If I remember well, Haitz article stated a trend of doubling performance/cost each 18months, and not a constant increase on performance, as we all know (Haitz included) the physical limits of energy conversion into visible light. For many years, LED manufacturers were focused on increasing the performance of LEDs, in order they can compete with other lighting technologies and goes getting into market niches. But once they have reached the 150lm/W figure, now they are focused of reducing prices at the same trend they increased performance before.
Until now, price per LED unit of a given power has dropped very little. Instead, the manufacturers offered a new device emitting more light, but at give or take the same price. It has been a slow decline in price with time, but very small compared with the increase in light emission. Now the focus is on offering the same amount of light at lower prices, and last releases of all main manufacturers clearly points to cheaper packages.
For one - that article is terribly misleading. PR probably wrote it and you know how that gets.
In a nutshell-
They are pointing 405nm lasers at a phosphor (instead of blue leds at a phosphor).
However, the numbers that they quoted (100lm/W vs 170lm/w of laser light efficiency), is almost comparing apples to oranges.
Their 100lm/w rating (which is a slightly outdated in the first place) is (probably, I can't predict from what sort of strange places PR pulls "competition" numbers from) a relatively honest figure,
but their 170lm/w rating is complete nonsense (please correct me if I'm wrong). From personal testing, I'm only getting 10-20% efficacy from a 405nm diode (~500mW in, 60mW out), and that's *before - so assuming they have some super premium LDs, I doubt it'll be higher than 20-30%.
Alright, now let's take a look at Cree's current Royal Blue emitters(the base of white emitters). Something modern, in this case the XT-E. at 1A it consumes approximately 3.25W (eyeballing graphs)- and (I'll choose a lower bin to not make BMW PR look bad) and puts out... 1.2W. that's 37% for a non super premium bin emitter (electrical efficiency)
Now, we already established that current incoherent sources are much more efficient - so how did the PR department get their super high 170lm/w figure?
From this information, it seems that their rep was quoting the lm/w figures of the led correctly, but was quoting the lm/w figure of the phosphor conversion of the laser.
I could be terribly wrong, it won't be the first, but BMW isn't a laser company. I doubt that they'll come out with anything super groundbreaking with laser tech.
Craig
I bet I know where that false figure came from. In one of the articles about laser converted white light which arises some months ago, there was a claim of a 70% improvement on light output per burned watt, but on very special conditions which may appear on car headlights: very high current density on the chips and very high power density on the board. On that conditions, LEDs suffer a strong decrease on efficiency, while lasers not, and thus when needed a powerful beam of light, on current state of the art of both technologies, laser converted white may be advantageous.
That situation just happen on conditions where LEDs gets low efficacy (not cited directly on the article) due the operating conditions.
The author of the article surely saw that figure of 70% improvement, took the typical efficacy of white LEDs as reference and calculated the supposed laser converted efficacy. Obviously, it gets a false figure, as the comparison surely was performed with LEDs emitting 50lm/W as best.