Atik 490ex question

[bottletopburly]

Having just acquired an Atik 490ex mono  and ccd is new to me but what gain and offset do you set it at , currently setting up a master dark and bias library.

[Budgie1]

You can't set the gain & offset with the Atik 4-series CCD cameras, they are factory set and you can't change them.

Something less to worry about.

It does take a bit of getting used too if you've been used to a CMOS camera, but it does make life easier and you may also find you won't need Darks with the 490EX either.

I currently have the Atik 428EX mono and the 490EX is my next planned large purchase.

[bottletopburly]

5 minutes ago, Budgie1 said:

You can't set the gain & offset with the Atik 4-series CCD cameras, they are factory set and you can't change them.

Something less to worry about.

It does take a bit of getting used too if you've been used to a CMOS camera, but it does make life easier and you may also find you won't need Darks with the 490EX either.

I currently have the Atik 428EX mono and the 490EX is my next planned large purchase.

Thanks budgie  all a learning curve, just getting my head around it I had come to the conclusion you couldn’t set it using Apt as gain /offset shows as N/A..

[carastro]

Agree with Budgie.  
 

 

[windjammer]

The ccd in the 490 is a sony icx814 - the well depth is 12,500e.  So a 16 bit a/d  will have 5 counts for every electron collected (65,636/12,500).  Put another way, the 16 bit space can count every electron directly with no need for analogue gain or offset.  These can be provided in your processing software.  Arguably, the camera is oversampled.  I have a 460, well depth 18,000 so 4 counts per e.  Depending on your view of Nyquist, you need two bits to count each e.

Cmos cams have been moving up to 50,000e so a 12 or 14 bit a/d needs gain/offset to move the sampling range to where the data is.  And even the 16bit cmos cams still have gain controls.  The whole world is waiting to move past 16 bit!

You will need darks IMHO, and if my 460 is similar, definitely if you bin beyond 1x1.

Simon

[bottletopburly]

6 hours ago, windjammer said:

The ccd in the 490 is a sony icx814 - the well depth is 12,500e.  So a 16 bit a/d  will have 5 counts for every electron collected (65,636/12,500).  Put another way, the 16 bit space can count every electron directly with no need for analogue gain or offset.  These can be provided in your processing software.  Arguably, the camera is oversampled.  I have a 460, well depth 18,000 so 4 counts per e.  Depending on your view of Nyquist, you need two bits to count each e.

Cmos cams have been moving up to 50,000e so a 12 or 14 bit a/d needs gain/offset to move the sampling range to where the data is.  And even the 16bit cmos cams still have gain controls.  The whole world is waiting to move past 16 bit!

You will need darks IMHO, and if my 460 is similar, definitely if you bin beyond 1x1.

Simon

Thanks I’ve been creating a bias and dark’s library all night 1x1and 2x2 .

[bottletopburly]

@Budgie1  @windjammer  would I be correct in thinking the Adu value for the 490ex would be approximately 8000 Adu  for taking flats ?

 

Dave 

[Starflyer]

36 minutes ago, bottletopburly said:

@Budgie1  @windjammer  would I be correct in thinking the Adu value for the 490ex would be approximately 8000 Adu  for taking flats ?

 

Dave 

With my 460EX I used to aim for 1/3 to 1/2 way along the histogram, 22000 to 30000.

[Budgie1]

3 hours ago, bottletopburly said:

@Budgie1  @windjammer  would I be correct in thinking the Adu value for the 490ex would be approximately 8000 Adu  for taking flats ?

 

Dave 

In APT I set the target ADU to 26,000, I can't remember what settings I used in NINA because I've only used it once, but it was similar settings to APT. 

[windjammer]

>>With my 460EX I used to aim for 1/3 to 1/2 way along the histogram, 22000 to 30000.

I use similar with my 460 - perhaps a little higher, 50% to 60% of full scale reading using the native Atik software, Artemis capture.

If your ADU scale is 0 to 65,536 (ie 16 bit), and full well is 65,536 then an adu count 30% to 60% of the full well value is what you are looking for.  If your ADU scale is different, then shoot for 30% to 60% of the saturated, full well value.

So, your 490 is 16 bit: the Atik software would have its histogram 0 to 65,536. So 20k to 40k counts would be fine.

I can see there is room for confusion between the well depth in terms of electrons, and the pixel value measured in terms of what the 16bit A/D returns.  I would guess the 'ADU' value your are looking for in your software is the A/D value in range 0-65,536.  So 20k to 40k counts - 8k sounds far too low unless your ADU scale is totally different.  What is the full scale deflection value - that would be the clue?

HTH

Simon

 

 

[bottletopburly]

2 hours ago, windjammer said:

>>With my 460EX I used to aim for 1/3 to 1/2 way along the histogram, 22000 to 30000.

I use similar with my 460 - perhaps a little higher, 50% to 60% of full scale reading using the native Atik software, Artemis capture.

If your ADU scale is 0 to 65,536 (ie 16 bit), and full well is 65,536 then an adu count 30% to 60% of the full well value is what you are looking for.  If your ADU scale is different, then shoot for 30% to 60% of the saturated, full well value.

So, your 490 is 16 bit: the Atik software would have its histogram 0 to 65,536. So 20k to 40k counts would be fine.

I can see there is room for confusion between the well depth in terms of electrons, and the pixel value measured in terms of what the 16bit A/D returns.  I would guess the 'ADU' value your are looking for in your software is the A/D value in range 0-65,536.  So 20k to 40k counts - 8k sounds far too low unless your ADU scale is totally different.  What is the full scale deflection value - that would be the clue?

HTH

Simon

 

 

Thanks Simon I was unsure what I was aiming for I will try 23k .

 

 

Dave