ZWO ASI224 Electrons per data

[cuivenion]

I want to use the SubframeSelector for selecting the best lights and weighting. I'm using the ZWO ASI224 for DSO and I had a conversation here:

https://stargazerslounge.com/topic/255509-what-settings-for-zwo-asi224/

Short story is, apparantly bias frames can't be used for calibration with this camera because the Sony sensor has an internal calibration that kicks in when the exposure time changes. This means I have to take darks or dark flat immediately after lights or flats without changing the exposure time.

I'm assuming that I can't use bias when using the Basic CCD parameters script to determine my Electrons Per Data, so I'm wondering what to do?

Any help appreciated.

[cuivenion]

Hi in addition, I have found this table on cloudy nights that seems to give me my answer:



Does this look right? Does the electron per data change according to the gain used? Or is it a single value for the camera?

[bulrichl]

Hi cuivenion,

--- from stargazerslounge ---
Last tip - don't use bias frames - take as many dark frames as you can at exactly same settings - without touching any of controls - just cover telescope with a cap.
-----------------------------

What he probably wanted to say is: he advises against
scaling the MasterDark (in PixInsight: use of Dark frame optimization). And if you don't use Dark frame optimization you don't need to take bias frames because they are not needed for calibration.

ZWO has published the data for the ASI224 on this website:
https://astronomy-imaging-camera.com/products/usb-3-0/asi224mc/

The table that you found on cloudynights contains these data, only "Offset" and "MinDN16bit" are additional.

Read noise is calculated from the difference of 2 bias frames, so yes, of course: if you want to calculate the read noise for your specific camera, 2 bias frames (per gain-setting) are needed (e.g. for the application of the Basic CCD parameters script).

Bernd

[cuivenion]

Hi, dark frames won't remove amp glow if you optimize them. I don't think hes referring to that. He is is quite specific in saying that there bias changes every time the exposure or gain is changed.

With the above table would my Electrons Per Data be 0.80 because I used 150 gain? What do I do if I've used two sets of data at different gains?

[bulrichl]

Hi, dark frames won't remove amp glow if you optimize them. I don't think hes referring to that. He is is quite specific in saying that there bias changes every time the exposure or gain is changed.

OK, and did you check what he suggests? Are the results that bad?

I don't own a ASI224 but a ASI294. However, with the ASI294 offset is applied automatically, and there is no problem at all with changing bias.

With the above table would my Electrons Per Data be 0.80 because I used 150 gain? What do I do if I've used tow sets of data at different gains?

Roughly, yes. According to the graph on the ZWO website it is approximately 0.9 electrons/ADU at a gain of 150.

Two sets of data at different gain of course would have to be calibrated separately, with Masters that are generated from calibration frames with suitable gain/offset (same settings as at light frame acquisition).

In practice you will not change gain continually, but you will have to generate Master calibration frames for some exposure times that are used more often.

Bernd

[cuivenion]

Hi, I did subtract one set of bias from another as suggested. There was one column still there along with the random noise.

[bulrichl]

There was one column still there along with the random noise.

What's that supposed to mean? I can't make head or tail of it.

Bernd

[cuivenion]

Hi. I followed the advice about comparing bias in the Stargazers Lounge thread. Here's what I did:

1. Took a set of ten bias.
2. Shut down the camera.
3. Restarted the camera, took another set of ten bias.
4. Integrated both sets using these settings:

5.Subtracted integration2 from integration1 and integration2 from integration2 with these settings:

6. These are the results side my side. They're very different, when they should be identical:

[sharkmelley]

Did you really do the following?
       5.Subtracted integration2 from integration1 and integration2 from integration2

If so then integration2-integration1 will be random noise with a lot of clipping to zero and integration2-integration2 will be precisely zero.  So I'm guessing I misunderstood you 

In any case try running the Statistics process on both integrations so we can see mean, median, stdDev etc.

That bright column is interesting though.

Mark

[cuivenion]

Sorry, to clarify:

I subtracted integration2 from integration1.

And then separately subtracted integration2 from integration2.

Both were separate processes.

[cuivenion]

Here's the statistics:

[sharkmelley]

Let me try and clarify what I meant.

After ImageIntegration you had two files:  Integration1.xisf and Integration2.xisf

Can we see the statistics on those two files before you do the bias subtraction.  And if you click on the spanner icon of the Statistics Dialog you can select stdDev in addition to avgDev.

Mark

[cuivenion]

No problem thanks for the help. Here's the files:

[cuivenion]

Just to add, each stck consisted of 10 frames.

[bulrichl]

The statistics look perfectly similar. In order to generate the difference try the following PixelMath equation:

integration1-integration2+0.1

Let PixelMath create a new image, 'Diff'.

Take a look at the histogram of 'Diff' in HistogramTransformation with suitable zoom setting. The histogram should look like a Poisson distribution, in particular it should be be symmetric respective to the y axis. If this is the case, everything is fine.

Bernd