ZWO 1600MM Super Bias - Issue

[Mark Haythorne]

A number of firsts here for me.

New filters/filter wheel, oag, guide camera, relative new 2nd hand scope & Mount.

And the other night everything worked a treat, didn't take long to actually start imaging, until my filter wheel hit the tripod legs and so I rotated it 180 deg (and is now not correctly framed, given me a chance to practice in PI), and then started again, DOH no flats for original shots.

 In the end I managed a few Ha and OIII captures....

Then took Flats, Flat Darks, Darks & Bias.

Created Master Dark, Master Dark Flat & Master Bias, then a Super Bias, and then a Master Flat.

I've just bought PixInsight and Warren Kellers Book 'Inside PixInsight'. and having followed it through to the point of Cosmetic Correction (hadn't opened an image after calibration, noob alert) and was shocked to see the following image.

 

 

I read on here that the 1600 doesn't need a Master Bias, so I went and did the whole thing again, without the bias obviousily and got the following, nothing else done to it apart from calibration... Phew

 

Edited November 20, 2019 by Mark Haythorne

[wimvb]

A master bias is only needed if you want to scale darks that were taken at another exposure than the lights. But with cmos you can't scale darks, so you don't need a master bias. 

Only use darks, flats and flat darks. The latter to calibrate the flats. Any bias signal is already in the darks and flat darks as well as their corresponding lights. Pixinsight now has an option to use flat darks in the batch preprocesser, afaIk. 

[blinky]

What is a dark flat and how do you take one?

[wimvb]

A flat dark is a dark frame taken at the same settings as the flat frame that it will calibrate.

You take them the same way you take darks, but use the exposure, gain/iso, and temperature that you used for the flats. 

[blinky]

ah, ok got you

[Laurin Dave]

Mark

Nice Bubble and M52..   Whilst better your last image shows some signs of amp glow (the two slight glows on the right hand edge) which with the asi1600 should calibrate out if your darks match your lights exactly in exposure length temperature gain and offset..   I’d check two things firstly that they do and secondly that you have  “optimise”  unchecked in Pi’s image calibration process.. 

hth 

Dave

[sharkmelley]

13 hours ago, wimvb said:

But with cmos you can't scale darks, so you don't need a master bias. 

Can you clarify what you mean here.  I agree that with set-point cooling, darks don't need to be scaled (if you match exposure times).   But that doesn't mean dark scaling can't be done for CMOS. 

Mark

Edited November 23, 2019 by sharkmelley

[wimvb]

1 hour ago, sharkmelley said:

Can you clarify what you mean here.  I agree that with set-point cooling, darks don't need to be scaled (if you match exposure times).   But that doesn't mean dark scaling can't be done for CMOS. 

Mark

You can do the scaling, of course, but the general experience of cooled cmos users (ZWO mainly) is that scaled darks won't calibrate out amp glow properly. So far I haven't heard about anyone saying "scaled darks work for me." A more cautious phrasing in my post should probably be "scaled darks generally won't work  for CMOS".

[vlaiv]

4 hours ago, sharkmelley said:

Can you clarify what you mean here.  I agree that with set-point cooling, darks don't need to be scaled (if you match exposure times).   But that doesn't mean dark scaling can't be done for CMOS. 

Mark

Next to issues with amp glow (which is not really amp glow in CMOS) there is often issue with bias levels.

I think it has something to do with the way cmos cameras work / fast readout. There appears to be two different modes of readout that one can't control - at least that is what I've read (might not be related to this issue after all) - short one, less then 1s where timing is kept in sensor itself and long one - where application controls exposure length.

In any case, bias levels can be different in these two regimes (or in general - bias level can depend on exposure length). Not all cameras show this behavior, but I had such issue with my ASI1600 - bias sub had larger mean ADU then 10s dark sub - and that of course should not be the case.

Another thing that can happen, but I doubt it happens with serious CMOS cameras, I only had it with guider cams and small sensor - automatic bias offset. Sensor does internal offset calibration on power up. That means that bias is not stable across bias power cycles and that you need to always take bias / darks at the same time (or rather after) lights / flats. No reusing of subs for later.

Because of all of those issues - one should not scale darks with CMOS unless they determine that they can do it properly with their sensor data (best way would of course be to take two sets of darks and one set of bias and try dark/dark calibration and examine result for any DC offset and patterns - it should be pure noise with 0 mean ADU).

One way to do dark scaling is to use two sets of darks of different exposure and not dark and bias. For example 2 min dark and 1 min dark can be used to get: bias as 2 * 1_min - 2_min and dark current only as 2_min - 1_min. After that it is easy to do scaling.

Edited November 23, 2019 by vlaiv
correction

[George 47]

Mark,

Do try the new weighted batch preprocessing script as it is a real improvement  on the older BPP.

I thought the problem with dark scaling CMOS is there are two populations of pixels which accumulate at different rates,  the dark current and the FPN. Easier to just use same time dark exposures.

Edited November 23, 2019 by George 47

[wimvb]

All this hassle with bias/no bias, dark scaling or not, is why I've never really tried anything other than calibrate my lights and flats with their respective darks. Seems to work ok. Actually, I took bias very early on when I got my ASI camera, but very soon settled on darks only. 

[gorann]

29 minutes ago, wimvb said:

All this hassle with bias/no bias, dark scaling or not, is why I've never really tried anything other than calibrate my lights and flats with their respective darks. Seems to work ok. Actually, I took bias very early on when I got my ASI camera, but very soon settled on darks only. 

Well, if you keep your optical path clean, checking and air-puffing each filter before the session, and have a relatively small chip (like the ASI1600) in relation to the image circle of the scope, you do not need anything but darks.