LED Torch lumens = 85% of bulb lumens?

As compared to incandescent's established 65%?

I know the % loss is less than incandescent, but has the approximate amount been calculated?

Brightnorm

Don't think so.

We need someone with an IS to stick in a bare LED, then put the head back on and see what the difference is.

Something like the Regalight EDC would work because you can take the head off to expose the LED.

I think we are about there now that we are seeing some real IS measurements of some of our lights. Numbers are going well over 200 lumens at around 1A to the LED, some at 220 and 230 close to data sheet flux readings for Cree's, for example, if you subtract 15-20% from emitter flux at 1amp. See here: https://www.candlepowerforums.com/threads/211402 for some IS results. I think realistic number is 20% reduction considering losses from reflector, window, etc.

bill

This is very difficult to do for a LED that is already made into a pill. Even if we can take the reflector off the bare LED would have to be held up very carefully at the entry port of the sphere without any blockage from the flashlight host. I would have to clip leads to it from a power supply source. The host would get in the way if the reflector is not there to angle all the light into the port. I don't have a mount made for a LED module that is already made into a spring loaded drop in pill.

It would be a very interesting test. However, each reflector type would have to be tested separately. The smooth, orange peels, different shapes like for the Dereelight and Raidfire Spear throwers, not the same as the little ones in the D26 size pills. Could not assume one constant for all reflectors.

The only light I had that lent itself to do that was the maglight. With the head and reflector off and the bare LED could be shown directly into the entry port of the sphere just enough to get a peak reading while still attached to the main body of the light that I could hang on to.

After the Christmas holiday break I will look into getting this done.

Also once we get WBP's system fully calibrated and giving us good readings, we might be able to get more testing like this done. We are very close and he has some exciting data collected so far.

It will be interesting to see how much of the color spectrum gets lost by the reflector and glass windows as well (if any???).

The Regalight edc might be a good candidate for this test, it is very easy to remove the reflector and it has two different reflectors that it can be tested with.

It is not really possible to do measurement on the bare led, due to heat sink problems.

In MMrGman's IS tests
https://www.candlepowerforums.com/threads/211402
Mag-LED 4D______________50
Mag-LED 4D______________80 no reflector
That works out to 62%!!!

Someone compared the Fenix # with the spec and came up with 81%.

So it depends on how well built the light is rather than LED vs. Incan. I will stick with 65-80% unless someone actually done tests on the light I want to buy.
Light is light. Unless the lens is tinted the light transmission through it is the same.

Just dropping a AR coated UCL lens into a M@g makes a noticeable difference in output. Flashlightlens measured ~9% improving the ~65% to ~74% transmission. I have used them in my Minim@gs and full size m@gs.

That completed mag reading was with the cheapo plastic window and I couldn't get the entire beam into the port (which I have mentioned in the past numerous times) so I don't know how good the 50 lumen reading really is. I now have a couple of Ultra Clear or AR coated lenses for the maglights. If I could read the entire thing I might be able to get it up a little higher. So I wouldn't put the reflector at 38% loss.

It won't be sufficient to just remive the reflector and window from the flashlight as the empty head cavity will block light from the bare LED, giving false low readings.

I am very interested in this project and would donate some money in order to establilsh a good testing procedure and have the right, meaning the best components at our disposal.

For example ... I really would like to know how efficient TIRs really are. I am sure milkyspit has tons of SF TIRs of the current L1 lights lying around ... and could spare one for testing. McGizmo has his McR-xx line of reflectors that should be very good and available candidates.

We could also test various lenses ...

bernie

I believe I already addressed this above.

"It won't be sufficient to just remive the reflector and window from the flashlight as the empty head cavity will block light from the bare LED, giving false low readings"

Yes that would be a problem but we should be able to get around it.

I don't know if we should even try to come up with a general number. LEDninja has already shown how much it varies based on the build quality. However, if somebody were to work on a database with optical efficiencies for many flashlights, that would be really great!

I believe I already addressed this above.

Click to expand...

You didn't.


I don't think we need a database, this would be too much work I guess. Testing the best and the most common should give us an idea what is possible and what is the usual.

bernie

I am quite bored about that torch or bulb lumens issue. The only important thing is the comparable values, then if it's based on torch or bulb brightness doesn't matter. All manufacturers would use a standard so we could avoid that discussion about torch or bulb lumens.

Regards, Patric

That's funny, most of us reading and posting tend to do so in threads that don't bore us. Look a little more and you'll find some you like, but leave this one alone as you're likely to continue to find it boring. :wave:

Kiessling said:

I don't think we need a database, this would be too much work I guess. Testing the best and the most common should give us an idea what is possible and what is the usual.

bernie

Click to expand...

I agree. But I still think a database would be better, if not proportionally to the work put in. Which is why I'm suggesting somebody else do it, and not myself . Sorry...