94 replies to this topic

[GoldSpider]

I've had a lot of problems with flats on my 294 along with a lot of other people on CN, and it sounds like this might explain what's happening.  It's encouraging that ZWO is at least getting close to acknowledging a problem with the sensor.  Staying tuned...

[cuivienor]

I'm not sure whether flat problems are related to the same issue - what I see is very specific to narrowband, and red narrowband at that.

I will say my issues with flats in broadband with the 294 stopped when I took them above unity gain (I was first taking them at gain zero). It's a capricious little thing, that camera, but it can deliver very good results in broadband, from red zones or better.

[andysea]

I think it may be affecting broadband imaging as well and it's most likely the culprit for the red background artifacts that we see. The camera is capable of producing good images, especially of busy nebular regions. However I have always had issues with background artifacts when shooting galaxies where the background is mostly sky. It also depends on how far one is willing to push the data in order to get the very faint stuff.

[jgraham]

Like some, I really haven't experienced a lot of problems with the 294 and flats, but as a rule I try not to push my data too hard. On the rare occasions where I end up with a stubborn gradient a synthetic flat does the trick. This applies equally well to my white light and narrowband Triad images. I actually seem to have fewer problems with the Triad as I tend to get better flats and have fewer problems with gradients caused by my Bortle 8 skies. Just lucky I guess.

[17.5Dob]

@Ballyhoo: to be fair, I'm a bit crazy I'm now using the 071MC Pro instead, and it's getting good results on the Crescent Nebula

 

@17.5Dob: yeah, the 294 is actually quite good for its intended purpose: OSC in low light pollution (red zone or better - yeah, red zone is low LP for me). Very fun to watch the subs come in so quickly. And very long stacking times, you'll want to invest in a multi-core CPU and install Linux + PI Linux for best performance But it has to be used at high gains (including for flats), and it can feel like you're tinkering with it from time to time. In that respect, the 071MC Pro is very stable, if less sensitive.

The 071 is my current choice ATT.

The Nikon D5100 was legendary for it's AP quality. Cooling it will help even more.....

[Zebenelgenubi]

 I have been following these discussions on ZWOASI294MC calibration problems for almost a year now and have been puzzled since I haven’t been seeing the artifacts users are describing in images with my camera.
I decided to do some experimentation prompted by the discussion on this forum concerning narrowband filters.  Here are flats that I collected using a drawing light panel.  I show two cases: the first is a master flat collected using the camera without a filter.  The second is master flat with an OPT Triad Multi-Narrowband filter.    I used SharpCap to collect and average the master flats and adjusted the light intensity so that the white light histogram was at approximately half scale in the ADC output in both cases.  Notice that the OPT Triad flat is deficit in red due to the very narrow Ha bandpass.  I also show highly stretched and color saturated images of the flats as well as images of the extracted red channels.  While I do not see a pronounced X, there are low level artifacts in the flats with the Triad that do not appear in the broadband flats.
Seems to me that the artifacts ought to calibrate out in when these flats are applied in the conventional way (I normalize my light frames with monochrome master flats).  I searched through some of my images to try to find these artifacts but since I have been imaging in Bortle 4 skies most of my images are limited by the LP residuals.  I did find one very deep Ha image of the Horsehead region using the red channel extracted from an OSC image.  You can see a dark crescent shape in the lower right of the image that I had not noticed until comparing it to the extremely stretched Triad flat.  The artifact appears to be real but at a very low level in my camera.  I may have just stumbled onto a good one.

 

 

 

Edited by Zebenelgenubi, 14 June 2019 - 10:19 PM.

[cuivienor]

Wow thanks for that Zebenelgenubi!

The original Triad would be less affected, since green blue channels tend to hide the issue, while the newer Triad Ultra would show the problem more.

I think mine is a particularly bad sample (the thing is visible even with a very small stretch), most other 294 should be much less affected. From a flat that Sam sent me of another 294, a pattern was there but much less pronounced, and could likely be calibrated out. Mine is really, really bad (I'm getting an exchange to a refurbished 183MM Pro in the end, at my request. Will go proper narrowband. ZWO really stand behind their clients).

[AhBok]

Too bad about the 294. I really liked mine and wanted it to work so much. Mine was one of the bad ones. I worked with ZWO about 3 mos trying to find some combination of gain, temps, flats ADU, etc., but could not calibrate the purple and green “swoosh” out. I took some pretty pics of brighter objects , but when others commented on my beautiful black backgrounds, I knew it was because I was clipping data with my black point. I was patient working with ZWO, because I really just wanted a solution. In the end, ZWO and I agreed the camera would never be able to perform to spec and they (and OPT) were extremely supportive in letting me return the camera under warranty. I then purchased the ASI071MC Pro and could not be more pleased.

If, however, we could have solved the problem with the 294, I definitely would have kept it. It’s sensitivity is remarkable. I’ve never seen a camera that could work as well with high gain and short exposures. It’s definitely a wonderful EAA camera.

[Zebenelgenubi]

Wow thanks for that Zebenelgenubi!

The original Triad would be less affected, since green blue channels tend to hide the issue, while the newer Triad Ultra would show the problem more.

I think mine is a particularly bad sample (the thing is visible even with a very small stretch), most other 294 should be much less affected. From a flat that Sam sent me of another 294, a pattern was there but much less pronounced, and could likely be calibrated out. Mine is really, really bad (I'm getting an exchange to a refurbished 183MM Pro in the end, at my request. Will go proper narrowband. ZWO really stand behind their clients).

I have convinced myself that the artifacts that I am seeing in the Triad Narrowband images with the ASI294MC camera are due to a variation in the response of the Bayer filters over the sensor array.  The variation effects the Ha response of the Bayer matrix is the most obvious since 656 nm response is on the shoulder of the red Bayer filter. Minor shifts in the filter response due to process variations affect the filter gain on the shoulder.  Turns out, in my camera, the effect is negligible for broadband imaging and subtle for narrowband imaging.  I love this camera and am working on narrowband flat processing the would mitigate or eliminate the artifacts for imaging.

 

[cuivienor]

For information, I was contacted on FB to check the flats of another ASI294MC Pro, used with the OPT Triad Ultra filter. The pattern is there, much more subdued in the center, and very marked at the corners. Posting here with the permission of the camera owner.

Attached Thumbnails

• 

[Diellur]

Just read up on this thread, having started another one today to get some input on whether I go for the 294MC-Pro or the 071MC-Pro. As I may decide to shoot NB with an OSC in the future, this more or less makes my mind up and swings me towards the 071MC-Pro.

[cuivienor]

@Diellur: yes, if you intend to try NB with one of the multi-band filters that have a narrow bandpass in the red spectrum (e.g. OPT Triad and OPT Triad Ultra), the 071MC Pro is a better choice. The 294MC Pro is such a fun camera in broadband though, with its insane sensitivity.

[McErono]

I don't know if this is interesting or helping anything but I thought I show you my result with the 294mc pro and an optolong l-eNhance Ha, Hb and OIII NB filter. 120 unity gain masterflat debayered boosted autostretch in pixinsight. Camera was bought in Summer 18 directly on ZWOs website.

 

 

 

same stretch no filter

 

Edited by McErono, 07 September 2019 - 01:35 PM.

[Swordfishy]

 

I don't know if this is interesting or helping anything but I thought I show you my result with the 294mc pro and an optolong l-eNhance Ha, Hb and OIII NB filter. 120 unity gain masterflat debayered boosted autostretch in pixinsight. Camera was bought in Summer 18 directly on ZWOs website.

 

 

 

 

 

same stretch no filter

 

 

 

Very interesting observation. Now you have prompted me to try without.

[niccoc1603]

I did some tests with my ASI294MC PRO and Optolong L-Enhance
6s flats
 
RGB Master flat:
 

 
Blue channel:
 

 
Red channel
 


You u can notice a faint pattern in the Red left corner. If you take a closer look at the RGB you can recognize the same pattern, although very faint



It is barely noticeable so I am not too worried about it for now, and being on the side of the frame it will likely be cropped anyway. I will report back when I integrate and calibrate the lights

For comparison, this is the R channel from a Masterflat without filter

Edited by niccoc1603, 19 September 2019 - 05:11 AM.

[tedswanson]

It's been a few months since anyone has posted on this thread, but I'm having the exact same problem with the 294 and the ZWO Duo-Band filter.  My images have a pretty dramatic color gradient and I'm wondering if anyone came up with a solution (increased gain?) to solve the problem.  The gradient appears in lights and flats pretty equally.  Interestingly, if you rotate the filter the gradient moves with it.

[BigBanger]

Hi folks,

 

I’m joining this thread late in the game, but am doing so because I too seem to have a problem with ASI294MC-Pro camera. I did however do some interesting testing today that may (or may not) help shed some light on this issue.

 

I created indoor, un-cooled t-shirt-white-wall flats with 1) no filter (but with the filter drawer in place), 2) Optolong's 2” L-Pro and 3) Optolong's 2” L-eNhance filters and here are the results. (These single frame flat examples were de-Bayered, color-calibrated and stretched to better show the uneven backgrounds.)

 

The images were taken through a TS-Optics 115mm ED triplet with a TS-Optics flattener attached.

 

It looks like the uneven background has to do with the L-eNhance filter or some interaction between that filter and the ASI294MC-Pro camera. SO IT'S NOT JUST THE OPT TRIAD FILTER! (sorry for shouting) It seems to be a "generic" problem with this and some other CMOS cameras (I also had this problem with a previously owned Opt Triad and my Meade 70mm Quad astrograph) and any tri-band filter.

 

Comments would be appreciated.

 

Attached Thumbnails

• 

[ChrisWhite]

I cant understand why zwo is still selling this camera...

[Avenar]

Hi everyone,

 

I had a lot of problems with this camera, especially with the flat.
For now I can get good images only without calibrating with the flat frames: they make worse final elaboration, adding colored gradients, vignetting, also with the L-enhance or L-pro filters.

 

I would like to remind those guys who do flat at 25000 ADU for this camera and experience overcorretting, 294 has a 14bit sensor and its max dynamic range is 16364 ADU! A good flat will be done with about 8000 adu depending on the gain setting. I tried also with flat darks but no no significant improvement. Short or long exposures also not working for me... I have a newton 200/1000 and I'm using a led flatbox selfmade with a lot of layers inside to reduce brightness.

 

For example, this image was taken under an urban sky last week with 294 and l-enhance filter, only corrected with dark and pixhinsight. Not bad for a heavily lights polluted sky.

 

 

 

 

Attached Thumbnails

• 

[bulrichl]

I would like to remind those guys who do flat at 25000 ADU for this camera and experience overcorretting, 294 has a 14bit sensor and its max dynamic range is 16364 ADU! A good flat will be done with about 8000 adu depending on the gain setting.

Sorry, this advice is totally wrong. The sensor IMX294 has indeed an ADC with 14 bits, but the camera driver scales the intensity values (multiplies by factor 4). So the FITS files of the ASI294 contain intensity values in the range of 0 - 65535 (= 2^16 - 1), and not 0 - 16383 (= 2^14 - 1). That means: correctly exposed flat frames of the ASi294 shall have the peak in the histogram around 32000 ADU. When the MasterFlat overcorrects the vignetting, the cause is probably wrong calibration of the flat frames. Overcorrection will result, when a too high value was subtracted from the flat frames.

 

Bernd

[Avenar]

Hi Bernd,

 

First of all thanks for your advice, but I have I tried several exposures from 25000 to 5000 and the vignetting seems lower with lower values. At the monent in my setups, flats always make the images worse instead of "cleaning" them.

 

The calibration is done with flat darks, without bias with batch processing. I also tried doing it manually but nothing. I believe that the issue isn't during elaboration but of another nature... Wrong flatpanel? Internal reflections?

 

I have read so many threads about cmos sensors with flat problem but no one seems to find a definitive solution!

[bulrichl]

I don't claim that I mastered my ASI294 perfectly. In fact, I had severe problems with the "colored background artifact" (see the lengthy thread https://www.cloudyni...-asi294-mc-pro/ , examples of deep integrations are in posts #105 and #112. The images show STF auto stretched integration results without any further processing). I could not manage these issues appropriately and finally threw in the towel and purchased an ASI071 which I now use exclusively. However, I never had issues with over-/undercorrection of vignetting and dust motes with the ASI294. The most common cause of over-/undercorrection of vignetting and dust motes is wrong calibration of the flat frames.

 

I didn't use those special filters (L-enhance or L-pro) though.

 

Bernd

 

[Avenar]

Hi Bernd, 

I read part of the post,and your images are pretty close to mine, with same color artifacts.
I am sorry that you have changed camera, but I understand that it is not easy to solve:
in fact I haven't found an effective way out     
What I ask myself is: I see very nice photos taken with the 294, so, how do they get them?   

[Biggen]

Sorry, this advice is totally wrong. The sensor IMX294 has indeed an ADC with 14 bits, but the camera driver scales the intensity values (multiplies by factor 4). So the FITS files of the ASI294 contain intensity values in the range of 0 - 65535 (= 2^16 - 1), and not 0 - 16383 (= 2^14 - 1). That means: correctly exposed flat frames of the ASi294 shall have the peak in the histogram around 32000 ADU. When the MasterFlat overcorrects the vignetting, the cause is probably wrong calibration of the flat frames. Overcorrection will result, when a too high value was subtracted from the flat frames.

 

Bernd

I've been using a 25000ADU setting when creating flats with my ASI294MC Pro.  Wonder if I should up it a bit?

Edited by Biggen, 22 January 2020 - 09:20 AM.

[bulrichl]

I've been using a 25000ADU setting when creating flats with my ASI294MC Pro.  Wonder if I should up it a bit?

That's not the topic of this thread, so my apologies to the OP.

 

The following applies to an OSC camera (as this thread is for the ASI294):

 

It depends on the light source for the flat frames. If the peaks of the color channels decently coincide with one another, there will be no need to make new flat frames. However it is not uncommon to have one (or two) color channels that are considerable weaker than the others. Therefore it is worthwhile taking a closer look, namely at the histogram. Appended is a histogram of a MasterFlat (CFA) that was taken with an ASI071 and a LED flatfield box. It shows two separated peaks, the left one for red (CFA0) and blue (CFA3) channels and the right one for the green (CFA1 and CFA2) channels. Here are the statistics, also of the split channels:

MasterFlat_unity_0_0055s_10   CFA0         CFA1          CFA2          CFA3
count (%)   100.000           100.000      100.000       100.000       100.000
count (px)  16236096          4059024      4059024       4059024       4059024
mean        30636.359         22042.943    38595.879     38511.426     23395.186   <==
median      34149.705         21968.214    38465.472     38376.895     23287.535
minimum     19691.551         19691.551    34218.626     34149.705     20622.178
maximum     42730.629         24409.932    42730.629     42647.029     25913.595   <==

I deliberately exposed somewhat longer in order to get sufficient signal in the red and blue channel as well. Clipping (caused by a too long exposure) must definitely be avoided. That means there must be no pixels with an intensity of greater or equal about 80 % (about 52000 ADU) in the uncalibrated flat frames.

 
Bernd

Attached Thumbnails

•