Test of a 365 nm LED on ebay (powerful!)

[colinbm]

Cadmium (Steve) sent me a Convoy S2+ torch with his U-340 filter to do a comparison with the LED I have. To have fair results, I filtered my LED with a 2 mm thick ZWB2 and a 3 mm thick ZWB1. WB set on the unfiltered LED. The battery in the torch was fully charged, and I ran the LED at 1.3 A. LED on the left, torch on the right.

 

That was very charitable of you Steve, good on you...

[Stefano]

I would like to see the two LED's compared along, no lens, no reflectors, just the LED's, and maybe with only the filters added.

Just an idea, not important or required.

Ok, so I can do it with my 10 W LED, but what about the torch? Do you want me to use it without reflector and filter but with the LED assembly still inside the head or do you want a comparison between the LED assembly alone and my LED? If I leave the LED assembly inside the torch, the lateral portions of light will be cut out by the housing itself and I would get a cone of light, but I would lose output.

[Cadmium]

Remove everything from both LED's and try to compare them shining at some surface from the same distance.

Yes, the Nichia is deep inside the Convoy, so either make a similar tub around your LED, or remove everything and hot wire the Nichia to the battery.

It's just an idea, I don't know how easy it is to do... but get rid of the reflectors and lenses, then compare the light intensity.

[Stefano]

So, I removed the lens from my 10 W LED, I unscrewed the LED assembly out of the torch and I powered it with my lithium-ion battery, charged at 4.08 V, with the help of magnets and a ferromagnetic wire. Both LEDs were at ~7 cm from the paper. Both of them without any filter. Note that the Nichia LED inside the Convoy has a little lens, and of course I didn't even think about removing it.

 

Images at different exposures. Convoy LED on the left, 10 W LED on the right.

 

 

 

 

 

Then, out of curiosity, I ran the torch with my power supply, set at 4.2 V. It uses 0.7 A, so the power consumption is 2.94 W. Keep in mind that in my test above the battery wasn't 100% charged, but still the difference is quite evident.

[dabateman]

Cool test. So your 10W led now looks to be the champ, about 1 or 2 orders of magnitude brighter.

So the reflector cone inside the convoy really does help.

Keep that cone away from your 10W led. That might be too dangerous.

[Stefano]

My LED is at least 3 times brighter, probably 4 and probably not more than 5. That reflector would be too small for my LED, and the beam would diverge rapidly. I have a bigger one, but it would make a beam much larger than the Convoy one. Using one lens I got a 5º beam, which is not that bad (40 cm wide spot 4.6 m away). If I use two lenses I obtain that super high intensity burning spot, but it diverges very rapidly.

[Cadmium]

Stefano, Nice test, thanks!

[enricosavazzi]

Reading this thread, I was tempted to buy one of these UV LEDs to convert the mercury arc illuminator of my microscope (Olympus BX 50 with 100 W mercury arc burner) to LED for epifluorescence. The Olympus burner only lasts 200 hours and is quite expensive these days (roughly ten times more than a UV LED). Switching the burner on and waiting for it to warm up eats away half an hour to one hour of lifetime even for short epifluorescence sessions.

 

My idea is to build a LED support that replaces the burner and keeps using the illuminator optics and centering mechanisms. The modification should be quickly reversible for the rare occasions when a mercury burner may still be required. There is ample space within the casing for a heatsink at the rear of the LED. A small external power supply (or initially a small lab power supply) will be required for the LED, of course. The burner currently uses an external power supply roughly 30 by 20 by 15 cm, and replacing it with a small LED supply attached to the microscope would free up valuable real estate from my microscope desk.

 

My rough estimate is that at least 75% of the emission of the original burner (which is omnidirectional) is wasted within the illuminator casing and never reaches the collector optics. There is a concave mirror at the rear of the casing to reflect radiation back to the arc, which gives a modest improvement in efficiency. A LED with a roughly lambertian emission profile should be several times more efficient.

 

3 W emission at 365 nm might however be excessive for the job. I have no idea how much UV at 365 nm a 100 W mercury arc lamp emits, and it probably varies with pressure within the lamp and other parameters. Then a mercury arc emits lots of lines at other wavelengths, while the LED just emits a narrow-ish 10-20 nm peak centered around 365 nm, so the LED might be unable to excite certain dies (but the amount of shorter wavelengths transmitted by the microscope optics is perhaps very low in any case, so it might not matter).

 

A 365 nm LED replaceable with 380 nm, blue, cyan and perhaps green power LEDs might be more than versatile enough, and a 1W emission at 365 nm might be enough. Jury-rigging a 365 nm Convoy torch to shine into the collector optics of the illuminator could be a first test.

[Andy Perrin]

Enrico, the LED intensity should be more or less proportional to the current, so you could just drive it at a lower current (which may also extend the lifetime of the LED).