For those who own LHires III and interested in the very near UV world, let me publish here my measurement of a banknote checker (BLB-T5/6W).
Here I managed to calibrate and identify the Hg lines:
The 3 prominent lines I used for calibration (order 3):
3650.153
4046.563
4077.834
Some other visible, but somehow makes calibration incorrect (would need smoothing the background precisely):
3650.153
3654.836
3663.279
3906.37 (weak here)
4046.563
4077.834
Of course, there are some more weaker lines visible at higher resolution. Let's see how the interval look around the H,K lines:
- indeed, I do have ripples in UV, too (ATIK 414 EXm), the error is +/- 1.5% of the actual signal - whilst the internal lamp didn't show any ripple...
- internal lamp's flat shows only dusts + reflections from visible; now I see reflections must be removed if someone intends to use it
- internal lamp's flat becomes fully useless with a photometric slit (too big reflections)
I'm going to use now this pair of "BLB-T5/6W" tubes for creating flats, to correct my ripples (canceling the lines at 3901.87, 3906.37 manually on the flat in 1D).
Other experiments I did, to test materials as a diffuser:
- a white T-shirt removes UV signal; but some weak transparency or most likely fluorescence (!) it provides towards 4000A
- A4 paper fully removes signal below 400nm
- white plexi fully removes all signal
- clear plexi transmits mostly everything as-is; best diffuser I've found so far
Furthermore, this BLB-T5 tube is written typically to peak at 367nm. On my graph, peak is at 376nm. Therefore I suspect this must be the honey spot where my optical system become inefficient (including the 250/1000 newton + APM 2.66x barlow I used for now).
A positive surpise, that the 3 Hg lines at 3650, 3654, 3663 seems to be still focusable with the doublet, where the internal Ar lines become fully vanished!
It would be nice to see where the real peak should be, to do an assessment of efficiency how deep it's worth to go with this instruments.
About further flat lamps possibilities, I did find UV Leds - but their half width is only about 15nm, and many of them needed (with much more planning care and soldering skills). Found also no prepared "LED array" stripes for the UV - as it exists for visual range very cheap... I think it's for future. Currently I was able to disassemble the banknote-lamp quickly without any soldering or high DIY skills, already assembled it on a plexi. Hope I can use it for observations around H,K line at high resolution.
Of course, it would be nice to see other spectrographs testing with a similar lamp.
- Peter