Quote:
Originally Posted by White Rabbit
1. If I bin the RGB subs am I correct in quartering exposure time with relation to the duration of my Lum subs?
2. How do I calculate my sky glow?
3. What is the optimal FWHM I should looking for with my rig? Or is FWHM not dependant on equipment? I may have my wires crossed on this, I wouldn't be surprised.
4. How do I tell what colour my chip is least sensitive in?
5. Then, how do I know how much to increase that channels exposure time by?
6. I've read that the main aim of master calibration frames (darks, bias and flats) is statical average of noise, and that I should be combining at least 50 calibration frames or more as this will give a better statistical average.
|
1. Not necessarily, see answer #2
2. The lengths of your subs should be limited by sky shot noise (unless you are worried about saturation or blooming). That means the exposure is long enough so that the sky background (from light pollution or natural sky glow) is bright enough that the noise it introduces is greater than the camera read noise. There are online calculators for this such as:
http://starizona.com/acb/ccd/calc_ideal.aspx
and
http://www.ccdware.com/resources/subexposure.cfm
both these pages have instructions on how to measure your sky flux. Note it will change on a per filter basis, with L having the most flux, and RGB being about a 3-4 times less. On camera binning (for CCDs) combines the flux from 4 pixels into one readout, hence potentially 4x more flux for the same read noise. However not all chips have the same readout noise in 1x1 and 2x2 modes. For example the KAF16803 has been found to offer no advantage when binning 2x2 due to increased read noise in that mode.
3. Optimal FWHM or zenith seeing limit, in arcsec, is determined by your observing site. Under 3" is pretty good for coastal areas. Good sites have seeing better than 2", pro sites try to get down to 1" or lower. Bad or misaligned optics will make things worse of course. Look at unsaturated stars in your subs with MaxIm. If you entered your focal length it will tell you your FWHM in " directly, otherwise multiply by pixel scale in ["/pixel] = 206 * pixel size[um] / EFL [mm] = 1.56 in your case. You must measure seeing above 70deg altitude, it degrades noticeably as the altitude drops. This is due to the increase in airmass that your scope looks thru. The airmass for an object at some time is calculated by TheSky, look in the Object Information dialog. The FWHM degrades with approximately the square root of the airmass (exponent=0.6 in theory). Monitor this often and get familiar with what your skies+equipment give you. BTW with a refractor you may notice different values for R vs G vs B (chromatism), furthermore I find B frames degrade with altitude faster than R, so I try to shoot B higher up than G or R to try and even things out.
4. SBIG has a graph for your chip somewhere on their website, but the answer is blue. Compare the fat red line (your chip) to the thin blue/green/red lines (OSC filter bands) on the attached graph (I made it a while ago by pasting various QE curves onto each other)
5. Don't bother. Pick a length long enough for the blue channel to be close to sky limited. The L frames will probably need a different length (shorter) if near light pollution. Then take more blue frames than Red, and more Red than Green, perhaps a 7:6:8 ratio for RGB (guessing from the graph). This needs to be adjusted again depending on the object and what features you want to stretch. Getting the color balance right for DSO is *really* tricky. I have tried G2V and excalibrator but the results were inconsistent, especially in light polluted data or very deep data from dark sites. I think it is due to the great uncertainty in the bgd level offset. For example: how much of the glow is natural vs man-made, and natural glow has distinct spectral features, which ones were dominant that night? What were the actual RGB extinctions due to airmass (the curves actually change depending on aerosol content!) It doesn't take much error in the black point offset to greatly shift the color of the faint regions.
6. Take lots of bias frames, they are quick to do. Number of darks depends on chip temperature and subexposure length, see the ccdware online calculator. I usually take 24 to 48 bias and 12 to 18 darks.
See gcx.sourceforge.net/html/node7.html equation 6.10 for a formula to give you an idea of the balance between bias and darks.
Best,
EB