I've recently obtained M67 data from a system consisting of a CDK20, QHY268M camera, and Baader BVRI filters. The B filter does not have a red leak, based on V838 Mon data. However, I'm a bit disappointed in the transformation coefficients, as calculated by TG.
Tb_bv = 0.211
Tv_bv = -0.017
Tr_vi = -0.130
Ti_vi = -0.051
Tbv = 1.269
Tvr = 1.242
Tri = 0.868
Tvi = 1.037
with typical fitting errors around 0.01 from 3 independent calculations of the coefficients.
So I'd consider V and Ic to be close to the standard system, and Rc to be slightly farther off than I'd like. The one that seems worse is B, at least compared to the original Astrodon filters. That said, using a coefficient of 0.211 will transform ok to the standard system.
I also tried using a Baader Sloan u' filter as a replacement for the Baader Johnson U filter, and the transformations were reasonable:
Tu_ub = 0.008
Tub = 0.886
I like using the Sloan u' filter because it is blueward of the Balmer limit (364nm), while the Johnson U filter straddles it.
I don't have independent spectral scans of these filters, but they have some on their website:
I would not worry about the red leak that Galli shows, as the transmission is beyond the response limit of silicon (usually given as 1.08 microns). The filters do show some bumps and wiggles compared with standard glass filters, but that probably won't affect photometry.
This particular telescope has only Baader filters, so I can't compare actual throughput of the filters compared to my reference set of Astrodon filters.
I have equivalent data for other filter vendors that I'll post in the coming days.
Which transformation coefficients would be ideal, or at least acceptable, for a set of BVRI filters?
I have a mixed set - Baader V, Astrodons B & Rc - and I am about to buy a Baader Ic.
The ideal coefficients are as close to 0.0 for magnitudes and 1.0 for colors as possible. I think values up to +/- 0.25 from these ideals are fine with linear coefficients, but above these offsets, I'd consider moving to higher-order terms. This is a hand-waving answer - I have not calculated what effect a large coefficient would have on transformation.
Usually non-red stars are no problem, especially if the comp star(s) are chosen to have similar colors. Where transformation has difficulty is for very red stars. The problem is twofold: (1) your original coefficient calculation is unlikely to have very red stars, so transforming red stars will be an extrapolation rather than an interpolation; and (2) very red stars typically have a spectrum that is heavily influenced by molecular bands, so slight differences in the spectral response of a filter can make a large difference in the magnitude estimate. You can see the difficulty most easily by transforming V with respect to (B-V) using Tv_bv, and doing the same thing using Tv_vi. As the star gets redder, the difference between these two transformed values increases.
I suppose the "Appendix D: Supporting Calculations for Color Transformations" in the AAVSO Guide to CCD/CMOS Photometry with Monochrome Cameras (2022) would be an appropriate and practical guide for the characterization of a filter set.
Am I on the right path?
I have Baader B & V filters. Here are the coefficients that resulted from averaging two standard fields (NGC 7790 & SA41).
[R Squared Values]
I don't know enough about this to know if I should be "disappointed" or not. Thoughts?
I achieved similar results calculating transformation coefficients on my Baader BVRI filter set paired with a CDK14 and IMX461 sensor in a Moravian C5 camera - a larger version of the sensor(s) found in the QHY268 and QHY600 cameras.
Tb_bv = 0.159
Tv_bv = -0.012
Tr_vi = -0.078
Ti_vi = 0.006
Tbv = 1.206
Tvr = 1.181
Tri = 0.845
Tvi = 0.999
The modest negative coefficients Arne got make it seem like the R filter is in fact not far from Sloan r rather than Cousins R, i.e. the filter is shifted blueward from the nominal standard. This might also be the case for the I filter. Could you use the ATLAS 'refcat2' gri photometry in the M67 field to make a trial of this.
Seems like where one runs into trouble with the passband mismatches is with reddened stars, or where you have a mix of unreddened and much reddened comp stars. To some extent it may be possible to use transformations involving a quadratic or with two colors (probably need that for u ---> U anyways).
is there a minimum focal ratio required for the Baader or other filters currently available?
I am looking to use a 200mm f2.8 camera lens with a CMOS monochrome camera and filter wheel for bright star photometry. The fov will be about 6 x4 degrees and I don’t want filter band pass to vary across the image. Cheers,
Here is the basic idea of an interference filter: Bandpass Filter Kits (thorlabs.com)
As the angle at the edge of the light cone gets steeper, light hits those layers at less than normal incidence and distance traveled by the ray becomes more than 1/4 wavelength.
That becomes important when using a very narrow band filter like a bandpass of 5 nm around a laser line or a favored emission line as in solar filters.
For example: AOI_Technical_Note.pdf (omegafilters.com) Has handy F ratios, cone angle, NA data. NA data is more useful for NB filters fed by very steep cones like you may see going into or coming out of fiber.
And: Infrared Optical Filters - Angular Shift - Electrical Optical Components, Inc. (eoc-inc.com) If you know your light your light cone angles, you can calculate the shift from formulas here. The shift goes to longer wavelengths because the longer path through the 1/4 wave layers becomes a quarter wave at a longer wavelength.
Generally, angles of incidence of less than about 3 degrees produce only about a nanometer of shift on a narrow band filter. Probably more noticeable on a Sloan filter than a Bessel filter, and probably not really noticeable on either except at the band edge of a Sloan and then only if the edge shift happens to fall on an important emission line.
Baader says their J/C filters are good from f/15 to f/1.8. They produce two versions of their narrowband filters, depending on the focal ratio.
Don Goldman said that his Astrodon wide-band filters were ok at f/3.6, but was concerned at much faster than this. So my personal feeling, without testing, is that f/2.8 is a questionable focal ratio. This was one reason why I chose the ASA N8 telescopes instead of the Takahashi E-180 for APASS. You can estimate the bandpass degradation for the center of the field of view from the website Ray gave, but ray-tracing to the edge of the field of view is beyond my expertise.
The way professionals did it in the past was to insert extra optics that collimated the beam so that the filter saw parallel light (and was also temperature controlled). Then optics were added to refocus the beam onto the sensor. You can't do it in this case, so the best that you can do is either find someone with the right software, or do the experiment yourself, imaging an area in the Milky Way with lots of stars and doing a transformation in rings to see the change and whether it is measurable for wide-band filters.
Thanks Ray and Arne,
very useful information. Canon have documentation detailing the optical path of the lens I plan to use.
I could stop down the lens to, say, f4 but diffraction spikes start becoming prominent at slower f ratios. Cheers,
I've followed this thread with interest and followed through on Brian's suggestion. Using TychoTracker I calculated transformation coefficients for the Baader VRI filters, but used Sloan gri photometry data from ATLAS for each corresponding filter. Here are my results. On the left are my original BVRI transformations, and on the right, the same data transformed using Sloan gri.
Now to go start digging and understanding these results better than I do.
J/C VRI Sloan gri
Tv_bv = -0.012
Tr_vi = -0.078 Tr_vi = 0.022
Ti_vi = 0.006 Ti_vi = 0.103
Tbv = 1.206
Tvr = 1.181 Tvr = 1.896
Tri = 0.845 Tri = 0.848
Tvi = 0.999 Tvi = 1.339
Thanks for these checks, Mike. I'm not sure how to interpret the figures either. For the Tr_vi case, for instance, are you giving the coefficient as a function of V-I or actually g-i? In this instance it does seem as though the Baader Rc filter is in fact closer to Sloan r. Another case is Tvr, with the value 1.896. If this is for g-r, then the large value may in fact be about what one expects (rather than being close to the ideal 1.0).
Coefficients were/are calculated in that test as g-i and g-r. Data was collected using the BVRI filters, but when I did the transformations I used the applicable Sloan bands (g', r', i') for each filter.
I hope you all don't mind me breaking in to ask a four-year-old question: what two filters should I get for the upcoming Choice Photometry course?
I apologize I don't understand the data presented well enough to interpret it for my needs.
I'll be using a 6 inch SCT and Asi178 cooled mono camera if that is pertinent.
The course description says B and V or V and I.
I would probably go with V and I, but as they say: your mileage may vary! Either are good choices. It seems that most advice I've seen suggests V then B then I then R when expanding ones filter set.
I would suggest you ask the class instructor, Ed Wiley, before your purchase filters. You can look him up in this web page and send him a PM. Consider the QE of your imager and compare its efficiency at the B and I bandpass. Page 17 of this document I think: https://astronomy-imaging-camera.com/manuals/ASI178_Manual_EN.pdf
Also consider that with a 6" SCT OTA you will benefit from the higher QE of the B bandpass. As someone new to photometry I would recommend you begin with V and B filters, but certainly do ask Ed.
I was fortunate enough to purchase Astrodon Johnson/Cousins V, B, & Rc ("I" wasn't available) just before they went out of stock. They have been everything I needed for the past three years. I have been using V, B,Rc for overcontact binary research and Rc for exoplanet transits. V & B in support of various Alert requests. But, as such I am not qualified to answer which manufacturer's filters you purchase. To be specific I am suggesting your compare the B and I (near infrared) filter's bandpass to the QE of your sensor and in consideration of your current OTA. IMO: You will better served with V & B filters to start. But do ask Ed Wiley (WEY), he has vastly more experience than I and he is teaching your class. I'm also hopeful that others in this group with more experience than I correct any misconceptions I might have.