Quote:
Originally Posted by Slawomir
This interesting discussion has induced in me a question about the impact aperture has on imaging.
Given the same sampling (in arcsec per pixel), would a large scope with large pixels capture more detail in comparison to a small scope with small pixels?
I know that it would make a difference for point sources of light (stars), but not sure about extended objects such as nebulae.
I also understand that small pixels mean shallower well depth so you need to have shorter exposures, but in case of Sony sensors read noise is also significantly lower than that of large chips and their sensitivity is relatively high, so apart from lesser FOV, perhaps not as much difference after all?
Any enlightening comments will be extremely valued and appreciated.
|
if the atmosphere is the limiting factor (eg for scopes above about 6 inches in Australia), the scope size does not determine resolution. To obtain best possible image detail, the angular size of the pixels should be around 1/2-1/3x the seeing ( ie 0.67 - 1 arc sec in Australian 2 arc sec seeing). Oversampling does not get you more detail (it can't because the sky gets in the way and it doesn't take any account of what you do). Some find that oversampling provides for more freedom with deconvolution, but if you do oversample, you take a big hit in sensitivity - 2x oversampling requires 4x the time spent imaging. In terms of sampling under seeing-limited conditions, all that matters is the angular subtense of the pixels - for example, a 1m fl scope with 4.5 micron pixels will produce an identical image to a 2m scope with 9 micron pixels - both have the same pixel scale. The sensitivities of the two systems will depend on the aperture, but if both have the same aperture, the performance will be identical. ie a 1m f4 scope with 4.5 micron pixels is functionally identical to a 2m f8 scope with 9 micron pixels.
the debate about well depth is a furphy - part of the marketing by camera makers with high read noise chips. what matters is the dynamic range (the ratio of well depth to read noise) and most chips get to around 70-80 dB, regardless of well depth. If you have low read noise, you can use short subs and the average of many short subs with a low read noise chip will be the same as the average of fewer, longer subs with a noisier chip. this explains dynamic range fairly well
http://www.ccd.com/ccd111.html . Some typical dynamic ranges are 8300=69dB, 694=71dB, 3200=77dB, 16803=80dB. Low well depth can cause saturation problems if you insist on soaking the chip with photons in super long subs - something you had to do with older chips, but don't need to do if the read noise is low.
There is only a difference in system performance for extended and point source objects if the system is heavily undersampled - wide-field vistas showing a carpet of myriads of almost identical point stars come to mind.