Die luminance, emittance and advanced die heatsinking

[Dr.Jones]

For bare throw, the most important factors are lens diameter and LED die luminance (usually referred to as surface brightness) - and then of course all the losses and the measures to reduce them.
While the optics part is covered in other threads, I couldn't find much info on die luminance. Is there any list?

If not, maybe we could compile one.
One could try to deduce that from the specs, but the specified LED flux (lumen) includes light strayed within the LED, which does not contribute to throw, thus this method over-estimates luminance.

The best method seems to be Ra's proposal: Measure the LEDs beam intensity (maximum) with a lux meter at 1m distance, then divide that by the apparent(!) die size as seen from ahead.

Measuring apparent die size might not be that simple, so I measured some typical LEDs with 1-2 methods:
a) measuring the width using an 100x low-NA microscope with scale and subtracting the contact pads (not good for XP-G, since the apparent area isn't a square any more)
b) taking a close-up image from the LED and counting pixels.

With the two XR-Es I applied both, and the results were within 3%

Apparent die sizes:
XR-E EZ900: 2.07mm²
XR-E EZ1000: 2.76mm²
XP-G: 3.50mm²

For luminance in cd/mm² just divide your measured beam intensity (it's proper unit is candela (cd), equivalent to lux@1m) and divide it by that area.
For standard units (cd/m²) multiply by 1 000 000.

Maybe we can collect some...

EDIT: Since gathering that data currently doesn't go that well, I'll post some die emittance (lumen per mm²) data meanwhile that I derived from the specs.
Theoretically the luminance is the emittance divided by pi, but as I wrote above this doesn't account for emitter package internal stray light.

The XR-E EZ1000 is reported to perform slightly (1-2%) better than stated at 1A.

 

[Walterk]

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

What do you mean with apparant size.... is that as how it looks to the eye when looking at the led ?
As opposed to cutting the led in half and measuring the actual size?
(Beacuse then you could also use the data-sheets dimension sketch.)

 

[Ra]

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

What do you mean with apparant size.... is that as how it looks to the eye when looking at the led ?
As opposed to cutting the led in half and measuring the actual size?
(Beacuse then you could also use the data-sheets dimension sketch.)

That's exactly what is ment: The dome already acts as a kind of precollimating lens, so you need the apparent die-size. After all, it is the apparent surface that creates a certain amount of lux@1meter..

That also was one of the major conclusions we ended up with in the "formula for calculating throw"-thread..


Regads,

Ra.

 

[Ra]

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

For bare throw, the most important factors are lens diameter and LED die luminance (usually referred to as surface brightness) - and then of course all the losses and the measures to reduce them.
While the optics part is covered in other threads, I couldn't find much info on die luminance. Is there any list?

If not, maybe we could compile one.
One could try to deduce that from the specs, but the specified LED flux (lumen) includes light strayed within the LED, which does not contribute to throw, thus this method over-estimates luminance.

The best method seems to be Ra's proposal: Measure the LEDs beam intensity (maximum) with a lux meter at 1m distance, then divide that by the apparent(!) die size as seen from ahead.

Measuring apparent die size might not be that simple, so I measured some typical LEDs with 1-2 methods:
a) measuring the width using an 100x low-NA microscope with scale and subtracting the contact pads (not good for XP-G, since the apparent area isn't a square any more)
b) taking a close-up image from the LED and counting pixels.

With the two XR-Es I applied both, and the results were within 3%

Apparent die sizes:
XR-E EZ900: 2.07mm²
XR-E EZ1000: 2.76mm²
XP-G: 3.50mm²

For luminance in cd/mm² just divide your measured beam intensity (it's proper unit is candela (cd), equivalent to lux@1m) and divide it by that area.
For standard units (cd/m²) multiply by 1 000 000.

Maybe we can collect some...

Oh yes, we can !!!

You have found quite a nice and accurate way to measure die sizes.. That's something we can work with !
I just received some led's, hope to be able to measure them soon..

I'm looking into the relationship between the several led-constructions and their overdrivabillity, which also is directly related to surface brightness.

Example: I still do not know why Cree choose to a base of aluminium-oxide.. Which has much lower thermal conduction than aluminium, let alone copper.

Some numbers on thermal conductivity: In W/mK

Al-oxide : 30
Aluminium : 250
Copper : 401
Silver : 429

And to make a remark that certainly will raise some eyebrouws:

I'm testing led's right now with a revolutionary heatsink material, with a thermal conductivity of.... 2000 ! Yep, that's a two with three zero's !!


Laterrrr..

Ra.

 

[Th232]

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

Some numbers on thermal conductivity: In W/mK

Al-oxide : 30
Aluminium : 250
Copper : 401
Silver : 429

And to make a remark that certainly will raise some eyebrouws:

I'm testing led's right now with a revolutionary heatsink material, with a thermal conductivity of.... 2000 ! Yep, that's a two with three zero's !!

I think diamond has a thermal conductivity about 5x that of copper, have you found a block of diamond?

 

[Ra]

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

I think diamond has a thermal conductivity about 5x that of copper, have you found a block of diamond?

Uhh... If this was a quiz, you would have been excluded from it..

You're right !! Diamond... CVD or industrial diamond to be precise..
It's a disc, 20mm diameter and 1.2mm thick..
The idea is to spread the heat from the led over a bigger surface, and then take it away with a copper heatsink..
That means that I can overdrive the led much more than with a conventional heatsink..


Regards,

Ra.

 

[Th232]

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

That is very neat!

I dread to ask, and I understand if you don't want to say, but how much did it set you back?

 

[Ra]

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

That is very neat!

I dread to ask, and I understand if you don't want to say, but how much did it set you back?

Almost nothing!! A friend of mine could get his hands on it for me.. And in return, I'm going to build him a telescope..

To all:

Note that this is a one time oppertunity for me !! So I must say no beforehand to any questions that start with: "I do not dare to ask, but is there a possibillity that...." !! Sorry...


Regards,

Ra.

 

[Th232]

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

Good trade!

 

[saabluster]

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

The best method seems to be Ra's proposal: Measure the LEDs beam intensity (maximum) with a lux meter at 1m distance, then divide that by the apparent(!) die size as seen from ahead.

Although you may have inferred it you did not directly state. Are you saying to use a lens to project an image of the die or not? If you do not use a lens the project an image then it seems to me all conclusions would be faulty. I have been doing this testing for some time although I have a specific distance I like to test at and I see no need to do any dividing but maybe I am missing something. Assuming you are indeed using a projected image of the die I have one little fly in your ointment. The luminance varies across the surface significantly. If you take the peak number and go with that you may be skewing the results.

You need super high quality lenses for this to really work. I found that some specialty projector parts that have 5 lenses in it which have extremely low aberration and allow me to get very clean measurements.
Interestingly the Nichia Raijin warm white has a higher surface brightness than the XP-E R3 at lower output levels.

Another means of measuring surface brightness would be to take images of the LED itself. I did not make this up. I read it somewhere but I don't remember where anymore. I'm sure it would require some specialized equipment though.

Uhh... If this was a quiz, you would have been excluded from it..

You're right !! Diamond... CVD or industrial diamond to be precise..
It's a disc, 20mm diameter and 1.2mm thick..
The idea is to spread the heat from the led over a bigger surface, and then take it away with a copper heatsink..
That means that I can overdrive the led much more than with a conventional heatsink..


Regards,

Ra.

I thought about doing that myself. I even looked at trying to find a chunk of SiC to use. In the end I don't think it is worth the effort as you can get a heatpipe that would only cost a few dollars and do a far better job at thermal transfer while weighing much less. The only advantage I could find to using diamond was that I could directly mount the LED die without the heatsink becoming part of the electrical path.

 

[Ray_of_Light]

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

http://www.electrochem.org/meetings/scheduler/abstracts/214/2252.pdf

Anthony

 

[McGizmo]

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

Hi guys,
This is probably a stupid question but that has never stopped me before. Is it possible with a lens group to condense the apparent image of the LED die into a smaller image somewhere forward on the Z axis and then treat this convergence point as the source to be collimated? Clearly we can take parallel sunlight and condense it to an intense point with a single lens.

I guess in other words, is is possible with a lens group to create a smaller apparent image enjoying greater surface brightness than the original LED source?

 

[saabluster]

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

Hi guys,
This is probably a stupid question but that has never stopped me before. Is it possible with a lens group to condense the apparent image of the LED die into a smaller image somewhere forward on the Z axis and then treat this convergence point as the source to be collimated? Clearly we can take parallel sunlight and condense it to an intense point with a single lens.

I guess in other words, is is possible with a lens group to create a smaller apparent image enjoying greater surface brightness than the original LED source?

No it is not possible.

 

[Ra]

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

Another means of measuring surface brightness would be to take images of the LED itself. I did not make this up. I read it somewhere but I don't remember where anymore. I'm sure it would require some specialized equipment though.

Yep, Dr.Jones did that in the first post.. What you need? A camera...


Regards,

Ra.

 

[Ra]

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

http://www.electrochem.org/meetings/scheduler/abstracts/214/2252.pdf

Anthony

Hi Anthony,

Your very short post is more informative than some much longer posts I've seen lately!! Thanks ! :bow:


Regards,

Ra.

 

[Ra]

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

I thought about doing that myself. I even looked at trying to find a chunk of SiC to use. In the end I don't think it is worth the effort as you can get a heatpipe that would only cost a few dollars and do a far better job at thermal transfer while weighing much less. The only advantage I could find to using diamond was that I could directly mount the LED die without the heatsink becoming part of the electrical path.

Uhhh, SiC ??? :duh2:

SiC has a thermal conductivity of 120 W/mK.. Copper has 401 !

http://www.accuratus.com/silicar.html


Regards,

Ra.

 

[saabluster]

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

Yep, Dr.Jones did that in the first post.. What you need? A camera...


Regards,

Ra.

No he mentioned using it to measure die size. I am talking about taking images that will tell you the actual brightness by comparing the brighttness of one image to another. I am looking for the article but the first time I found it it was quite buried in the halls of the net.

 

[saabluster]

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

Uhhh, SiC ??? :duh2:

SiC has a thermal conductivity of 120 W/mK.. Copper has 401 !

http://www.accuratus.com/silicar.html


Regards,

Ra.

Sorry I should have been more specific. Moissanite

 

[Ra]

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

Sorry I should have been more specific. Moissanite

Better... But:

Moissanite has a thermal conductivity between 360 and 460 W/mK...
Depending on the cristal-structure.. So practicly the same as copper..

http://en.wikipedia.org/wiki/Silicon_carbide


CVD Diamond has 2000..


Ra.

 

[gcbryan]

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

Just curious...so is the primary limitation regarding driving an emitter conducting the heat away or is there something inherent in the die design?

Meaning...I understand when driving within an emitter's spec that conducting heat away is crucial to maintaining that output but when pushing above spec is the main limitation still just conducting heat away or is there an internal limitation to the emitter itself? I assume the answer is yes but I'd appreciate a little clarification.