Affiliation
British Astronomical Association, Variable Star Section (BAA-VSS)
Thu, 07/27/2023 - 20:46
I'm looking to return to photometry after a 10 years with a very user friendly Optec SSP3, an enforced break and hopefully moving into
camera photometry with the greater potential for remote operation.
A fundamental question seems to be CCD or CMOS.
This is a much more technical area for me so simple (as possible) advice would be very welcome.
Kevin
Hi Kevin,
Congratulations for getting back into photometry!
Here are some pros and cons.
CCD pro
usually larger pixels
decent read noise
decent QE
always 16-bit pixels
can be inexpensive if purchased used
CCD con
few new cameras available, so you may have trouble getting service in the future, depending on the vendor
for the same pixel size, higher read noise and dark current than a CMOS camera (though these are really not important for wide band photometry since the sky or star noise will dominate)
new cameras tend to be more expensive for the same sensor area
CMOS pro
lower read noise, lower dark current, higher QE than CCD
can take very short exposures, not limited by shutter speed
lower cost than equivalent CCD sensor size
CMOS con
no mechanical shutter, so darks require covering things and/or blank filter taking up a filter wheel slot
random telegraph noise, which requires multiple images and stacking to fully remove (though low level, and sky/star noise will dominate)
some early generation sensors (such as the IMX183) have amplifier glow
some early generation sensors (such as the IMX183) are 12-bit only
So both sensor types have good points and bad, though most of the "bad" is mitigated by taking exposures through wide-band photometric filters like Johnson/Cousins or Sloan. If I were starting out today, I'd probably select a camera that uses the IMX-571 sensor, such as the QHY268 or the ASI-2600, as this sensor is the best CMOS value right now, and you can use 1.25" filters with some corner vignetting (or 36mm filters with no vignetting).
Arne
Hello Arne,
Many thanks for your clear reply which I can actually follow without a glossary.
It's a steep learning curve requiring mountaineering equipment.
Having just moved to a site with a pretty good (albeit sea-level), my first instinct was to dust off the OPTEC SSP3, get a C8 with a decent mount and get started again possibly within a few weeks. Then gradually try to get into CMOS photometry perhaps with a smallish APO (80mm ish), whatever I can afford. People seem to be doing useful work with similar set ups (but CCD's).
Speaking to people here who were doing imaging from the comfort of their homes made me wonder if I could do that.
Getting on a bit now, I realise I may need a bit more comfort and more importantly be more
inclined to do useful obs with a remote set-up.
I'm just starting to get into learning about camera photometry, and most replies are way above my level.
The AAVSO webinars are also well suited to where I am. Working my way through them. Lots of ah ha moments.
Would you be happy for me to post your reply to the BAA forum post on this debate?
I am hoping to get some local help over here in UK with our "special" weather which might also be a relevant factor.
Lots of people active in imaging which seems to be a good starting point as they have to solve many of the same problems.
There may well be some people using your recommended models, hence the reason for posting your comments.
Thanks again. I will almost certainly have further questions.
Regards
Kevin
Hi Arne,
As more and more folks are using CMOS cameras for exoplanet observing, we are assessing the importance of including some measure of random telegraph noise into the photometric error computation along with the normal values for readout noise, dark current, etc. In doing an AAVSO search, I see in your response to Kevin that you mention it. So, my question is: in doing time-series diff. photometry, what is your opinion about how much we should try to quantify it as part of the overall noise calculation and, if so, how could that even be done.
Thanks,
Dennis
Hi Dennis!
RTN only affects a fraction of the pixels, kinda like hot pixels in a CCD. It is all based on impurities and the generation of excess charge. For those pixels that exhibit the feature, you get a mean value and then two bands - one above and one below the mean value - when the pixel is viewed over a period of time. The problem lies in its name: random. It is very difficult to theoretically determine the size of the effect, as it differs for every pixel. The best way to handle it is through the typical empirical methods when doing a time series - it is built into the measurement error and does not systematically affect an average. Note that its size will be different for the comp star than for the target star too! Luckily, it is usually a VERY small effect for bright objects. I'm starting to look carefully into this interesting effect and may have updated information in a few months.
Arne