NINA Flat Wizard Issue

[MikeeJC]

I'm experiencing a strange issue when starting to take dark flats in NINA's flat wizard.  Having looked back I realise it's always happened but only just noticed after switching to ASIFitsView to check the images prior to stacking.

The first dark flat taken always has a green bayer matrix visible when stretched which the other frames don't.  Also the histgram is showing elevated figures suggesting there's some light present.  Although the scope is covered with its lens cap and a folded dust sheet, also taken at night so there's no light entering the OTA.  If there were light leaking in then all the darks would be the same.

I'm using a RisingCam IMX571 OSC camera which appears to working well.

I know I can just delete this single 'incorrect' dark flat frame but it would be nice to figure out why this is happening.  Any ideas greatly appreciated.  Not sure if it's a NINA or camera issue.

I've attached an image of the first flat dark and also the subsequent one to illustrate.

Clear skies!
Mike

Edited December 4, 2022 by MikeeJC

[The Lazy Astronomer]

I'm not sure what's going on here, it could be a NINA or driver problem, but your second image has average and minimum values of 0, which implies your offset is too low. There should not be 0 value pixels even in a dark frame.

[Gary Clayton]

Are you using the native or the ASCOM driver, try swapping them to eliminate driver issue. 

[MikeeJC]

On 05/12/2022 at 14:01, The Lazy Astronomer said:

I'm not sure what's going on here, it could be a NINA or driver problem, but your second image has average and minimum values of 0, which implies your offset is too low. There should not be 0 value pixels even in a dark frame.

Hi.  I run the gain at 100 with an offset of 20 which seems to be popular for the IMX571 sensor and the images come out pretty good.  I will have a play at changing them a bit and see how the average and minimum values change.  Thanks.

On 05/12/2022 at 14:11, Gary Clayton said:

Are you using the native or the ASCOM driver, try swapping them to eliminate driver issue. 

Hi Gary.  I'm using the native driver as from what I can remember (it's been a while) it had better functionality and control compared to the ASCOM driver.  That said, I'll try changing to the ASCOM driver and see if I get the same result.  Thanks.

[ONIKKINEN]

12 minutes ago, MikeeJC said:

I run the gain at 100 with an offset of 20 which seems to be popular for the IMX571 sensor and the images come out pretty good.

20 is way too low, stick to the default 768 offset to make sure no pixels are at 0. There are no real world downsides to having that extra offset as you can afford to lose that few hundred ADUs from the 65k the camera has to work with. Assuming you are using HCG mode too? If not, maybe better to, as LCG mode has limited benefit compared to HCG and only in some very specific situations (if you dont know you need it, you will not need it or benefit from it). It doesnt really matter if the images come out good or not, 0 value pixels are physically impossible to calibrate properly and you will have problems with flats and darks calibration if you have them (cant divide or multiply by 0 so software used for calibration uses a very small number, like 0.0001 or something, but not 0. Its a problem because the real pixel value could be -100 or -200 but there is no way to know since it was clipped).

I would still recommend taking all darks indoors with the camera off the scope and completely covered as just a few photons creeping in somewhere will ruin them. As for why this happened, not sure. I have noticed that sometimes immediately after shining something bright down the tube, like a flat panel, the next frame will have some ghosting of signal as if the light was still partially there. This is always the first shot after, so was your failed one the first darkflat after taking the flats? Dont know of a solution to this, other than checking each first shot of a sequence when taking darks but for what its worth this happens quite rarely.

[MikeeJC]

49 minutes ago, ONIKKINEN said:

20 is way too low, stick to the default 768 offset to make sure no pixels are at 0. There are no real world downsides to having that extra offset as you can afford to lose that few hundred ADUs from the 65k the camera has to work with. Assuming you are using HCG mode too? If not, maybe better to, as LCG mode has limited benefit compared to HCG and only in some very specific situations (if you dont know you need it, you will not need it or benefit from it). It doesnt really matter if the images come out good or not, 0 value pixels are physically impossible to calibrate properly and you will have problems with flats and darks calibration if you have them (cant divide or multiply by 0 so software used for calibration uses a very small number, like 0.0001 or something, but not 0. Its a problem because the real pixel value could be -100 or -200 but there is no way to know since it was clipped).

I would still recommend taking all darks indoors with the camera off the scope and completely covered as just a few photons creeping in somewhere will ruin them. As for why this happened, not sure. I have noticed that sometimes immediately after shining something bright down the tube, like a flat panel, the next frame will have some ghosting of signal as if the light was still partially there. This is always the first shot after, so was your failed one the first darkflat after taking the flats? Dont know of a solution to this, other than checking each first shot of a sequence when taking darks but for what its worth this happens quite rarely.

Hi Onikkinen.

Thanks for the info.  Yes, I always use HCG mode.  I haven't tried LCG despite the sensor analysis stating you get a much higher full well depth in LCG at the expense of much higher read noise.

My dark frames have been taken indoors completely covered as you mention and all look good (apart from min=0 also) but the dark flats are taken outside as they are straight after the flats, which I take just before a session as this is more convenient than after as I usually leave it running overnight while I'm in bed.  As you mention, it is always the first dark flat frame taken directly after the flats which has these elevated values.

I'm a little dubious of the figures ASIFitsView histogram is giving compared to the visual histogram, although that could be me just not understanding them correctly.  I've attached a couple of screenshots of a light and flat frame showing FitsView histogram.