Die luminance, emittance and advanced die heatsinking

[Al Combs]

Parallax has nothing to do with our problems here.

The other problem with doing it on the top so that it works for a reflector is that creating a reflector that close to the die means that in order for the silvered part to not block even part of the die from the reflector it would have to be so small as to not even matter if it was even possible at all(would depend on the reflector).

What you describe as "other problem" is a fair description of parallax.

I also have not seen anything approaching a 50% loss from removing a dome. Ever.

Well technically a 50% increase was what I said. That's the equivalent of a 33% loss. Osram lists in their pdf doc's the Ostar LE W E3B with the silicon dome as having an optical efficiency of 50 lm/W at 350 mA vs the Ostar LE W E3A that has an optical efficiency of 36 lm/W at 350 mA. That's an efficiency increase of 39% by adding the dome.

The top bin (WX) of the Luminus CBM-360 is 4,300 to 5,100 lumens @ 6.3 amps. The CBM-360 is output binned at full power. The CSM-360 WV bin is rated at 3,600 4,300 lumens @ 3.15 amps. On page 9 of the CSM-360 specs is a graph called, "Relative Luminous Flux vs. Forward Current". From that I extrapolated by pixel couting a 193.6% increase from 3.15 to 6.3 amps or 6,969.6 to 8324.8 lumens. That's an increase of 62% on the low side of the range vs 63% on the high side. The CBT-90 is binned at 9.0 amps and shows a ratio increase of 1.34 at 13.5 amps. The top (WR) bin is 1,750 to 2,100 lumens for an adjusted 13.5 amp output of 2,345 to 2,814 lumens. The CST-90 has the same top bin as the SST-90, namely the WN that is 1,000 to 1,200 lumens but at 3.15 amps. It's 13.5 amp multiplier is 340.8% for a total output of 3,408 to 4089.6 lumens. That's an output increase of 45%.

They do not change the phosphor based on whether it is domed or not.

Here is a picture you posted of an SST with and without the Wavien reflector. You commented on the "warming of the color". I think it's a reasonable inference that if a secondary fluorescence from the royal blue peak causes a color shift to the warm, the only way two LED's, both with and without a dome, could have the same color is by using a different phosphor. That's just an educated guess on my part. If you have inside information from the LED manufacturers that allows you to say that's not so, I'd love to hear it.

You are really getting off track here. The loss in lumens from removing the dome is due to an increase in internal reflection.

I don't think it's off track at all. Newbie correctly pointed out the test would have been more accurate if a single XR-E had its dome removed. But jtr1962's test results indicate a 30% increase in output for the normal XR-E over the XR-E with no dome. If you say that's nothing but internal reflection. I both disagree and ask that you explain that statement. Is it based on something other than your best guess? The CSM-360 being 60% brighter than the CBM-360 has to be due to something other than internal reflections.

You might possibly see some increase with an experiment like this but I wouldn't expect it to be much at all.

And that is just your guess. Which as I said I respectfully disagree with. For the sake of people who read these posts and just assume an unchallenged rebuttal is an admission of error.

Finally I have some personal experience of my own. In a failed experiment on two DSVNI P7's I bought from PhotonFanatic, removing the dome on one made it's output much dimmer than its twin(?) brother. The hotspot was much smaller but overall output when the reflector housing was removed from the MagLites I had them in, removed that advantage. It was both much dimmer and of a different color than its twin, after the removal of the dome.

 

[saabluster]

What you describe as "other problem" is a fair description of parallax.

Maybe you are not following my line of reasoning well enough. It would be best if I made a drawing i suppose but I can see no application of parallax in this situation.

Well technically a 50% increase was what I said. That's the equivalent of a 33% loss. Osram lists in their pdf doc's the Ostar LE W E3B with the silicon dome as having an optical efficiency of 50 lm/W at 350 mA vs the Ostar LE W E3A that has an optical efficiency of 36 lm/W at 350 mA. That's an efficiency increase of 39% by adding the dome.

Obviously by my response I assumed you were speaking of the lumen penalty in removing the dome not the benefit in adding a dome. Going in the other direction certainly helps your case but even then 39% is not 50%.

The top bin (WX) of the Luminus CBM-360 is 4,300 to 5,100 lumens @ 6.3 amps. The CBM-360 is output binned at full power. The CSM-360 WV bin is rated at 3,600 4,300 lumens @ 3.15 amps. On page 9 of the CSM-360 specs is a graph called, "Relative Luminous Flux vs. Forward Current". From that I extrapolated by pixel couting a 193.6% increase from 3.15 to 6.3 amps or 6,969.6 to 8324.8 lumens. That's an increase of 62% on the low side of the range vs 63% on the high side. The CBT-90 is binned at 9.0 amps and shows a ratio increase of 1.34 at 13.5 amps. The top (WR) bin is 1,750 to 2,100 lumens for an adjusted 13.5 amp output of 2,345 to 2,814 lumens. The CST-90 has the same top bin as the SST-90, namely the WN that is 1,000 to 1,200 lumens but at 3.15 amps. It's 13.5 amp multiplier is 340.8% for a total output of 3,408 to 4089.6 lumens. That's an output increase of 45%.

You can just throw all your figuring out the window on this one. First off you used two different LEDs with datasheets from different dates. The thermal resistance does not match between the two.
Second I am calculating roughly a 20-29% increase in adding the dome.

There is also one other thing you may be missing. The difference is not as great at low power densities but as power rises and the package is stressed to the limits the extra amount of light lost in the domeless package dies as heat and that heat will make the spread between the domed and domeless larger not the optical losses or gains in and of themselves.

Here is a picture you posted of an SST with and without the Wavien reflector. You commented on the "warming of the color". I think it's a reasonable inference that if a secondary fluorescence from the royal blue peak causes a color shift to the warm, the only way two LED's, both with and without a dome, could have the same color is by using a different phosphor. That's just an educated guess on my part. If you have inside information from the LED manufacturers that allows you to say that's not so, I'd love to hear it.

That is not a reasonable inference. The phosphor does not need to change to get a different color only the amount or percentages of phosphor used. Besides the effect seen from adding the collar is far more drastic than would be seen by just what bounces back from internal reflections due to no dome.

I don't think it's off track at all. Newbie correctly pointed out the test would have been more accurate if a single XR-E had its dome removed. But jtr1962's test results indicate a 30% increase in output for the normal XR-E over the XR-E with no dome.

I confirmed jtr1962's results in my own tests and I used only one LED. That said we are still at 30% not 50%.

If you say that's nothing but internal reflection. I both disagree and ask that you explain that statement. Is it based on something other than your best guess?

It is based on my knowledge of light extraction and most certainly is not a guess. The information is on the web for all who care to look for it.

And that is just your guess. Which as I said I respectfully disagree with. For the sake of people who read these posts and just assume an unchallenged rebuttal is an admission of error.

Finally I have some personal experience of my own. In a failed experiment on two DSVNI P7's I bought from PhotonFanatic, removing the dome on one made it's output much dimmer than its twin(?) brother. The hotspot was much smaller but overall output when the reflector housing was removed from the MagLites I had them in, removed that advantage. It was both much dimmer and of a different color than its twin, after the removal of the dome.

As far as whether or not the aluminum foil test would work or not...well yeah that is a guess. And a good one at that.

Showing me a picture of a chopped off P7 dome is hardly refuting my assertion that the foil will have little to no effect to increase surface brightness though.

 

[Al Combs]

...

I think this is the point where we'll have to agree to disagree. :wave:

 

[saabluster]

I think this is the point where we'll have to agree to disagree. :wave:

So be it.:wave:

 

[mpteach]

I'm making a led worklight and might use a tir or aspheric. There will be 3 groups, the large main group on top for ceiling bouncing to light up the whole room and weaker groups on the left and right sides with elliptical optics for occasionally shining down the walls to find flaws.The main beam has to be just wide enough of an angle to diffuse lots of lumens off a normal height ceiling without glare or afterimages in my eyes from the reflection. Also there should be almost no spill in that beam, too much light going sideways from the worklight will also cause glare in my eyes, especially when standing near the light.

I was originally thinking of 9 cree xml and 9 tir 16mm optics for the main array, with a 2" square piece of glass and a hinged shroud on top to absorb the spill.

Now im wondering if i could use one of those really big bridgelux leds and one big aspheric lens instead of the small xmls and tirs.

I want a large percentage of the beam to have even brightness but don't mind many small shadows or color spots since its just reflecting off the ceiling!
The smaller the diameter of the glass the less area i have to clean. This light will have cleaning supplies in it.

Ra how did you make a diy reflecting collar? Also how much did the color temperature shift?

 

[Pizeer]

Re: Die luminance (surface brightness, important for throw)

So you were testing Graphite crystals? How did that turn out?