Interestingly, Luminus’s SST10 blue datasheet indicates shift in peak wavelength as function of both current and temperature:
Here are measurements for each of the seven stepped levels and turbo. Interestingly, at higher levels the spike is even lower than 450nm.
Thank you for compiling these color tests–I really appreciate it! Here’s an overlay of the first with the last:
So it looks like current density does play a significant role in the peak wavelength shift. Though the difference is quite minimal between adjacent levels, so I don’t think current density alone explains the SFT40 and SFT70 difference.
Also, there is still some difference between your level 1/7 and the sample tested by koef3. koef3’s sample has a peak wavelength 5nm even longer, with a bit less cyan dip:
I did notice that wavelength binning for blue LEDs spans a 15nm interval from 445 to 460nm; perhaps different SFT70 samples came from different bins of blue.
thank you! which light is the test host?
besides individual LED variation…
do both tests use the same host protocol?
ie, bare led, or reflector w glass, or Tir w or without glass?
Oh that’s a great point, koef3 uses a bare LED in an integrating sphere while @crbnfbr presumably measured the beam from a secondary optic, so conceivably some of the spectral differences could be due to optical setup. I don’t have a spectrometer but have visibly observed how much the choice of secondary optic can influence the spectrum.
Just as a reminder,I also tested the SFT-40 3000K under same conditions (1 W power, same spectrometer, bare in sphere) as well
Light was DA1 with 10394 optic.
I know my using a light vs koef3 using an integrating sphere isn’t an apples to apples comparison. I was mainly trying to show the shift in the blue spike with different current.
