Why is no one talking about the 2400MC Pro?

[oymd]

Why hasn’t this camera taken off?

Its full frame, and considerably cheaper than the 6200?

It’s always discounted on FLO, so seems it’s a slow seller?

Why isn’t it as successful as the 2600 or 6200?

[vlaiv]

Maybe because of people's perception of what is important in sensor?

It has large pixel size (and everyone is chasing "high resolution" these days - so prefer smaller pixel size). It is "only" 14bit ADC while 6200 is full 16bit.

It has slightly lower QE of the two - ~80% vs 91%

People using full frame sensor are either using small expensive APO refractor that can illuminate such large sensor - and that means TAK for example - so I guess money is no issue there, or are using large reflector telescope like ODK / RC - and mono has edge there.

First would rather go for 6200mc due to smaller pixel size, latter will probably go for 6200mm version.

People with scopes that can't fully illuminate full frame will naturally go for APS-C size or 2600 model.

[oymd]

2 hours ago, vlaiv said:

Maybe because of people's perception of what is important in sensor?

It has large pixel size (and everyone is chasing "high resolution" these days - so prefer smaller pixel size). It is "only" 14bit ADC while 6200 is full 16bit.

It has slightly lower QE of the two - ~80% vs 91%

People using full frame sensor are either using small expensive APO refractor that can illuminate such large sensor - and that means TAK for example - so I guess money is no issue there, or are using large reflector telescope like ODK / RC - and mono has edge there.

First would rather go for 6200mc due to smaller pixel size, latter will probably go for 6200mm version.

People with scopes that can't fully illuminate full frame will naturally go for APS-C size or 2600 model.

Makes sense what you said Vlaiv!

doesn’t that suggest that ZWO did not do enough research before releasing the 2400?

As you have nicely analysed, the 2400 is caught somewhere in no man’s land?

[vlaiv]

2 hours ago, oymd said:

Makes sense what you said Vlaiv!

doesn’t that suggest that ZWO did not do enough research before releasing the 2400?

As you have nicely analysed, the 2400 is caught somewhere in no man’s land?

I think it is down to people's perception more than anything.

14bit ADC is really no biggie. It is in no way restrictive and people happily image with 12bit ADC with cameras like ASI1600.

As far as pixel size is concerned - well, I think that people are caught up in mega pixel craze. It is due to marketing departments of both mobile phone manufacturers and DSLR manufacturers that push large mega pixel counts as something very good. I think that 6µm pixel size of 2400 is just right for OSC full frame sensor.

Such sensor is really suited to 10" or lager scopes. Say you have scope like that that is F/7 or F/8 - that will give it 1800 - 2000mm of focal length. Let's go with lower number of 1800mm.

1800mm and 6µm gives 0.69"/px - and because this is OSC sensor and we should really use super pixel mode rather than interpolation - actual sampling rate is twice that so 1.38"/px. That is very good sampling rate in my view.

If you use 6200 with 3.76µm pixel size with such focal length - you'll get base 0.43"/px. That needs to be binned x3 or x4 to get good sampling rate. How many people actually does that?

As you see, for someone with large scope that can illuminate full frame - if looking for OSC camera - 2400 makes sense - and they save money over 6200 as well. QE difference is rather small to be deciding factor.

[barbulo]

Agree, but when you go from an APS-C 18Mpx to a lower resolution CCD it is hard. I say myself: don’t fall in love with mega pixel, there are other aspects to look at. 

Edited July 12, 2021 by barbulo

[vlaiv]

12 minutes ago, barbulo said:

Agree, but when you go from an APS-C 18Mpx to a lower resolution CCD it is hard. I say myself: don’t fall in love with mega pixel, there are other aspects to look at. 

There is a link between three quantities: Megapixel count, sensor size and pixel size.

In daytime photography we don't really think about limiting resolution and we don't really think of pixel size. In astrophotography - that should really be starting point - and next should be FOV. Once you set those two things - well, you don't really have a choice of pixel count - for a given pixel size and FOV - there is only so much pixels that you can fit inside.

[oymd]

4 hours ago, vlaiv said:

There is a link between three quantities: Megapixel count, sensor size and pixel size.

In daytime photography we don't really think about limiting resolution and we don't really think of pixel size. In astrophotography - that should really be starting point - and next should be FOV. Once you set those two things - well, you don't really have a choice of pixel count - for a given pixel size and FOV - there is only so much pixels that you can fit inside.

Vlaiv, considering the 6micron pixel size, do you think the 2400 is a good match for an EdgeHD1100?

[oymd]

P.S. I wish you can simplify the image scale analysis you mentioned above. I do not understand it, and I’ve tried to read about it, but can’t get my head around it?

In many posts, I see imagers mentioning that they are imaging at x pixel/arcsecond, or something like that. 
 

How do you reach that?

For example, I have an 294MC Pro, and an Esprit 100ED AND EdgeHD1100. 
 

How do I calculate and UNDERSTAND those measurements of image scale and under or over sampling?

thanks

Ossi

 

[vlaiv]

49 minutes ago, oymd said:

P.S. I wish you can simplify the image scale analysis you mentioned above. I do not understand it, and I’ve tried to read about it, but can’t get my head around it?

In many posts, I see imagers mentioning that they are imaging at x pixel/arcsecond, or something like that. 
 

How do you reach that?

For example, I have an 294MC Pro, and an Esprit 100ED AND EdgeHD1100. 
 

How do I calculate and UNDERSTAND those measurements of image scale and under or over sampling?

thanks

Ossi

 

Formula is rather simple:

206.3 * pixel_size / focal_length

Pixel size is in µm and focal length is in mm.

294MC Pro having 4.63µm pixel size will give 1.74"/px with Esprit 100ED (F/5.5 - so FL 550mm, if I'm not mistaken), while with EdgeHD1100 it will give 0.34"/px

Calculation is straight forward. Understanding it is a bit more complicated - and it involves an easy part and a bit more difficult part.

First let's go with easy part. Resulting number represents how much of the sky is covered with a single pixel. Distances across the sky are given in angles (angular size) and angles are measured in degrees, minutes of arc and seconds of arc. Math that connects those is simple 1° = 60 arc minutes (denoted with ' so 60') and 1' = 60" (again seconds of arc or arc seconds is denoted with "). So 1° = 3600".

Arc seconds per pixel helps you quickly convert angular size of object or FOV into number of pixels. Say you have galaxy that is 24' across. That is 24' * 60 = 1440" or 1440 arc seconds across.

When you sample it at 1.74"/px using your Esprit 100 - that galaxy will be 1440 / 1.74 = ~828px across. When you sample it with EdgeHD it will be 1440 / 0.34 = ~4235px across. You can do similar calculations with your FOV - if you say have 1° FOV with esprit - what is that in pixels? 1° = 60' = 3600" = 3600" / 1.74 = ~2069px across.

It can work in "reverse" as well. Since your camera has 4144px in horizontal - if you multiply that with 1.74"/px - you'll get ~7210" or about 2° (7200/3600 = 2). Indeed if you look at astronomy.tools with that combination - Esprit 100 and 294mc, you get that FOV is:

It shows also 2° of sky.

Ok, so that is easy part - how to convert angular sizes into number of pixels and vice verse. Now comes a bit more difficult part - that is under sampling / proper sampling / over sampling.

That is very technical stuff so we need not get deep into it, but here are couple of simple "rules of thumb" that you can remember:

1. Coarser sampling - meaning larger number in "/px expression gives you better Signal to noise ratio for same aperture and integration time. In another words - if you have 100mm scope and image at 1"/px and 2"/px for an hour, 2"/px will result in better, less noisy image, but object will be smaller in the image because it will be represented by less pixels.

2. For any given image there is optimum or proper sampling rate - which simply means that detail in the image that depends on hosts of factors like seeing, mount performance, scope aperture and performance is properly matched to pixel size. Larger pixel size - coarser sampling or larger "/px number (but we call that lower sampling rate) is called under sampling. Smaller pixel size, finer sampling or smaller "/px number (which we call higher sampling rate) is called over sampling.

Under sampling - is ok - nothing wrong with that, but object is not as large in the image as it could have been - image is less noisy

Proper sampling - is ideal balance between captured detail and noise in the image

Over sampling - is bad - it will not capture any additional detail because there is none but it will be noisiest of the three

To recap: Under sampling and Proper sampling - good, Over sampling - bad

How to know if you are over sampling or under sampling or sampling "real good" ? Well - you can't know until you already captured the image as every night is different - there are nights of good seeing and nights of bad seeing and that changes sharpness of the image. What you can do is figure out "average" sharpness of the image and then dial your sampling rate (or arc seconds per pixel) to match that.

This brings us to third rule of the thumb:

- scopes up to 70mm should not sample with sampling rates higher than about 2.5"/px

- 80mm scopes should not sample higher than 2"/px

- 100mm scopes should not sample higher than about 1.6"/px

- 150mm scopes should not go higher than about 1.4"/px

- attempt high resolution imaging with 8" or larger apertures on very good mount, and by high resolution imaging I mean ~1.0 - 1.2"/px. 99% of people will not be able to reach 1"/px in 99% of the time. You really need excellent scope, excellent sky and excellent mount in order to get image sharp enough to sample it at 1"/px

In the end - there is simple relation between FWHM of the stars in the image and sampling rate. For given FWHM of stars in arc seconds - you want to sample at maximum FWHM / 1.6 sampling rate. Anything over that will result in over sampling and lower quality (this only applies to long exposure imaging and not planetary - that is another game and has different rules).

If you are over sampled - like you are with EdgeHD 11" and ASI294 - you can bin your pixels in order to reach good resolution. That recovers SNR lost due to using smaller pixels (binning is like using larger pixels). With 0.34"/px - you can easily bin x4 to get to 1.32"/px (which in case of OSC camera means super pixel mode + software bin x2).

 

 

[Seelive]

32 minutes ago, oymd said:

P.S. I wish you can simplify the image scale analysis you mentioned above. I do not understand it, and I’ve tried to read about it, but can’t get my head around it?

In many posts, I see imagers mentioning that they are imaging at x pixel/arcsecond, or something like that. 
 

How do you reach that?

For example, I have an 294MC Pro, and an Esprit 100ED AND EdgeHD1100. 
 

How do I calculate and UNDERSTAND those measurements of image scale and under or over sampling?

thanks

Ossi

 

Without going into the maths, the (close) approximate formula for determinating the arcsec per pixel is:

206 x camera pixel size / telescope focal length

The formula requires the camera pixel size to be in micrometres (um) and the telescope focal length in millimeters (mm). So for your 294MC and Esprit 100 it would be:

206 x 4.63 / 550 = 1.73 arcsec / pixel

[oymd]

33 minutes ago, Seelive said:

Without going into the maths, the (close) approximate formula for determinating the arcsec per pixel is:

206 x camera pixel size / telescope focal length

The formula requires the camera pixel size to be in micrometres (um) and the telescope focal length in millimeters (mm). So for your 294MC and Esprit 100 it would be:

206 x 4.63 / 550 = 1.73 arcsec / pixel

Thanks. 
I know the formula. 
it’s understanding it properly and applying it to my imaging is what I struggle with. 
 

But @vlaiv has done it again and provided a comprehensive read that I’m about to read and study!

[newbie alert]

Without going into too much details, sampling is to do with you focal length and pixel size...

With regards to the sensor size you need to look at the image circle of your scope and adapter sizes, too large and it will vignette

Can your seeing conditions support your sampling that you're trying to achieve

[ollypenrice]

The crux of arcseconds per pixel is that it states how much sky lands on each pixel. The more sky, the more light - but the less resolution of detail.

I'd have this camera in a heartbeat, even though the just-under 6 micron pixels would be a little large for my Tak FSQ106. They'd still offer a 33% gain in resolution over my full frame CCD, though, and that still produces nice images at 3.5"PP. Also, that sampling rate would make it fast.

It would be a match made in heaven with my TEC140, though.

Relatively few optics will cover full frame and folks who do own such optics are likely to want full narrowband capability so this camera will not appeal to all who might be in the market for full frame. The other thing is that you need the computing power to stack and calibrate all those large files.

Olly

[DaveS]

This QHY Camera might be an interesting alternative.

[gorann]

Are you sure that more ASI6200MC than ASI2400MC are sold? I suspect that most people that decide to invest that much money in a full frame CMOS will go for mono (= the ASI6200MM) to get more versatility. That is at least what I did, and for OSC CMOS I have gone for the ASI2600 (perfect for RASA 8 for example). The smaller pixel size also gives more versatility since it can be binned. I can for example use my ASI6200MM binned 2x2 (=7.5 µm pixels) or even 3x3 (11 µm pixels) when I hang it on my 14" Meade ACF at 3.5 m focal length, while I use it unbinned on my Esprti 150.

Edited July 15, 2021 by gorann