ZWO ASI 294 MM / MC / MC-P / MM-P - strange matrix pattern issue (KNOWLEDGE BASE / EXPERT NOTE) tips and advices opinie

Shop online: ● ready for shipping  Shop in Warsaw (qty): 10+ ● ready for shipping  Shop in Chorzow (qty): ● not available

CHEAP DELIVERY TO EUROPEAN COUNTRIES CHECK IT OUT!

Recently, I received two ZWO ASI 294 MC cameras for testing – one cooled and one uncooled – and that’s how I will refer to them :)
 
In short: when used with narrowband (NB) filters, these cameras produce a strange background pattern, which can be calibrated out, but NB images from these cameras, especially of faint objects under dark skies, will generally not be usable. What’s more, the worse and less optimal the calibration attempts are, the worse the results will be.
 
Briefly (a more detailed write-up is in progress, but it will take time): the cameras clearly behave differently under high-light conditions (full spectrum) and in narrowband conditions. This suggests that the sensor design – perhaps “backlit” – or the filter structure is non-uniform when viewed… for example, quite close to the H-alpha band. The topic is widely discussed online, but rarely in a deeper, comprehensive way with an attempt at thorough analysis.
 
Taking advantage of easier access to the devices in question, I performed several tests. EDIT1: it is clear that the problem is not limited to color cameras – I was able to test the ASI294MM camera as well, both in bin1 and bin2 modes (hardware-supported in this camera). The resulting flat also contains a certain pattern (visible after histogram stretch). The problem does not depend on the CFA (color filter array)! These are pure fluctuations in sensor efficiency in the Ha band. Very interesting – I also wonder how they vary depending on the band! Probably only SONY knows the real answer to this question.
 
1) In the case of the cooled version, I did not observe any issues with the cooler:
- the sensor ran for 12 hours at -15°C
- no dew, frost, or grease leakage occurred
- power at nearly 100% maintained the set temperature within +/- 0.2°C
 
During that time I took about 30 dark frames, 1200 s exposure, gain 120 (lowest HCG, HCG vs LCG), offset 50. In the darks, amp glow is clearly visible in the cooled camera (294mc-p, amp glow), while in the uncooled version there is very high dark current (aka thermal noise) Hot Dark, which, if wrongly calibrated with a flat, will certainly cause a “mess” in the background.
 
2) In both cases, flats taken with the setup: large APO + LED panel + cameras produced “normal flats”, confirming online reports that bright broadband targets will be well captured.
 
3) At this point, I concluded that the problem is mostly on the user’s side, arising from acquisition mistakes, setup construction, calibration, and normalization errors.
 
4) However, trusting the popular online opinion stating that the effects appear in “certain” bands, I decided to repeat the flat experiment using a 7 nm H-alpha filter.
 
5) This test was surprising – the flat background contained a pattern I call the “butterfly”.
 
6) I do not have a database on this, but I ran several experiments:
- heated and cooled the camera four times, repeating series of a hundred flats (-15°C)
- then took another hundred flats at +25°C to see if the pattern would change significantly
- took a flat with the same setup but using the 174MM camera instead
 
7) Conclusions:
- cooled flats are statistically identical and calibrate each other with about 35 levels accuracy out of 16 bits (65536), also shown after stretch
- an uncooled flat differs from a cooled flat by about 100 levels – still within tolerance despite the difference (read the stats!), also shown after stretch
- the flat from the ZWO ASI 174 MM is simply “normal” :/ (apart from dust visible with a strong stretch)
 
It is clear that the pattern is not strongly temperature-dependent, but rather wavelength-dependent. For the uncooled camera, doing long exposures would be very challenging – it is a “planetary” camera, and using it for long exposures requires a proper understanding of calibration, not blindly repeating internet pseudo-theories. You would need good dark calibration because the thermal noise is significant. Flats are essential. Flat darks are essential. With the cooled version, things are easier since darks (to remove amp glow) can be taken comfortably during the day.
 
Comparing the ASI294MC-P (cooled/uncooled) to the Atik383 is strange (such comparisons have appeared) since a CMOS sensor and a cooled CCD are entirely different things with different calibration methods.
 
Endnote
I do not have extensive research or months of use behind me – I take flats usually when changing something in the setup or doing an important session. In my opinion, the best feedback comes from users who use the device naturally over many months.
 
If I were to help in such a situation, I would need a precise description of calibration, master integration, and the raw files themselves. The topic is undoubtedly interesting, the “background pattern” does exist, but in my tests it is calibratable. Obviously, I did not perform many tests since I only had the devices for a few days.
 
My concern
If the “butterfly” is strongly wavelength-dependent, calibration might not be sufficient when all wavelengths in the spectrum are equally represented in the flat but unevenly represented in the light frame. This is difficult to test – it requires time, different astrophotography sessions, and dark-sky trips. Even if such an effect exists, it would be hard to measure and could be hidden “within typical amateur astrophotography errors.” Furthermore, for narrowband this problem does not occur because the light spectrum is inherently very narrow and well represented in nature on the objects for which we use it, e.g., Optolong L-Extreme. Here, proper calibration (but not via a t-shirt) is enough.
 
It would be interesting to get support in such tests from a user who understands the discipline. This is certainly not a camera for easy use on very faint objects (it works well on bright ones), but it can be handled, and it works properly. Background artifacts are definitely a weak point, but they do not make it unusable – they simply show that calibration is not “a fairy tale about noise removal”!

Author (or: complaints to)

Przemek M.

Glossary

The above text is written for practitioners involved in astrophotography who understand the technical language. Every professional field develops its own hermetic jargon – astrophotographers are no different, much like physicists, doctors, or lawyers.
 
Below (in progress) will be an introduction to the Astrophotography–Polish glossary.