ASI183 or???

[LewisM]

Doing the maths for my 450mm focal length setup and typically good seeing, it would seem between 2 and 3 uM pixel size would be ideal. This somewhat limits choices, yet the ASI183 at 2.4uM comes to mind.

I currently use an 8300 chip camera with 5.45uM pixels and not been unhappy but will a 2.4uM sensory deliver sharper results? (FSQ85 - 450mm - f/5.3. Seeing here is typically 2 FWHM or less)

EDIT: just to re-iterate, I have ZERO issues with guiding, the mount or the scope. ZERO. Perhaps my focusing could be better (hand focused with a Bahtinov) but this is not what I am asking about.

[gregbradley]

The ASI533 might be the better choice but its only in one shot colour. The ASI1600 is the most popular CMOS camera. Not expensive (well compared to SBIG and FLI) micro 4/3rds sensor size, slightly lower QE at 60% but a reasonable pixel size that suits a lot of scopes.

I am using the 183mm. You can always bin 2x2 if the seeing is not as good. It doesn't seem to affect the resolution much.

The 183 has amp glow so you need to settle on settings and keep it simple otherwise callibration becomes complicated.

Also put the settings you used in the file names of the images you capture so you know which darks to use.

The amp glow does callibrate out very well so long as the darks are exactly matched.

There is a mono 2600 equivalent i QHY coming out in October. That has 3.76 micron pixels, no amp glow, 16bit and 84% QE in an APSc sized sensor. It will be very popular and will sell out for a while. Its basically a scaled down version of the ASI6200. I am tossing up whether to get that or the 6200 mono. 122mb files though don't excite me with the ASI6200 but maybe the extra field of view is worth it.

Greg.

[LewisM]

I thought about the 533 but it’s not a lot different to the 8300 I suspect. Some but worth it?

I only use OSC so I am ok with that

[Atmos]

It'll definitely give sharper results but I'm wondering if the 2.4 micron pixels might be a little small due to diffraction setting in. Not that it really matters as much if you're only wanting to go down the OSC path as that will fatten the stars a bit anyway.

[multiweb]

What don't you dither and use drizzle in PI?

[LewisM]

Oui oui Marc, mais...

[multiweb]

Mais quoi? C'est ce que je fais et ca marche. Economise ton argent pour le vin au lieu d'acheter une autre camera qui ne sera pas d'aussi bonne qualite que ton 8300. :P

[gregbradley]

I'd say its a lot different. Its BSI (backside illuminated) sensor. So it has large wells for its size and 84% QE which is higher than any regular CCD. The KAF8300 QE is around 59%. No amp glow which is one of the more annoying aspects of some CMOS sensors. 3.76 microns is a good size for shorter focal lengths.

As to diffraction I thought that was a lens quality when you stop down too much. Perhaps it happens with small pixels. the 2.4 micron pixels on the 183 are fine with my RHA at 1159mm focal length. I find its quite sharp. I also find it tends to give rounder stars because each star has more pixels showing it.
It does however make it harder to focus. It takes me longer to focus with it.

3.76 microns is probably a better fit pixel size for that scope. The 183 would suit a widefield refractor better but I would want a large sensor for my refractor because I got it for wide images.

Also these CMOS sensors are usually more sensitive in narrowband than regular CCDs.

The ASI2600mc seems to be the one shot colour a lot want. APSc sized, same 3.76 micron BSI sensor, same 80+% QE, very low dark current and read noise. I've seen some excellent images from them.

Greg.

[LewisM]

L'argent n'est qu'un ennui

[gregbradley]

I had fun translating that Marc. My wife and I are trying to learn French.

Greg.

[multiweb]

:lol::thumbsup: Bonne chance. :)

[LewisM]

I learned French for 6 years. Fat lot of good it did for me. I can’t even insult Marc properly

[Atmos]

Votre mère était un hamster et votre père sentait le sureau.

[gregbradley]

Hahah LOL

[JA]

Out of curiosity you could try a Nikon1 J5 mirrorless camera (now discontinued) it has a 1"-type 20.8MP BSI CMOS sensor (5568 x 3712 pixels) with 2.4 µm pixels : and if purchased used, would be possiblty be 1/5th of the price (if lucky) somewhere from $250-500, depending on extras. (They were ~$1500 originally with the top spec. lens kit). It has an unusually clean image for a sensor of that size (same size as ASI183 : 13.2 x 8.8mm)although not as good as it Nikon Fullframe brothers but certainly better than some older fullframe and APSc cameras. It has a fold out screen so you could go possibly go totally portable (sans computer) in the field.

I'm interested in a project with 3 or 4 ASI183 cameras or similar on 400mmm to 600mm focal lengths (similar focal lengths to your area of interest) so that's why I'll be trying the Nikon 1 J5 just to see possibilities before possibly shelling out for 4 astrocams.

Best
JA

[multiweb]

You're getting better. :thumbsup:



:lol: Took me a little while to click.

[LewisM]

Meh... ma mère a enlevé les noix de mon père

[Slawomir]

One thing to keep in mind is decreased sensitivity for a sensor with smaller pixels. KAF-8300, in spite of having lower QE, will be accumulating photons at about 3 times faster rate than 183 when used with the same telescope. I'm not sure (I simply do not know) if the RN difference would level things out though.

[LewisM]

I have an SX SXVR-H9C I can compare with (even larger pixels than the 8300)

[gregbradley]

No I don't think so. My experience so far with a 183mm compared to both a KAF16200 6 micron and 16803 9 micron (I did have a KAF8300 a few times as well) is that its more sensitive especially in Ha and O111 and S11.

The KAF sensors are probably somewhere like 40% QE or less for Ha whereas the 183 is up around 70%.

Its definitely faster than the CCDs despite smaller pixels.

Greg.

[gregbradley]

A picture is worth a thousand words:

ASI183mm Pro 2.4 micron pixels on AP RHA 305 F3.8 in semi rural skies:

https://pbase.com/gregbradley/image/170949826 ASI183mm Pro

FLI Microline 16200 (6 micron pixels at a dark site):

https://pbase.com/gregbradley/image/170949827 CDK17 ML16

As you can see the 183mm image is much the same brightness despite being taken on a 305mm scope in semi rural skies versus a 435mm scope at a dark site. The CDK image should be much brighter than the 183mm but its not.

This matching pixels to optics, in my opinion, is a very limited theory and there are other factors just as important like QE, read noise, dark current, dark skies, seeing. The effects of mismatched pixels don't seem to me to be that great in reasonable seeing.

Greg.

[Slawomir]

Greg - in this scenario it would be much more fitting to compare SNR from your both cameras when operating on the same scope (same length of total exposure), as Lewis is contemplating replacing his KAF8300 with ASI183 while keeping the same scope.

I believe is you stick ASI183 onto your CDK17 you will see a much weaker signal than from the KAF16200. Alternatively, if you attach KAF16200 to your Honders I suspect the SNR will be stronger than with the ASI183-Honders combo for the same length of total exposure and for the same object.

[LewisM]

Decided to continue on with the ST-8300C and the SXVR-H9C. Both aren't tiny pixels, nor are they enormous, but I'll just try my best.

[multiweb]

Debayer in PI. It's smart enough to drizzle off the original bayer matrix. Shoot shorter subs but more of them.

[LewisM]

Agreed Marc - last time I imaged, I did that - dither, debayer, drizzle integration etc. And that was using a 9 micron sensor camera and no horrific results - well, no more horrific than my usual mess.


And the new EZ Suite really makes my mess just that little bit tidier.

[gregbradley]

Yeah I wanted to do that but I don't think I have the same data from both cameras on the same scope. I'll look again.

Greg.

[codemonkey]

The small pixels might give you sharper results, they'll almost certainly get you rounder, less blocky stars... I agree with what's already been said though, I'd try dither + drizzle first which will definitely round up those stars.

What's your typical guide error (total RMS in arcseconds)? What do you use for autofocus (hardware and software)?

[LewisM]

Get perfectly round stars. I do dither (up to 5 pixels) and drizzle integrate in PI. Here's an image from last time I imaged last year :lol: https://www.astrobin.com/full/mv7lg0/0/


Guide error.... errr.... don't you recall my guide error...essentially, well, none...


Autofocus? No way Jose! Bahtinov and Eyeball Mk.1 for me (zoomed in 300% in MaxIM 6)


Here's my guide error...yeah its in pixels, but I don't use PHD...MaxIM does a FANTASTIC job. Suavi, it is STILL not guiding on a hot pixel :P (another shot when there was a minor burp here: http://www.iceinspace.com.au/forum/attachment_browse.php?a=225766)

[Slawomir]

Fantastically flat guide graph Lewis. I presume you are still binning Lodestar 2x2 on a 150mm FL guide-scope :question:;):P

On a more serious topic - at about 2.5" pp you would most likely recover some resolution by drizzling sufficient number of dithered subs.

[LewisM]

Drizzling certainly does. Next time I actually get to image - maybe next year at this rate - I'll be doing it all the time.

[codemonkey]

lol I had forgotten about your hot pixel guiding.

I asked about focusing and guiding because in my experience these are critical to get the most out of your scope and if there was any question there, I'd be focusing (pun intended) on those before even thinking about a different camera.

Maybe you're an expert turner of focus knobs, but personally I could never manually focus as well as a properly-configured auto focus routine can do. In addition to that, if your stars aren't looking like little squares at 2.48"/px with a Tak, I'd be guessing you're probably not getting the best focus that you could. You're too undersampled for them to look perfectly round (in single subs, prior to drizzle integration).

As for guiding... it's common to assume that if your stars are round, your guiding is fine, but that only suggests that both axes are equally good (or bad). I can very easily misconfigure guiding to significantly increase my FWHM while still getting round stars. In my experience, solid guiding is the next most important thing.

If there's any question on either guiding or focusing, I'd encourage you to to investigate those first.

[The_bluester]

Regards drizzle versus smaller pixels. My old ASI294 on my 80mm is just about bang on 2 arcsec per pixel, I recently used drizzle in APP on a set of subs I had and ran it a couple of times at different settings to work out where the noise started to become objectionable. First light with my ASI2600 (3.7 versus 4.63um pixels to make it 1.6 or so arcsec per pixel) was of the same target, unguided and with spacing still to be sorted, but to me the star profiles looked at least as smooth with 20 subs and all the tracking and spacing errors as the drizzled stack of 100 subs.

I suppose it says you can get a result both ways.

[Slawomir]

Paul - I believe 3.7micron pixels are at about diffraction limit for your 80mm, so stars should be round in the subs already, as opposed to 4.6micron pixels that, with good guiding, should give somehow blocky stars. Wondering if FDHM got better (smaller) with the new camera.

[The_bluester]

I am not paying any attention to image stats yet, not until it is properly set up spacing wise and guided. Subjectively the stars look nicer already with native resolution instead of recovering it via drizzle on more undersampled data.

[Camelopardalis]

Both ways are increasing your sampling, which is why both work :thumbsup:

[The_bluester]

Yep, just now having done it both ways I reckon my preference is for good sampling to begin with rather than retrieving more data via drizzle, the noise difference is noticeable, at least to me.


The downside being that the better sampling increases data storage requirements and processing times!

[LewisM]

Zero issues with guiding or mount. Maybe my focus isn’t exact but I am not a fan of adding yet more bits on. Been there with AF before, didn’t float my boat - I like actually being out with my scope doing stuff. Automated to me takes away most of the fun of astronomy, and besides, not in it for image glory and fame.

Only reason I was considering a small pixel cam is because of my typical low total time data sets. I am happy with my round stars as is, and if only slightly under sampled given our typical good to excellent seeing here, I can live with it. Better than being over sampled :)

As Marc said, make the best of what you have.

[DavidTrap]

Have you tried putting the y-axis of your guiding graph on arcsec of error rather than pixels - lets compare apples with apples?

DT

[LewisM]

David doesn’t mean anything to me honestly and I don’t use PHD. So long as it works I don’t chase numbers.

My graph in MaxIM in arcsecond isnt much different. It’s not the mount or guiding I am even remotely interested in, it was simply can I maybe get the stars etc sharper given all else being to my satisfaction.

[LewisM]

PS- I have done 20, 45 and 60 min single subs with this same rig with zero trailing, same star shapes. Also did the same with an FSQ106ED

[Camelopardalis]

Especially true with a camera like the 183, 40MB files each, that soon clogs up the SSD :lol:

[The_bluester]

Just had a look, the 16 bit FITS from my ASI2600 are just a sniff under 50MB each. The output seems very smooth compared to the ASI294 so hopefully I can get away with reducing the number of subs.