UVIR-CUT filters from Alibaba... or Optolong L-Pro?

[JA]

Originally Posted by bojan

Thank you for the offer, I will PM you when I am ready

OK sure.

Originally Posted by bojan

As for comparison....
the attached images below were taken with the same lens, left one is with unmodified 60D (last year), right one is with 450D FSM + L_pro (last night).

Some stars (with L-Pro) have halo, some others (of same or even higher magnitude) don't.
There are no halos on corresponding star images taken with unmodified 60D.
The difference is most likely in spectral type and excess of IR (most likely) or UV (I don't know yet what L-Pro filter does compared to Canon filter in 60D)
So, it seems more filtering will be needed to fix this (IR? I will know when replacement UVIR-cut from Rocolax arrives).

If you use a normal camera lens you could possibly rule out the 450D haloes as being caused by excessive L_pro UV bandwidth, by also stacking a lens mounted UV filter on to your lens, which should remove most of the UV and then check again for haloes against another non-UV filtered image, preferably taken at same time with same exposure.

In a round about way you could do something similar for the Infra Red end of the spectrum as well if you have some sort of photographic IR filter (like an R72 or the like). Except that this time you'd be looking to possibly find the halo in the Infra Red image. If you find the halo in the IR image then bingo.

Another possibility for the haloes, relates to the 450D's Full Spectrum Mod - When the UV/IR filter was removed was it replaced with clear glass of some type or left empty? My thoughts would be that if left empty or if the replacement glass is not of the right refractive index, it could cause a tiny deviation in the intended focus of infra red (if enough IR is coming through to be apparent) , in that it may not focus soon enough, which one would probably see as a sort of chromatic aberration (except not Red, Green, Blue, Magenta..., as IR is not represented by the sensor in that way, but mostly as deep deep brown/red or grey and would represent as a halo. Or maybe there's just too much Infra red bandwidth picked up by the sensor, which the lens is not able to focus closely enough to the visible light focus, either way I could certainly see it would represent as a halo of some sort (and a slightly defocused image) - see below... (coming soon .... )

EDIT: .... Now added as diagram below (that was fun !!!)

[bojan]

My modification was simple - no glass plate instead of removed filter.
Yes, I think you are right, it is IR.. I looked at couple of stars with halo and all of them have colour index >1.5, which means they are really red.
However, I am not sure the observed effect is because of removal of the glass, because uvir-cut filter is 2mm thick.. it is at the different location but still...
Most likely it is because this lens is Canon FD 400mm f/2.8L, from '80-ies last century, not corrected for IR.
And spurious far-IR bandpass of the uvir-cut filter (I will have a look with spectrograph tonight if it exists or not) allows too much IR.
BTW, all bright (saturated) stars have the ~same diameter, which is also indicative of something un-expected and counter-intuitive going on here.

[bojan]

There is also a tiny possibility of internal reflections inside L-Pro filter....

All will be clear once I receive the replacement for UVIR-Cut filter...

[bojan]

Rocolax UVIR-cut filter replacement arrived yesterday, I tested it with spectrograph and it turned out it is almost exacly the same as previous one...
So, most likely this problem I have is not about manufacturing quality but rather it is because distributors do not really know what they are selling...
After some more googling, it seems that this filter (with cutoff at 650nm) is pretty standard...
It seems there are two other cutoff wavelengths set at 700nm and then 800nm...

[bojan]

The new uvir-cut filter (680nm, from LenOpTec) arrived today.. it looks very good, H-alpha is passed almost un-affected (green plot, compared to Rocolax which is 650nm)
Halos disappeared with this filter... so the problem was caused by IR (between 670-700nm and possibly beyond 1100nm).

[bojan]

I did another filter test last night with Canon 450D (FSM) and Canon FD 400mm f2.8 (@f4).
The combination performs very good IMO, with focussing with Bahtinov mask (the star to focus on should be G~F type.. I also tried with Betelgeuse but the Bahtinov pattern was somewhat ambiguous and resulting focus was a tad off... obviously the lens issue).

The uvir-cut filter comes from LenOpTec (Beijing, the contact person was Oswyn Tu, info@lenoptec.com), the service was exceptionally good (they posted the item 2 days after receiving payment via PayPal).
Two filters plates 1.1mm thick (I asked them to cut it in round shape, 45mm dia, so I could mount the filter into Canon filter holder to replace "standard" neutral glass) cost me US$50.00 +US$25 for postage via TNT.
I imagine the other shapes and sizes are not the problem for them.

Compared to Hoya, Baader Planetarium and Astronomik prices for such filters, LenOpTec filters are absolute bargain... and the published data very closely match my own measurements.

[JA]

Good Stuff - Looks like you're in business

Best
JA

[bojan]

Yep...
This is nothing much, but illustrates the result.
Second attachment is 100% size crop from 16x120sec stack.