Hello! I've been getting the itch to try another telescope building proect. I've been reading about folks who have made F3 mirrors and used a ParraCorr for use.
I do not underestimate the difficulty in polishing a fast mirror to an acceptable figure
I've read a little about narrow band filters changing band pass with steep light cones. Are photometric filters affected in any way by fast optics and steep light cones other than possible vignetting? Thank you and best regards.
Arne (and others) can certainly give a more quantitive answer but fast optics will affect interference filters. Old-style colored glass filters worked by absorption, with the layers of different colored glass absorbing lower or higher frequency light, leaving a window in the desired wavelength range. The f-ratio didn't cause any issues. However, modern interference filters work by having carefully spaced layers of materials with different properties. These layers block certain wavelengths while allowing others through with almost no loss. The spacing of the layers is critical and generally tuned for nearly parallel light (ie: a very large f-ratio). The light cone from a fast telescope will enter the filter at a steeper angle, increasing the effective spacing of the layers and changing the wavelengths they transmit. Having said all that, the real-world effect is probably pretty small: basically it would shift the effective bandwidth a bit but you should be able to correct for that if you transform your data. Even if you aren't doing transformations, no one's filter, camera, telescope, and site are a perfect match to the standard magnitudes.
Narrow-band filters like those used for imaging faint nebulae might be more problematic since you'd likely lose some of the emission line data due to the filter's wavelength shift.
Shawn does a nice job of describing the issue. Don Goldman told me once that he thought his filters were ok down to f/ratios around 3, but I'd be cautious anywhere close to that speed. I've stayed in the f/4 range for most of my fast-optic photometry.
You have to do ray-tracing, but one of the problems is that the cone of light that makes a star image in the field of view will be different at the center than at the edge, and so the effective central wavelength and bandpass width is not constant across the image. Transformation only takes care of part of this. For narrow-band filters, this can be important. For sensors with micro-lenses, fast f/ratios can modify the quantum efficiency. How fast is too fast? I think each situation will be unique. Really precise, calibrated photometry would require placing the filter in parallel light and with temperature control.
When I was working on my dissertation (many years ago!), I had to worry about the central wavelength of narrow band filters, as I was imaging targets both in our galaxy as well as in external (red-shifted) galaxies. You can tilt filters to change their central wavelength, and also change their temperature. Both methods modify the effective thickness of the interference layers. Tilting is the most similar effect to fast f/ratios.
Both Baader and Custom Scientific, and others, still produce Bessel colored glass filters. Perhaps you should consider using Bessel filters rather than be left hanging about whether interference filters are degrading your results. Of course there are some downsides to Bessel filters, but these may just be part of the cost of doing business (photometry) with fast optics.
hmmm, I am used to gaussian beam optics, Has anyone tried putting a concave lens in front of the filter to parallelize the congerging beam?
Is that what the Paracore is doing?
A typical effect of off-axis rays on interference layers can be seen on any manufacturer's plots of transmission verses angle for AR coated optics. Usually shown for a single wavelength like 532 or 1064 nm. ( a fairly narrow bandwidth of << 1 MHz). These are also good websites to learn how coatings are done and how they work to make narrow band filters.