There is no simple answer to your question I recommend the reading of this comparison:

http://www.teledynedalsa.com/imaging/knowledge-center/appnotes/ccd-vs-cmos/

But I am enclosing the conclusion of the article just in case you do not have the time for rerading the complete reference. Here it is:

Choosing the correct imager for an application has never been a simple task. Varied applications have varied requirements. These requirements impose constraints that affect performance and price. With these complexities at play, it is not surprising that it is impossible to make a general statement about CMOS versus CCD imagers that applies to all applications.

CMOS area and line scan imagers outperform CCDs in most visible imaging applications. TDI CCDs, used for high speed, low light level applications, outperform CMOS TDIs. The need to image in the NIR can make CCDs a better choice for some area and line scan applications. To image in the UV, the surface treatment after backside thinning is key, as is the global shutter requirement. The need for very low noise introduces new constraints, with CMOS generally still outperforming CCDs at high readout speeds. The price-performance trade-off can favor either CCD or CMOS imagers, depending on leverage, volume, and supply security.

Hi Orlando, thank you for the answer, yes I have seen that article. I guess I will have to buy a camera with the chip that fits the resolution required and test it. The small pixel size (1.3 micrometre) is a necessity for the lightbending experiment, but then also sensitivity is important. One can get a fair amount of detail from the manufacturer's specifications, but in the end testing the chip will show what is what. If I can take images at say 20 fps, I can do some frame stacking/lucky imaging.

Hi Ludwig,

regardless of the choice between a CMOS and CCD, a pixel size of 1.5 or 1.3 micron is very small and, as far as I know, is not easy to find devices with pixels so small. Another limitation of so small pixels is the limited Full Well Capacity, which involves a small Dynamic Range. I do not know your experiment, but generally what is relevant is the Image Scale (arc sec/pixel) that also depends on the focal length of the instrument: you can get the Image Scale you need even with larger pixels, but by increasing the focal length. 20 fps is a quite fast readout rate for a full frame CCD, but it also depends from the number of pixels that you have to read.

Thanks for the answer, the plate scale is 132 arcsec/mm and image scale 0.2 arcsec/pixel. The closest what I could find is the Sony IMX121 CMOS, 1.55 μm,size 4096x2160 at 21 FPS. I would probably have to test it at times of full moon, as the starfield at the limb is approximately as bright as just outside the corona during a total solar eclipse. Then I can experiment with frame rate and exposure times. Reading this requires USB3 and fast solid state disk.

Dear Ludwig,

I am working on eclipse data, for Baily beads analysis to measure the solar diameter, see e.g. http://link.springer.com/article/10.1007%2Fs11207-009-9378-x

and http://link.springer.com/article/10.1007%2Fs11207-009-9400-3

My recommendation is to test your equipment with lunar occulations as in the attached file (communication presented to LARIM 2013 IAU meeting).

About the CMOS vs CCD debate I have two experiences:

one about delta scorpii magnitude evaluation with 0.5 s fixed pose CMOS expositions (Sanyo GV9 camera)

http://arxiv.org/abs/1112.2356

and

http://arxiv.org/abs/1109.5865

basically the stars were saturated and the measurements were not all good; naked eye evaluations were more accurate within 0.1 mag.

in the references of these two papers you find Lanciano & Fiocco saying that Nikon CCD G channel is equivalent to V Johnson... this may be interesting to be tested in CMOS using Landolt reference photometric stars.

the other experience is with the same Sanyo Camera for Venus occultation and for Arcturus occultation:

you can check the youtube test video here

http://arxiv.org/abs/1310.6557

CMOS are faster than CCD. For improving the research in Eclipse timing (and solar diameter measurements from ground) you can get 60 fps for unlimited time, or even 300 fps for 10 s (Sanyo CG10) and 1200 fps (casio Exilim) for commercial camera.

No CCD at this time have such timing velocity.

No astronomical camera have these video capabilities, the limit in the eclipses has been reached in the Australian one of Nov 13, 2012 by the group of Serge Koutchmy. Photometric measurements have been made by the french CNRS either in 2010 and 2012 eclipse, with the results still in analysis.

Costantino

Hi Constantino, thanks for the excellent information. I have looked at your papers and found it quite interesting. I will do some tests with lunar occultations as you suggest. Of course I am hoping to capture stars during the total solar eclipse down to a magnitude of 10-11. There are brighter stars (only a few) in the star fields at the time of the eclipse, but I want to capture hundreds of stars instead of just the bright ones.

Repeating Sobral eclipse relativity tests made on Hyades?

Recently I read some books in Portuguese on this eclipse finding some aspects that I did not know like

http://pt.wikipedia.org/wiki/Museu_do_Eclipse

by the way let me invite you to observe the Nova Centauri 2013, so bright as 3.6 magnitude in these days. Some information here:

http://www.aavso.org/possible-nova-centaurus#comment-7527

there is a great lack of observers in Southern Hemisphere.

greetings

Costantino

[email protected]

Hi Constantino, thanks for the information, it is a good thing that the museum was built in remembrance of the Sobral total solar eclipse experiment. I will have a look at Nova Centauri 2013. I suppose it is difficult for you to find dark skies from Rome!

I am planning to repeat the Sobral eclipse experiment, but of course using modern technology, no photographic plates, but a CMOS chip. The observation strategy will also be quite different (small star fields etc.)

Oi Ludwig, I observed from Porto Alegre-Sao Leopoldo and from Rio de Janeiro the Nova Centauri:

here the first results

http://adsabs.harvard.edu/abs/2013IAUC.9265....1G

http://arxiv.org/abs/1312.4848

Constantino, seems it requires a username and password to read.

this paper is open access http://arxiv.org/abs/1312.4848

while the first is the IAU Circular saying that Nova Cen 2013=V1369 Cen

I cannot access too

here another nice paper on the Nova Cen 2013:

http://www.aavso.org/nova-centauri-2013-another-bright-naked-eye-nova

it is still a very easy object even with binoculars, since Beta Centauri is in the same field of view.

You have other infos here:

http://www.aavso.org/v1369-cen

regards

Costantino

Thanks, will have a look at these.

http://www.youtube.com/watch?v=DB740lEkEkg

Most of CCDs have high sensitivity. It is important for obtaining of faint signal. CMOS sensors can be appropriate for bright objects.

http://www.steves-digicams.com/knowledge-center/how-tos/digital-camera-operation/ccd-vs-cmos-whats-the-difference.html