 Watts, due to narrow bandgap of Gallium Arsenide (the same could possibly be said for the Infrared Luxeon LEDs).

Off topic, although relevant, I have a Luminus Device SST50 LED - the brightness is also affected by the current consumption (amp vs volt at the LED dice). I have shoved 10+ Amps at 3.3 Volts from the junk PC power supply into SST50 LED (mounted on an Athlon XP heatsink I tapped to fit the LED starboard), and it was freaking bright.

 

[The Lightning Stalker]

The LEDEngin LZC-C0U600 were looking very
nice until further investigation into the
power binning. For my $340, I could be
getting an LED with as little as 2.4W of
output. That is still huge compared to the
competition, but wow that price!

 

[ped]

It's like saying "Lamborghini Diablo can go from 0-60mph in 2.4 seconds. My ford explorer is also a car. Therefore, my car is capable of similar acceleration."

I'm forced to agree with Cyp.

A crude but accurate analogy. Data, or it didn't happen.

Ped

 

[phenol]

according to its datasheet, the output at 70C of LZC-C0U600 collapses to 50% of what it does at 25c. the situation with the 4-die uv led is about the same as they are obviously using multiples of the same 365nm chip. For those who want to squeeze the last mw from the poor sucker, the heatsinking +thermal mngmt implications are dire. I bolted my 4-die ledengin led to a 400-gr fan-cooled Al heatsink so it stays below 40c at 23c ambient.

in contrast, the 400nm leds they have seem much less sensitive to heat-at 100c its flux drops to 60%.

 

[LarryDFW]

Here are a couple of low-output photos of earlier LED's in the same host:

41% efficient LED's which due to the application,

can be used without any lens .

 

[Cyparagon]

The only LEDs I see on their page that read 41% are 2.5W devices. 7 times 2.5 is not 50W input.

 

[rhd]

The only LEDs I see on their page that read 41% are 2.5W devices. 7 times 2.5 is not 50W input.

The problem, is that you're not familiar with China math.

Hop on eBay and search for "10w ir led emitter".

You'll be amazed at all the laws-of-physics-defying LEDs that not only don't account for efficiency losses, but in fact purport to provide 10W of output notwithstanding their draw of less than 10W of raw input. It's phenomenal.

 

[LarryDFW]

There are plenty of foundries that deliver high quality LED's .

Several LED manufacturers now has some UV bins that are over 50% efficient .

The seven in the photo above are 4 watt power input chips from an earlier design .

 

[Cyparagon]

So... you were questioned on the power and efficiency, so you posted a setup... and shortly after that you admitted was <30W. I can't tell if you're backpedaling, or outright lying about this whole thing.

 

[LarryDFW]

A high-efficiency UV produced by Nichia.

Efficiency of 48.6% . . .

http://www.nichia.co.jp/en/product/uvled.html

 

[diachi]

A high-efficiency UV produced by Nichia.

Efficiency of 48.6% . . .

http://www.nichia.co.jp/specification/catalogue/2014_02/NICHIA_LED_Catalogue2014_02_08.pdf

Are they the diodes you are using? If not - there is no use linking to them.

 

[LarryDFW]

"Nichia now has some UV bins that are over 50% efficient"

I am using some of these now.

Other LED manufacturers have some UV LED's with comparable efficiency.