First, Marduke, thanks for the clarification on your data source.
The data is entirely from Cree technical data sheets, and is 100% correct, and completely corresponds to independent testing by jtr1962. All I-V curves used are for the dice used to produce the white LED's, so they are completely comparable.
Really? I thought they were for the packaged emitters; if you mean you looked at the EZ900 datasheet for both, I'd have to wonder how you plot the same data twice and don't get the same curve. If, OTOH, you simply mean (as I'll assume you did) that the datasheet for white/blue
and green XR-Es is showing I-V curves and Vf values for only white/blue dies, with no explanation or warning in the datasheet, sorry, but I'll need more than your say-so.
As for agreeing with jtr1962's results, not so much. See
here and
here, 73 and 75 lm/W on a XR-E at 1A; your graph shows 68 for the XR-E, and 73 for the XP-E, and the rest of the data similarly fall at or above the XP-E curve in your graph. The results in Vf
are closer to the XR-E's datasheet values, but this was almost two years ago -- Vf has generally improved with time -- and regardless of the exact details, when you say "completely corresponds to independent testing by jtr1962.", I guess I would expect some correspondence in the quantity you're graphing... Am I missing some other results that might actually help confirm your graph instead of refuting it?
It is by no means "some random graph".
When posted by someone who probably doesn't know, and definitely doesn't state, the data source, it is. Graphs serve only to reveal or communicate facts already contained in the source dataset, and trying to prove the validity of the underlying dataset with a graph that shows the same effect only works if the graph is based on an independent dataset. Now that you've stated the source (again, thank you!), it's no longer a random graph, but a presumably accurate representation of (unfortunately) the same dataset under discussion. Thus using it as support for comparing the datasheet numbers directly is begging the question.
If you don't believe it, show me some proof that any of it is incorrect. Better yet, plot it yourself THEN criticize how my data is wrong.
As I've said, I don't believe those data are wrong
per se, but they don't belong on the same axes. So, no thanks, I won't plot them on the same axes and criticize myself.
Or easier yet, just look up the other independent measures of Vf all throughout the forum (a casual search yields dozens of individual results) and you can verify that in terms of Vf, XR-E > XP-E > XP-G.
Stuff like
this, or old data from before the XP-E even came out? Since a cursory scan shows no clear trend to either side, I'm not really interested in doing a detailed study of these. I think it would be fairly easy to justify seeing whatever conclusion you like, and rather difficult to extract reliable data.
So in summary there are THREE independent sources of the information that all concur. Do you still think EVERYONE but you is wrong?
- Comparison of datasheet values/curves:
"Everyone" conveniently assumes unlabeled values/curves in a datasheet for green and blue/white parts apply only to the blue/white parts, even though other values/curves are labeled as color-specific; this assumption also implies the datasheet has no data for the green parts. As this runs contrary to logic, and no supporting evidence has yet been offered, yes, I think "everyone" is wrong.
- Alleged results from jtr1962 which "completely correspond" to your graph, which is based on the above assumption.
I'm not seeing such results -- not to say they aren't there; perhaps I overlooked them. But until I do see them, yeah, I'll stick with "everyone" is wrong because of A.
- Random posts where people have measured actual Vf values
Great idea, but the actual data are widely scattered and influenced by a large timescale and a known downward trend in Vf; results are inconclusive. So I'll still stick with "everyone" is wrong because of A.
