Wed, 02/23/2022 - 01:41
Are there design flaws in using a dichroic or trichoic prism assembly to split incoming light from an OTA and imaging with multiple sensors similar to how microscope assemblies are able to image in multiple bands at the same time? Is it just the cost and complexity of such an assembly being out of reach for amateurs especially compared to OSC or DSLR setups? What made me think about it was reading about the trichroic prism setups for the old television camcorders that have 3 ccd's and the prism to split out the light into RGB.
There have been many professional designs using beam splitters or dichroic/cold mirrors to image simultaneously in multiple passbands. Examples are ANDICAM on the SMARTS consortium 1.3m telescope, MuSCAT3 at Las Cumbres, ULTRACAM on the WHT, the original MACHO microlensing experiment, etc. Usually they use dichroic beamsplitters so that there is a wavelength split, with all of the blue light, for example, being reflected and all of the red light passing through. This gives higher throughput at any specific wavelength than to use a half-silvered mirror or prism to reflect, say, 50% of all light one way and 50% the other way. The usual problems are optical aberrations such as astigmatism, and polarization. With amateur setups, one of the big issues would be back focal distance. This is certainly a technique to explore. APASS took a different approach, by using two telescopes on a single mount to image in two passbands simultaneously.
Arne
Arne, thanks for that information. I hadn't read about any of those specific projects. If someone was to design a setup to image in the B V and R bands simultaneously, would the overlap in the bands cause issues with the data where one band would get all the light from a particular wavelength due to the beamsplitter and the other would not? I'm not an optician so it's harder for me to understand if the beamsplitters truly reflect all light of a particular wavelength or if that's a percentage that can be modified. It seems like if someone had to have custom optical elements created that would put the cost out of reach of most amateurs versus something that could be designed with off the shelf optical elements.
Typical mirror/prism beamsplitters split all of the wavelengths simultaneously. You often use them in a 90/10 configuration, where 90% of the light goes to a science instrument and 10% goes to an acquisition/guide camera.
For wavelength splits, you use a dichroic beamsplitter. This acts like a long-pass filter, where all of the blue light gets reflected and all of the red light gets transmitted. For MACHO, that created two wide bandpasses which they labeled blue and red. For finer separation, like to split for standard system bandpasses, the overlap you mention is the tricky part. Johnson/Cousins is really hard to do this way, so most of the professional systems use the Sloan system, which was designed to not have any overlap between filters.
So doing a blue/red split is easy, and you can follow each beam with a true Johnson B and Cousins Rc or Ic filter cleanly. Edmunds sells what they call "cold mirrors", which split at specific wavelengths and so you can select an appropriate filter. An example is
https://www.edmundoptics.com/p/45deg-aoi-250mm-diameter-cold-mirror/19159/
Usually you have to create an optical system, where you first make a parallel beam (much like what is done in a spectrograph), do the split, and then add refocusing optics for each channel. You can do it in a converging beam from a telescope, but I worry a bit about optical quality and the correct bandpass as a function of field position. Designing such optics is beyond my capability. I bought some cold mirrors and some Meade flip-mirror accessories once, with the thought of making a 2-channel imaging system for an AAVSOnet site, but never followed through! This was based on a design by Smith, Genet and Heather, at a 2006 SAS meeting:
https://ui.adsabs.harvard.edu/abs/2006SASS...25...87S/abstract
Good luck! If you have any success, you should post on the Instrumentation Forum.
Arne
Great stuff! Thank you again. The paper basically had all the questions that popped up in my head when thinking about it. So maybe the "ideal" setup is to either keep it at 2 bands, or make a binocular setup with 2 scopes and 4 cameras to get all 4 bands at once. I know that most of this is beyond my current capabilities too. Mostly I'm just curious why it's not a more popular thing. Now I know why, and it's up to me to learn more about it if I want to give it a go.