Need help with QHY268M and Narrowband Flat Field Calibration

I have the 268M, QHY filter wheel, but Optalong filters and have not had this issue, though I only have Ha for NB. If it's the camera, and the problem disappearing when you swap it out seems to point that way, is it possible that the problem exists with broadband filters too, but is 'buried' under the wider signal response?

Cheers,
Scott

Almost looks like it's reflecting light from behind a filter and back at the camera.   That honestly looks like an outline of the sensor.  I'd say Jeff's idea above is a very solid possibility.

For what it's worth, I, as well have never had any issues with narrowband flats, I hardly even needed them until I started picking up dust.

Curious, could the filter wheel be out of sync?  Have you looked down the tube of your scope, at the sensor and made sure there's not a blockage caused by the filters?

Pursuant to Jeff's post above, I had the exact same issue from that missing plug that is supposed to fill that hole. Just to clarify, that hole actually isn't for the attaching the desiccant tube, it is for some other unspecified purpose, as the desiccant tube is threaded, which this hole is not.

My little black plug that is supposed to fill this hole was missing on arrival which I did not realize and so I had lots of issues from this light leak, and interestingly, it REALLY showed up dramatically for my Hα filter compared to my other filters. This little hole that Jeff shows above does not even exist in the design of the OSC version of the camera, and no mention in any documentation from QHY about it. Mine is taped over now and all is well with all my frames (NB flats included…)

ML

Scott Badger:
I believe that hole is to allow blowing air across the glass window in front of the sensor if dew forms on it and/or fill the chamber with argon to help prevent dew. See this thread:
https://www.cloudynights.com/topic/859159-calling-all-qhy268m-users-light-leak/

Cheers,
Scott

Yes, that is the thread I started when I had the issue. The theory about the hole being used for dew reduction or argon filling is interesting, but I never found any official documentation verifying that fact. In any case, pursuant to this thread, this hole, if unplugged, is able to cause artifacts that resemble what the OP is dealing with. 

ML

Lynn K:
Don't use Bias subtraction on your Flat frames. Not sure why , but it screws up the QHY268 flats. Also, flats can be clipped to the white, resulting in total saturated areas.  

Make sure your black point pedestal is set high enough. I use MaxinDL, other programe may set this in other ways. I just had large white arches in my flats. I set the pedestal higher and is solved the problem. The black will be darker with narrowband,  and that be more of a problem, not sure though.

Lynn K.

I’ve had the exact opposite experience with the 268M - NOT using bias subtraction causes problems with flats. Although I technically do “flat darks” and match the length of my flats to the bias.

Sorry for the delay in responding to these posts. I've been slammed with responsibilities at work and at home. Here on the East Coast it's been cloudy for weeks. Supposed to clear up tomorrow night here in North Carolina. We go through so much planning and waiting for the right conditions, only to have it wasted due to hardware issues that seem to be beyond our control. 

I'm hoping QHY will have a solution to this issue. Otherwise, I'll just part ways with the camera, stick with ZWO in the future, and try to be more careful going forward. 

Brian P
Curious, could the filter wheel be out of sync?  Have you looked down the tube of your scope, at the sensor and made sure there's not a blockage caused by the filters?

l've watched them rotate into place. I even removed my OAG from the equation. No obstructions in the light path. Everything lines up properly. 

Jeff Reitzel:
Well I was hoping for an easy fix for you.  Other possibly dumb questions on my part but just brainstorming a bit. I am not familiar with the Antlia NB 36mm filters so are they threaded filters by chance? Do they have filter masks on them? Are the filters seated well in the filter slots regardless of how you mounted them? What you show looks so much like light reflecting off the back of the filters I'm just trying to guess possible sources of unfiltered light getting to the camera side of the filters. Including using threaded filters and not putting a screw in each of the 3 holes around each filter that goes completely through the filter wheel. Have you tried turning the NB filters over to see if it improves? To me it does not make sense that this issue would only be there for NB if the camera were faulty. Scott may be right that it is masked with a stronger broadband signal. 
Best of luck,
Jeff

The 36mm filters aren't threaded, they are held in place by a couple of set screws with little cardboard washers. The 36mm filter wheel doesn't have any threads where the filters go. I've tried flipping the filters over, didn't seem to make a difference. 

I do agree that the uneven illumination is being masked to some degree by the broadband signal. Now that I know what I'm looking at, I do see evidence of the uneven illumination. This QHY268M produces great results with LRGB, but broadband calibration is required! No way this camera can be used without flats as other have suggested. I suspect this is connected to the microlensing errors, which causes uneven illumination across the sensor. 

Lynn K:
Don't use Bias subtraction on your Flat frames. Not sure why , but it screws up the QHY268 flats. Also, flats can be clipped to the white, resulting in total saturated areas.  

Make sure your black point pedestal is set high enough. I use MaxinDL, other programe may set this in other ways. I just had large white arches in my flats. I set the pedestal higher and is solved the problem. The black will be darker with narrowband,  and that be more of a problem, not sure though.

Lynn K.

I noticed the BIAS frames from my QHY268M looked strange compared to my ASI1600. They looked more white than black, if that makes any sense? I was suspicious of them at first and stopped using BIAS frames for a while. I investigated everything I could think of, even isolated the power source for the camera, thinking it was picking up interference, but it wasn't. After trying out a few different processing methods I started using BIAS frames again. I can say it's not hurting anything, This camera has so little dark current that they are almost optional. 

Lynn K:
Don't use Bias subtraction on your Flat frames. Not sure why , but it screws up the QHY268 flats. Also, flats can be clipped to the white, resulting in total saturated areas.  

Make sure your black point pedestal is set high enough. I use MaxinDL, other programe may set this in other ways. I just had large white arches in my flats. I set the pedestal higher and is solved the problem. The black will be darker with narrowband,  and that be more of a problem, not sure though.

Lynn K.

I’ve had the exact opposite experience with the 268M - NOT using bias subtraction causes problems with flats. Although I technically do “flat darks” and match the length of my flats to the bias.

I think that's the best approach, but for simplicity I've increased the exposure time for my BIAS frames. I used to go as low as 0.0001, but now 0.1 seems fine. It's hard to get Luminance flats above 1 second, normally 0.6 or so. Also seemed to make my BIAS frames look more normal to me, though can't say if it's better. 


I tried narrowband SkyFlats again, just to experiment, tried at different brightness levels, different angles, and with and without a light diffusor. The only noticeable difference was if I set the gain to 0. If I had to quantify it, might be ~10% improvement. The SkyFlats basically look the same as the ones made from my Light Board.

Scott Badger:
A couple things.....

Flats are never optional. Darks, maybe with some cameras like the 268M, but that hasn't been my experience, and flat darks can replace bias. Flats take care of vignetting, donuts etc. Without them you could probably deal with the vignetting with a tool like DBE, but the dust donuts are going to require clone stamping, and we all know that's a road to perdition..... : )

I believe that when you autostretch a bias frame, because of the way autostretch works, the range of dark gray to nearly white depends on how much of a 'pattern' and noise there is in the signal. Less pattern/noise, more white. Better sensor?

I've only heard that you should set the bias exposure as low as possible. In NINA I set it to 0, which results in an exposure much less than 0.1s, though I don't have a sub handy to say what it is exactly.

Why are you trying to get flats above 1s, or targeted to any particular exposure time? Exposure time is immaterial, it's where the peak of the histogram falls (about in the middle) that matters.

Flat darks vs bias. Lots of conflicting advice out there, so I can't comment conclusively, but I raised the question myself a while ago and the advice I got was that for the 268M, flat darks wouldn't give me much of an advantage over bias, so I stuck with bias frames since they're so much easier. Here's the thread: https://www.cloudynights.com/topic/833261-dark-flats/

Gain for all calibration frames should match the lights.

How are you calibrating your flats? I only use only a master bias and I don't calibrate darks.

Cheers,
Scott

Hi Scott, I agree on the use of Flats. Dark Flats or Bias could be considered optional with this camera. 

Your point about the auto-stretch makes a lot of sense. Since there isn't as much noise it's going to stretch it further which would explain the whiteish color. 

By definition, BIAS frames are supposed to be as short as possible. It would be more accurate to say I'm in the middle between BIAS and Dark Flats by using an exposure time of 0.1. I can't say it's actually better, but it works well for me. 

Good question about why my luminance has a longer exposure for Flats. Previously I worked with Flat exposures as low as 0.1 - 0.2 seconds. Flats will still work with really short exposures, but by lengthening the exposure just a little, it seemed to capture a fuller characterization of the imaging train. I should mention, I use NINA's Flat Wizard with the default settings. 

Using different gain settings to correct calibration issues is last resort, technically I don't think gain should matter for Flats, assuming the same gain was used for the BIAS, but using different settings adds unnecessary complexity in my opinion. 

This is a diagram on how I calibrate Flats. I only use BIAS frames for Flat calibration. Same as you.