Which gain for a CCD in Canopus and AAVSO batch ?

Affiliation
American Association of Variable Star Observers (AAVSO)
Tue, 01/05/2021 - 21:57

Dear all,

Could you help me abouta question about the gain of a CCD.

I use Canopus to dp my aavso reports.

In the configuration it is asked to enter the gain of the camera iin e/ADU to calculate S/B ratio.

With my CCD Atik 16 HR, the gain is 0,3 e/ADU by the manufacturer

The Chip Kaf 8300 has 0,37 e/ADU (recent CCD and expensive)

And the ST7, which is an old camera has a gain of 2,3 e/ ADU.

The difference is of a factor around 8.

This lowers the S/B ratio, as gain is a parameter of S/B ratio

I did photometry with canopus and get 0,015 mag error with 2,3 e.ADU and 0,55 mag error with 0.3 e/ADU.

The star was mag 8, not faint, really bright.

How is it possible that an old camera could have a less gain than new cameras.. ?

I saw in different discussion that some speak of the inverse of the gain.

Maybe the gain of some CCD could be the inverse of the gain of other cameras ?

Thanks

Jean

 

Affiliation
American Association of Variable Star Observers (AAVSO)
I measured the ADU from…

I measured the ADU from exposure time

 did a graph with ten exposures and find a saturation at 55000 ADU

Knowing that FWC is 19500 electrons, so 1/gain is 55000 / 19500 = 2,8 ADU/e-

and gain would be 1/2.8 = 0.3.

So it seems that 0.3 be the real gain.

But that means too that FWC of ST7 being around 80 000 and FWC of 16 HR is 19 500

the FWC of an old camera is better than a newer camera... 

For a given exposure, S/B of ST7 is better than S/B atik 16HR... how is it possible.... ? 

It is very strange for me...

Are my calculations right  ?

Jean

Affiliation
American Association of Variable Star Observers (AAVSO)
Gain and anti-blooming

Jean,

I'm not sure what you mean by S/B, but I think the main problem is that you are comparing a non-anti-blooming camera with large pixels to an anti-blooming camera with small pixels.  Smaller pixels have smaller Full Well Depths, and anti-blooming CCD's make the usable well depth, for photometry, even smaller. 

The KAF 8300 is a large anti-blooming CCD with small pixels.  The anti-blooming circuitry will cause the response to become non-linear well before the actual FWD is reached.

To keep the camera response linear, manufacturers can set the gain so that the Analog to Digital Converter is saturated before the chip response goes non-linear.  For example, with the gain set to 0.37e per ADU a 16-bit ADC will saturate at about 24,000e.  Thus, cameras using the 8300 anti-blooming chips can do good photometry, but at the expense of decreased dynamic range when compared to NAB cameras. 

Most ST7 cameras used in photometry are non-antiblooming.  The NAB KAF 400 chip is usually specified as having a FWD of 100,000e.  The manufacturers of NAB cameras usually set the gain so that the FWD of the pixels and ADC saturation happen at about the same level.  An unbinned KAF 400 CCD with a gain of 1.5  will saturate a 16bit ADC at ~98,000e. 

Phil

 

Affiliation
American Association of Variable Star Observers (AAVSO)
thanks Phil, I understand…

thanks Phil, I understand what you mean, but my camera in NAG, as the ST7

I wonder why old camera as the ST7 (years 1990) have a gain of 2,3 e/ADU and the newer camera like Atik 16HR (years 2010) and others have a gain of 0,3 e/ADU/.

There is a huge difference, my atik has a gain 8 times less than ST7...

 

Affiliation
American Association of Variable Star Observers (AAVSO)
Gain

...sorry Jean.  I misunderstood your original post. 

Consider the KAF 400 chip with 9 micron pixels as used in the ST7 :      SBIG gives 100,000e as the Full Well Depth.  My ST-402 (NAB KAF 402 CCD) camera has a gain of 1.47e per ADU when unbinned.  Knowing this we should expect the pixels to saturate at about 68K ADU's.  ( 100K  /  1,47 ).

Since the ADC saturates at 65K, the ADC will saturate well before the pixels.  (When the pixels are binned, the camera automatically changes the gain to 2.3.)

If the gain on a NAB ST7 is set to 2.3 we should expect the pixels to start saturating at about 43K  ( 100K / 2.3 ), long before the ADC saturates.  If you do a careful linearity test on the ST7 you should see the response start to fall off after this level is reached.  In the plot, the trend line will start to curve away from the linear response.  In their ST class camera manuals, SBIG states that the ST7 is non-linear above 40K ADU's.

If your NAB Atik camera has a FWD of 19,500e, I would expect that the pixels are relatively small compared to the ST7.   With a gain of 0.3, the pixels will saturate at 65K ADU's  ( 19,500 / 0.3 ).  This is the gain that causes the pixels and the ADC to saturate at about the same ADU count.

Small pixels give small FWD's.  Small FWD's need low gain settings to keep the response linear until the ADC saturates.

Phil