Heat issue using XP-G compared to MC-E

[videoman]

I was wondering if I would be better off using say 8 XP-G's ( highest bin) driving them at 1A or 4 MC-E's ( highest bin) at each die 700ma) for heat issues. I see that both scenarios would produce about the same lumen output (2500 lumens) but which really better (less) for heat produced in the same heat sinked enclosure? I really have no problem with heat with the 4 MC-E's but I would sure like to use the XP-G's later on.Also, which would be better for battery efficiency ? I would be using 2 shark bucks 3A versions in either case from a 14.4 V DC source.

 

[John_Galt]

I'm pretty sure that running 8 XP-G's would both reduce power consumption (8000 ma as compared to ~11000) and heat, as less energy would be consumed to produce the same amount of light.

That would also reduce power consumption, greatly increasing run-times.

 

[Gryloc]

Hello Videoman! You have brought up an interesting idea of replacing the once powerful and compact (and expensive) MC-E with more smaller and more efficient emitters (cheaper as well).

So, I did some quick math based on some actual measurements based from jtr1962 :twothumbs. First of all, jtr1962 recently tested a MC-E K-bin flux 4A-bin tint (Neutral White), which matches atleast the flux bin of the neutral MC-E emitters you have (which we have used before). Secondly, I made some estimates of the performance of the XP-G based on the XP-E. I was able to determine the approximate "lumen vs current" curve by pretending that two XP-E dies were wired in parallel (acting like the new 2mm^2 die). With a little shift to fit the numbers into the R5 bin (offered at Cutter), the curve seems to match the curve offered by Cree in some of their documents and datasheets pretty closely. Jtr1962 commented to say that the estimates were pretty close, but I failed to consider the different "junction to pad impedances" in degrees C per watt (C/W) between the XP-E and the new XP-G. However, he mentioned that the error is not too noticeable when he did some calculations. Knowing that actual measurements for the XP-G will be done soon enough, I do not plan on adjusting and re-posting the images of the graphs. Below, I posted my findings. Consider that the Vf measurements and flux measurements may not apply to your own sample of emitters due to variations in Vf and flux (contained withing the bin range). Even though your results will vary, the numbers still provide a decent estimate that should be useful during light design...

-So, from jtr1962's measurements of the MC-E K-bin neutral white (here in post #336), four MC-E emitters powered at 750mA per die will produce approximately 2871 lumens (calculated using linear interpolation) while consuming 39W of power. In this case, these emitters are operating at an efficiency of 73.7lm/W.

-From my estimates of the XP-G R5-bin (here in post #491), eight XP-G emitters powered at 1000mA each emitter will produce approximately 2770 lumens while consuming 26W of power. In this case, these emitters are operating at an efficiency of 107.5lm/W

This difference in power usage is mostly due to the efficiency with the number of dies used in each case. If you can get XP-G emitters in neutral white, or if you can properly filter the light to get the color you desire, then you will find that the XP-Gs could work better due to using less power. You would have to experiment because both would suffer from decreases in output from heat and optical losses. However, the XP-G would need more filters at this moment until the neutral white emitters are released, and the MC-E may suffer a bit more from losses due to heat. These would be things to stress over if the performance and costs of the two emitters were close, but the XP-G is clearly more efficient, and much cheaper.

The decision is up to you because I am not sure how the XP-G's efficiency would be affected from color-shifting filters. For those looking for higher raw output, then the XP-G is a good choice.

Here are re-posts of my XP-G estimates (look at the far right set of data in the charts). If you have questions, please refer to the previously mentioned post (follow the above link) as I explained things more clearly.


(click for full-sized view)


(click for full-sized view)

Cheers,
-Tony

 

[Curt R]

There is also a huge difference in light output quality as the MC-E with most optics produce a Lumen to Candela output of around 6 to one and up to about 20 to one with the best of them. The XP-G because of the smaller point source of light offers a conversion of 6 to one with a tiny 10 mm diameter optic and jumps to 48 to one with a 20 mm optic. More Candela equals more throw.

Curt

 

[videoman]

Tony, thanks, that's what I wanted to hear. I just can't wait to make some prototypes as soon as Cree has the neutral whites available to Cutter. I like the efficiency advantage as we can all use as much battery juice as possible. I also saw Cutter has them in square boards which I am planning on laying them right next to each other for compactness. Question is do I have to space them out say a few mm apart or it is OK to just have them touch together and use a blob of solder to series them. That would save a lot of time wiring them. They also have 3 on a star or round board already series configured (useful for my 12 led model), many possibilities. Any suggestions as to the ideal layout keeping heat down as much as possible? I am thinking of 2 light models, one with 8 XP-G's and the other with 12.
Thanks for taking time to answer.