ASI585 strange banding in blue chanel

Calipso232

I recently noticed that my ASI 585 camera has strange banding in the blue channel. Below is an example. Has anyone encountered this before and is there a solution? (Exposure time = 180s, camera temperature approx. 4 degrees C).

w7ay

Calipso232

The dark-light almost horizontal bands of the first image could be the "amp-glow" that you read about. It eminates from a warm spot on the sensor.

Amp glow is usually removable by taking dark frames at the same sensor temperature, and at the same expsoure time and camera gain. I don't believe your camera has a cooled sensor, so using an arbitrary dark frame (taken at a different temperature) can make the problem worse instead of helping.

If you cannot switch to a different camera, you can try taking dark frames right after taking your light frame, so that the temperatures are somewhat similar.

The color gradient in the second image is quite likely (with the light pollution nowadays) to be from the pollution gradient in the sky. When you took that image, the left side of the image was not closer to the horizon, was it? Or in a direction that is closer to city lights.

The right side is nice and dark and could be away from city lights, or closer to the Zenith, or wait for M45 to move to that darker part of the sky to look for nebulosity. If the sky background is brighter than the nebulosity, it is close to impossible to bring out the nebulosity (SNR-wise).

Better light pollution filters can help a little, but if you have uneven pollution, there is not much you can do about it except to move to a different location. And if the filter cuts too much blue region, the M45 nebulosity will be cut too. You just have to wait until the object has moved to a different part of the sky (like up closer to Zenith) that is not polluted, or has less gradient.

Otherwise, there are tools that can be quite effective in removing sky gradients, like on AstroPixelProcessor:

https://www.astropixelprocessor.com/community/main-forum/gradient-and-noise/

I don't know what error the third image is showing. I simply see some clumps of uneven illumination. Could be attributable to a poor sensor that is not meant for long exposures. I also see an almost-horizontal band in the nebulosity below the horse's head. I just looked at an image that I took some time back, and that band is not present, so it is something from the camera:

Did you check to see if is not some finger oil smear across a part of the long edge of the sensor? Dark frames will not be able to remove a defect like that.

Perhaps check what an ad-hoc flat frame shows (i.e., just a simple unsaturated daytime image with a T-shirt). If it is not a 100% opaque obstruction, a good set of flat frames (taken at any temperature and exposure value) will mostly fix the problem, albeit with a loss of SNR in those smeared regions.

Chen

Calipso232

Hi Chen

Thank you for your answer. The gradient in the photo is not a problem, it's just an unprocessed full moon photo. I'm more concerned with the slightly diagonal stripes visible in the photos. These bands only occur in the blue channel, the red and green channels do not have this problem, which is unlikely to be due to amp glow, since amp glow manifests itself on all channels. The photos I posted were already stacked, using 20 darks and 20 biases, taken just after the lights were made, with a temperature error of no more than +/- 0.5 degrees C. In single frames, this effect is also visible. Since this effect occurs only in the blue channel in the RGB frame, it causes color casts visible in the second photo. I also cleaned the sensor and the AR window, but it had no effect. Below I also mark what I mean, because in the first two photos it is actually poorly visible (this area is in the same place of the sensor, the photos are simply rotated 180 degrees in relation to each other).

w7ay

Calipso232 I'm more concerned with the slightly diagonal stripes visible in the photos.

That is why I suggested taking some unsaturated "flat" images, if those defects do not show up in your dark frames.

If those also show the streaks that are parallel to the long edge of the sensor, it could be something in the sensor, or on the Bayer layer right above the sensor.

And if you see them on the flats but not darks, try inspecting the sensor really carefully to see if there is any swaths of oil or something acrosss the long edge of the sensor. A set of flats should be able to correct the problem too.

If you don't see them on either the flat frames or on the dark frames, there could be something wrong with that camera when used with 180 second exposures. Perhaps that camera is not intended for long exposures. The stuff you circled are definitely not up on the sky.

One other thing -- do you have a good IR-Cut filter in the imaging train, and close to the sensor? That camera has a significant near-IR response in the blue channel. Though right offhand, I can't think of an IR leak mechanism that will create a pretty high resolution streak that is fixed relative to the sensor.

Chen

Calipso232

I will try to make a set of flats and analyze the darks in my free time and let you know the results. As I wrote, I cleaned the sensor and the AR window and did not see any dirt on them. As for the filter, I use the ZWO IR/UV cut filter, but I've also made frames without it and the effect is the same.

Tech@ZWO

Calipso232 You also can rotate the camera to see if these stripes change positions.
If they change, which means they are on the telescope or other optical adapters. if they are not, which means they are on the camera.

Calipso232

I managed to make flat frames and I'm posting one example, successively red, green, blue and RGB channels. I don't see this problem on dark frames.

w7ay

Calipso232 I don't see this problem on dark frames.

Dark frames should never give a gradient unless there is something terribly wrong with the sensor.

However, they can appear in flats (not unusual, actually) if your illumination is not perfectly non-directional, or non even.

So, the flat (blue component) does show the almost horizontal stripes! I took a part of your blue frame, and increased the contrast:

Do you have a second camera that you can use to eliminate the rest of the optical train as the source of the problem with 100% certainly? There is defilitely a problem somewhere between the flat panel and the sensor.

But... since the stipes had earlier rotated with the camera angle, I suspect the problem is on the surface of the sensor, or a defect in the sensor itself.

Your dealer should be willing to replace the camera if you send them the above flat, and another one with the camera rotated by 90º.

It should show up even more cleanly by stacking a bunch of flats to remove the camera noise.

However, the above also means that a good set of flats will remove the problem from your images.

Side note:

I actually use the Daystar FlatCap with my refractors when I take critical flats, since they further diffuse the directionality of light another by 4 degrees, on top of a Gerd Neumann flat illuminator that is placed right at the objective so that the illumination itself is way out of focus.

In my experience, tablets and tracing backlights are terrible for taking flats since they are usually not even, plus the red, green and blue LEDs are too narrow-banded to illuminate a narrow band filter (a good H-α filter will show very little response with an LED backlight). The non-LED (like EL panels) backlights are better since the spectrum is more broadbanded. A incandescent lamp is actually really good if you can diffuse it first, since those produce really wideband Planck blackbody spectrum. A 200 watt incandescent bulbs will have more blue component than a 60 watt bulb.

Chen

Calipso232

Thank you Chen for your suggestions. It occurred to me that I could make flats on the camera itself, taken out of the optical path, and then it should eliminate the problem with the optical path, although it seems to me that the problem in the optical path should manifest itself in all channels. I ordered the ASI 533MC Pro camera because I thought it would be a better option for DSO than the typical planetary ASI 585 MC camera. Then I can definitely eliminate the problem with the optical path, these cameras have a very similar sensor size. It will probably all end up with contacting my local dealer and having the camera repaired or replaced.

w7ay

Calipso232 It occurred to me that I could make flats on the camera itself, taken out of the optical path

No. Those won't be flats then. You can only do that with darks.

Flat frames depend on having the same geometry (same vignetting, same focus position, etc).

Dark frames depend only on the camera.

Chen

Calipso232

I just expressed myself wrong. I meant to make flats on the camera itself, taken out of the optical path, to determine where the problem is, in the camera itself or in the optical path. I made just such a flat frame (photo below) and it clearly indicates a problem with the camera itself.

w7ay

Calipso232 I made just such a flat frame (photo below) and it clearly indicates a problem with the camera itself.

If you have removed that camera and allowed light to enter the sensor, and you see patterns on the exposed frames, the error has to be right at surface of the sensor or inside the sensor. Even the glass window that is a couple of millimeters above the sensor surface would not be able to show banding structures.

Have you followed what I sugeested earlier to look for smears on the surface of the sensor? Just use a loupe --everyone who deal with filters and sensors should own one, and nowadays there are cheap loupes that come with their own rechargable illumination; I recently couldn't resist picking up this illuminated loupe:

https://www.amazon.com/gp/product/B01LWRRLV3/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1

There are many at Amazon and other places. I only wished these illuminated loupes also have UV lamps, since you can pick up smaller stuctures more easily. I still keep some UV illumination around -- but be really careful when you use UV lamps. Read all the cautionary notes.

There are also cheap "USB microscopes."

Since the problem is only present in the blue component, it sounds supiciously like a defect is in the blue micro filters of the Bayer layer.

If it is not on the surface of the sensor, it would be in the sensor itself (but that should be more invariant to color component). The problem may also not be covered by warranty since this camera is not meant for long exposures. A friendly dealer may still be willing to send you a replacement to test if it is any better.

Chen

Calipso232

Thanks again to Chen for your help and suggestions. I think that I will probably buy this magnifying glass, the more so that even with shipping to Poland, the price is attractive. As for the camera, I already sent it back to my dealer under warranty. I'll see what his decision is. If the problem was on the sensor itself, from what I noticed, there is no direct access to it. There is a protective glass in front of the sensor itself, in addition to the protective glass in the camera housing. I don't know why but staking with flats in DSS didn't work. As I wrote earlier, I will wait for the seller's decision and then I will think what to do next. Yesterday I received the ASI533MC Pro camera and it will be my DSO camera. So far I don't have the weather to test it. The ASI585MC will only be used for planetary imaging, which is what it was designed for.