Sensors

[gonzostar]

Hi I am getting my self confused again. Looking for a newCanon  DSLR suitable for my set up. which is a 102 ES APO refractor on a AVX  mount

My question to do with sensors, Should i be looking at the amount of megapixels or  pixel sizes?

Cheers 

Dean

 

[wimvb]

Pixelsize translates into "resolution" or pixelscale, how large a patch of sky each pixels collects light from (arcseconds/pixel). For galaxies and planets, you're better off with smaller pixels that have the possibility to reveal more detail (weather permitting). For faint extended dso's with not so much fine detail, you want large pixels which gather more light in any given exposure time.

Number of pixels is then determined by sensor size. Larger sensors capture more sky than smaller, although most common dslrs have about the same sensor size (APS-C). Some cameras have full frame 36 x 24 mm. But most scopes won't illuminate such a large sensor.

Number of pixels in itself isn't important. As said, that is determined by pixel size and sensor size. But more Mpixels make for larger datafiles, which can become a problem if you're using a modest computer and software.

Choosing a dslr for astro imaging, I would want low (read) noise, possibility to turn off in camera noise reduction tools, availability of supporting software (Canon, Nikon), and a good responsivity for red (Ha).

If you're buying new, you should definitely also consider a dedicated astro camera, either mono or osc. Being able to cool the sensor makes a huge difference. Mono has the advantage of easy switching between rgb imaging and nb imaging. Astro cameras are generally more expensive per cm2 sensor size than dslr.

[gonzostar]

Thanks Wim clearly explained as usual 

the canon 450d which i am borrowing 12 mp and 5.2 microns, 

canon 1300d for example is 18mp and 4.3 microns

Canon 200d is 24 microns and 3.7 microns

Would you say the 1300d is more suitable for AP then 200d as a comparison between these two?

To be honest i  know that mono and filters are so much better. However i an concerned its such a minefield and so much can go pear shaped! Especially if you have to set up fresh each night. 

Cheers

Dean

[wimvb]

Yes, I know. Mono can be a leap of faith. If you're more comfortable with dslr, just stay with that.

Since your 3 options are all canon, they are most likely very similar to handle. The sensors should also be the same size (give or take a mm). Larger pixels can be better. Consider how your guiding is. It's no use going for small pixels, if guiding can't keep them in position.

Apart from pixel size, I would prefer the camera that has lowest noise at high iso. Generally, noise increases with iso. If a camera has lowest noise at the highest common iso setting, it is likely to also have the lowest noise at lower iso. It's not a certainty of course. Rather common sense. Finally to consider is weight. A lighter camera puts less strain on the focuser.

[gonzostar]

Cheers Wim this is info i needed. Maybe will think about mono if i have a permanent setup

As for guiding i can easily do 5min subs. Havent tried longer as yet. This part due to LP in the area. Especially now the council have put a led light on the SW horizon!

Probably feel that the cameras with 3.7 maybe to small for my  set up and guiding?

Cheers

Dean

[drjolo]

I would keep an eye on camera noise and quantum efficiency (sensitivity) - pixel size is important, but not critical. If your optics provides stars with diameter of 6um and guiding keeps them elongated like 10% then it does not really matter what pixel size you use to sample this image. In my opinion if 3.7um pixel camera is both less noisy and more sensitive then it would be better choice than 5.2um camera. You can always resize image down, so the optics and guiding defects will be the same as for larger pixel camera. 

[gonzostar]

Thanks Lucas Cloudy night here tonight so will also add QE and camera noise into my thoughts. Brief look the 450d QE is only 33% some models i have had a look at have a QE of around 40%. Will that difference make a huge effect on images?

I havent as yet had any complaints about elongated stars for a year now   and have being doing 300s light frames.

Cheers

Dean 

[drjolo]

Yes, modern DSLRs tends to have better QE, lower noise and (as well) smaller pixels :)  QE effect is simple: assuming all other parameters the same, you will get the same result in shorter time. Or better result in the same time. Some Canon DSLR and even more Nikon DSLR have QE close to or above 50% https://www.sensorgen.info/

[wimvb]

27 minutes ago, drjolo said:

Yes, modern DSLRs tends to have better QE, lower noise and (as well) smaller pixels QE effect is simple: assuming all other parameters the same, you will get the same result in shorter time. Or better result in the same time. Some Canon DSLR and even more Nikon DSLR have QE close to or above 50% https://www.sensorgen.info/

I don't think we can trust the numbers. As long as a QE of 476 % isn't commented on, I wouldn't trust the other numbers either. Especially since there are more cameras with a QE in excess of 100 %. And many with QE in excess of 80%, which while not impossible, is not very probable either.

[gonzostar]

My head is defiantly getting dizzy now!  

Basically newer model, not a hybrid cmos sensor like the 750d, and fingers crossed :'0m

[drjolo]

sensorgen uses data from dxomark to calculate QE. They are mostly at least comparable with each other. But for some cases it leads to bizarre results, when dxomark data is not reliable. Modern DSLRs reach QE of level 60% (like Canon 7DII) or 65% (Sony A7s). This is also the same level of QE that OSC with BSI CMOS sensor reach (like QHY183C). 

[gonzostar]

And similar prices, however i  am thinking like me keeping it simple  DSLR it is rather then OSC.. Although i cannot find the numbers for read noise and QE for the 200d yet

[drjolo]

Yup, I also looked for a while for 200D and 1300D and found nothing :( 

[gonzostar]

Thanks for looking  

[wimvb]

3 hours ago, gonzostar said:

My head is defiantly getting dizzy now!  

Just as I was about to bring up the concept of full well capacity. ?

Ah well, maybe another time.

[gonzostar]

Ok. New day, new head ready for full well capacity 

For what i basically understand the more subs you make there is a increase in signal you want rather then noise?  

[wimvb]

?full well relates to how many photons (electrons, actually) a pixel can hold before it is saturated. Pixels fill up with electrons created by target photons, light pollution photons, and electrons generated by the electronics (dark noise). The larger full well capacity, the longer you can and should expose your single frame. During that exposure time, the well will fill up with all these electrons. The nice thing is that during (post) processing, you will remove (subtract) the light pollution and dark noise part, but keep target electrons. And since a larger full well allows more of these, you're set for a better quality image. If read noise is thesame, a camera with larger full well capacity will result in images with a larger dynamic range. Which is a good thing. (A simplified explanation)

[Davey-T]

Also worth checking the autofocus system which seems to have been causing problems on some models for astro' imaging recently, not sure which models, there's a post on here somewhere about it.

Dave

[gonzostar]

Thanks Wim for that clear  explanation. I am getting some more understanding now. On the sensorgen website should i look closely at the max saturation column (e-) or min read noise more closely?

Thanks Dave, Will look for links later. Can the AF bi turned off of or should i rely on backyard EOS live view, I tend to eyeball stars on 450d live view

 

Cheers

Dean   

 

[Davey-T]

21 minutes ago, gonzostar said:

Can the AF bi turned off of or should i rely on backyard EOS live view, I tend to eyeball stars on 450d live view

It's not that you're using the AF it's the way the AF detectors are laid out on the sensor that produces lines across the image that are tricky to remove in processing.

Dave

[gonzostar]

Ah think that was me when i tried the 750d. I think it was topic lines across image

[wimvb]

4 hours ago, gonzostar said:

On the sensorgen website should i look closely at the max saturation column (e-) or min read noise more closely?

Both. You have to expose long enough for the pixel values to clear the read noise floor, but not do long that the full well is reached. From a light polluted location, clearing the read noise floor is not a problem. Filling the pixels with lp photons isn't a problem either. I would prefer a deeper well, even if the read noise is somewhat higher.

[gonzostar]

Thanks Wim How can i tell how full a well is? To be honest i think i shall be ok with a max of 300s lights frames, as i am managing this with the old DSLR The big question now is which one? Deffiniatly not the ones with the cmos-hybrid sensor!

[wimvb]

30 minutes ago, gonzostar said:

How can i tell how full a well is?

The same as in a daytime image: overexposure. Blown out parts in the image.

[gonzostar]

ah, my usual astropics then ?