Larger lens vs more LEDs...

I got pretty deep into researching a project long ago which fell by the wayside. I was just thinking about trying to tackle it one of these days...

Basically im trying to decide if, given a certain amount of space to work with (including heatsinking) Id be better off with more LEDs with smaller lenses or if larger more efficient lenses would make up for a missing LED or two.

I was planning on running XM-L2s and driving them fairly hard because the duty cycle would be low for this application. Probably would use a dimmable driver.(I cant remember many specifics on what I was planning anymore). I would probably need as much throw as I could get out of it but a wide, flat shaped beam would be desireable as well. I would probably be using as big of TIR type optics as I could and was looking to use 6 LEDs. Some combination of elliptical and tight spot optics would be used. If I used large reflectors Id only be able to fit 4 of them or maybe two reflectors and three TIR optics.


Would you choose a higher number of LEDs or better optics?

Hi,
have you tought about the new cree Xhp?
There is the Xhp50 with 2500lumens, and the Xhp70 with 4000lumens
Alltrough they are a bit expensive one of the things I like about them is the low termal resistance.
On a Xml and Xml2 it is around 2.5ºC/W..... and the Xhp70 has a termal resistance of 0.9ºC/W (Xhp50 comes in with 1.2ºC/W)
This numbers are fascinating, let`s say you use 6x xml on full power compared to 2x Xhp70 or 3x Xhp50
6 xml = 60W*2.5 =150ºC @ 6000lumen
2 xhp70 = 64W*0.9 = 57.6ºC @ 8000lumen
3 xhp50 = 57w*1.2 = 68.4ºC @ 7500lumen

This shows that using xhps you will need less heat synking, actualy about half. Apart from that they are much more eficiente.
Depending on the beam you want, I don`t understand it in the previous thread
The xhp50 has the same size as the Xml on the sides, this means a normal xml reflector will fit on to the Xhp50 (with optics it might be more triky because the xhp50 is slightly higher )
For the Xhp70 I have a very nice optic wich I use in my dive torch, it makes a good sized spot for diving, and also a spill for night diving. This is the optic : https://www.led-tech.de/en/High-Pow...Mobdar-extra-narrow-4,5°-LT-1211_106_122.html

Can you explain a little better what are the light angles you need? It always depends on the particular optic.... for exemple I have an optic wich makes about 15º for 7xml and the complete optic is only 39mm in diameter

I hope I have helped you any further
Regards
Mark

m.pille.led said:

This shows that using xhps you will need less heat synking, actualy about half. Apart from that they are much more eficiente.

Click to expand...

Well, it would if you hadn't made a serious error in your calculations. A gut-feeling check should have told you that 150C for the XM-L is not workable, and not a reasonable result.

Since the thermal paths are in parallel, the thermal resistance should be divided by the number of LEDs. Making this correction brings the numbers to:

XML(2): 25C @ 60W
XHP70: 28.3C @ 64W
XHP50: 22.8C @ 57W

These numbers pass the gut-feeling check (which admittedly is a skill developed over decades of engineering work). However, even these numbers have a substantial error. They assume that all of the electrical power delivered to the LEDs is converted to heat, and that's far from true.

The radiometric efficiency (not the same as the luminous efficacy) of an LED is the ratio of how much power is in the optical output to how much power is in the electrical input. This is the fraction of electrical input power that doesn't get converted to heat in the LED. Of these LEDs, the XML is worst. XML2 is probably next. XHP50 is probably highest. Obviously (from previous statements) I don't know the exact numbers, but they're probably in the 30-40% range. Maybe not game-changing, but not insignificant either.

Unfortunately, radiometric efficiency values are generally not published. If you know what you're doing you can estimate values based on luminous efficacy and color specs that are published, but it will never be better than a rough estimate unless you have lots of data and good skills.

Also, some fraction of the optical output of the LED is converted to heat in the optics section of the device. If your reflector is only 90% efficient (which is pretty good) then 10% of the optical power that hits it is converted to heat in the reflector. TIR optics are also typically 80-90% efficient. And your front lens typically only passes 80-95% of the optical power that hits it. The rest is converted to heat. And in some lights there's a fraction of the light that never even hits the optics (instead hits other internal parts of the light), and most of this is converted to heat.

Depending on your light design and application, these additional power losses may or may not contribute to your heatsink requirements. In hand-held lights they probably do contribute significantly. In dive lighting, probably not. Others would be in between.