This dataset has some failings, however -- I was unable to record the spectra at the same horizontal location on the spectrometer each time due to the setup being a little... haphazard. So the curvature of the spectra will introduce some error. Still, all of the peaks were wider than the calibration peaks, and so at the very least this can give us a bottom limit for the bandwidth for each LED:
https://spectralworkbench.org/analyze/spectrum/12106
I.e. if the calibration peaks are narrower than the LEDs, then it's not just because the spectrometer was out of focus -- the LEDs really don't have that narrow of a bandwidth:
2 Comments
wow, good results jeff, as it happens, I'm about to do a similar test on some filters which we use for another experiment. It is rather important to get a narrow a band as possible from the filters and I've always suspected they are wider than desirable. LEDS are also an option for this experiment, so great to see the bandwidth comparison.
Reply to this comment...
Log in to comment
Thanks for posting this Jeff. The results are interesting. Most of the LEDs that I've seen have band widths in the 5nm - 10nm range. It's important to remember that band width is usually calculated as the full width at half maximum (FWHM) [http://en.wikipedia.org/wiki/Full_width_at_half_maximum] so what we see on the graph will appear (at least it does for me) broader than the actual specification. That said, some of these still seem to be broader than (what I expect are) the specs.
Adding a FWHM calculation in the spectral workbench might be a helpful tool.
Reply to this comment...
Log in to comment
Login to comment.