Reflections from a CDK 12.5

It’s a micro-lensing effect and the Moon or the scope isn’t culpable. AFAIK you can’t avoid it with a bright star in the field. Well, unless you mask the sensor but I guess this isn’t pratical…

Planewave recommended extending the baffle tube in some cases to block external light from the optical path. Worth investigating as well.

Like Andrea mentioned, it is a well-known phenomenon caused by the micro lenses on your sensor. Due to reflections between these lenses and the AR glass that covers the sensor, these patterns can occur.

Nothing you can do to prevent them, but there are ways to get rid of them. Be ready for some PixInsight tricks. I’ve written the process down that I used for processing my Cat’s Eye. It follows a technique originally described by Adam Block.

Just a few days ago, the MoteCorrector script was released. This may also do the trick. Obviously on the starless image.

My ZWO 294 MM PRO do the same, it´s micro-lensing. I knew that my camera produces this phenomenon but I thought that the 2600mm was exempt from it.

John Hayes · May 7, 2026, 03:33 PM

I agree that the repeating pattern is caused by reflections from the micro-lens array on the sensor; however, the strays are coming from inadequate AR coatings on both the micro-lens array AND one of the elements sitting above the sensor—usually the camber window. When you are stretching the data to look for the very faintest signal, even a tiny “mis-tune” in the AR coatings will cause this problem when you have a bright star in the field. Knowing the focal ratio of the telescope and the diameter of the strays, it’s a simple geometry problem to work backwards to figure out which surface is causing the strays.

Unfortunately, this is a very difficult problem to fix without throwing a lot of money at it. Down in Chile, both of my friends Adam Block and Mark McComskey have discovered similar stray light problems that vary in magnitude on their scopes. One night, we all imaged the same star using the same exposures to compare results. The same sort of pattern occurs in all three scopes but in my scope, it is all but invisible except with the most extreme stretch and even then it is hard to see. Mark sees the worst strays with the OIII filter and Adam sees moderately strong strays with all filters. As I recall, he took the problem back to QHY but I don’t know where that went and I don’t want to try to tell Adam’s story. The bottom line is that the only way to fix this is to fix the AR coatings, which isn’t easy. Garden variety AR coatings that operate with a reflectivity of ~ 1% are easy to produce; however, it’s a bit of a black art to get coating performance to R<=0.25% and this is an application where you want the best possible coatings. Once the coating is deposited, you can’t strip it and start over if it isn’t right. You have to start all over again. So, in practical terms, you are generally stuck with what you get.

There is one thing that I’ve thought about that might help, but I’ve never tried it. In my telescope, I use an ONAG guider, which forms a NIR filter. So, when we did our co-imaging experiment, I wondered if the ONAG might be the reason that I was seeing such weak strays. It could be that the strays are being caused by the reflection of NIR light where the AR coatings would have VERY poor performance. So, it might be worth it to try putting a filter with a cutoff at around 700nm in the system to see if it affects the strays. To be clear, this might be a wild goose chase but it would be a big deal if it actually works.

John

Hmm, the Chroma Lum filter cuts off hard right at 700mm. Any narrowband filter obviously cuts off a couple of nm outside of the target band, definitely below 700nm, and you mention a case where it shows up in Oiii. Wouldn’t adding another 700mm be redundant in that case? Or am I misunderstanding what you’re suggesting

John Hayes · May 7, 2026 at 11:04 PM

jego · May 7, 2026 at 09:25 PM

John Hayes · May 7, 2026, 03:33 PM

I agree that the repeating pattern is caused by reflections from the micro-lens array on the sensor; however, the strays are coming from inadequate AR coatings on both the micro-lens array AND one of the elements sitting above the sensor—usually the camber window. When you are stretching the data to look for the very faintest signal, even a tiny “mis-tune” in the AR coatings will cause this problem when you have a bright star in the field. Knowing the focal ratio of the telescope and the diameter of the strays, it’s a simple geometry problem to work backwards to figure out which surface is causing the strays.

Unfortunately, this is a very difficult problem to fix without throwing a lot of money at it. Down in Chile, both of my friends Adam Block and Mark McComskey have discovered similar stray light problems that vary in magnitude on their scopes. One night, we all imaged the same star using the same exposures to compare results. The same sort of pattern occurs in all three scopes but in my scope, it is all but invisible except with the most extreme stretch and even then it is hard to see. Mark sees the worst strays with the OIII filter and Adam sees moderately strong strays with all filters. As I recall, he took the problem back to QHY but I don’t know where that went and I don’t want to try to tell Adam’s story. The bottom line is that the only way to fix this is to fix the AR coatings, which isn’t easy. Garden variety AR coatings that operate with a reflectivity of ~ 1% are easy to produce; however, it’s a bit of a black art to get coating performance to R<=0.25% and this is an application where you want the best possible coatings. Once the coating is deposited, you can’t strip it and start over if it isn’t right. You have to start all over again. So, in practical terms, you are generally stuck with what you get.

There is one thing that I’ve thought about that might help, but I’ve never tried it. In my telescope, I use an ONAG guider, which forms a NIR filter. So, when we did our co-imaging experiment, I wondered if the ONAG might be the reason that I was seeing such weak strays. It could be that the strays are being caused by the reflection of NIR light where the AR coatings would have VERY poor performance. So, it might be worth it to try putting a filter with a cutoff at around 700nm in the system to see if it affects the strays. To be clear, this might be a wild goose chase but it would be a big deal if it actually works.

John

Hmm, the Chroma Lum filter cuts off hard right at 700mm. Any narrowband filter obviously cuts off a couple of nm outside of the target band, definitely below 700nm, and you mention a case where it shows up in Oiii. Wouldn’t adding another 700mm be redundant in that case? Or am I misunderstanding what you’re suggesting

Narrowband filters are very well controlled over a defined bandpass but I don’t know how far out they go. You are probably right that Chroma filters extend out past 700 nm but I’m not sure how far. As I said, this is long shot and unlikely to be the cause. If you are sure that the system doesn’t pass anything beyond 700 nm then I agree that adding another NIR filter wouldn’t do anything and it’s not worth pursuing. In that case, it’s just the coating quality that’s causing the problem.

John

I have seen the spectra charts enough times to know that my Chromas are flat as can be way out past the QE of the IMX455, and thought nah…but waitaminute, lets check optical density…

📷 IMG_3748.jpegWhats this! Dropping slightly short of 4.0, which is Not Good. You might be on to something…

Ha does not have this peculiar spike:

📷 IMG_3749.jpeg

Just remember that the filters have to be properly manufactured to work properly. So the question is: Is the problem related to poor AR performance…or could it be due to some leakage in the NIR?

John

Mike Mulcahy · May 8, 2026 at 03:41 PM

Here is the six image stack of the O3 filter taken in the same run as the Ha stack. Our next step is to confirm the Ha filter is inserted facing the correct way. Chroma filters are unmarked as far as direction and they must examined carefully to get it right. Stay tuned.📷 cdk 12.5 O3 filter.jpg

The orientation of the filter will make almost zero difference. Yes, there is a difference; however, it is extremely minor. That’s probably why Chroma doesn’t mark the orientation.

John

I think it will make a difference but will not solve the problem. My guess, chroma are 3mm thick, so you will observe the effects of the artifacts being slightly larger and slightly fainter, if you have the reflective side facing the sky. You can tell if you’re looking at the reflective side by looking for the inside edge of the filter when you look through it.

I think this is consistent with what I was suggesting. To be clear, I am not even attempting to correct John here, and if at any point I stumble in the right direction, it is by no means through any real expertise like he has, and corrections are welcome 😆

But, the reason I think it might help a bit, is because yes, energy remains the same, but the energy is spread over a larger area. Probably the closer the reflecting surface on the filter is to the sensor, the more severe the reflection artifact will be in the image. 3mm probably wont be a game changer, but better to have the filter as far away as possible and deal with bigger but fainter spots than smaller but stronger ones.

There are a couple of other things not in the diagram that I think are in play. The sensor window and the microlenses.

Suppose the filter is perfect in terms of blocking and is not leaking any NIR light. What I think will happen is that the incoming light will hit the sensor, and reflect back through the microlenses (little domes over the pixel photosites), up to the sensor window. The light that reflects directly out at 0 degrees will pass through the filter again, out the top of the scope.

But, because the microlenses are domes (this is to catch light that is heading towards the area between pixels, and angle it back into the pixel), it’s actually going to reverse that reflect some light OUT at all different angles.

When that returned light travels towards the filter, the steeper angles are going to blue shift outside of the narrowband, turning that filter into a mirror. Then it comes back down onto the sensor again.

The more I think about it, the less I feel like the IR leakage would have enough energy to cause the multiple artifacts. It would definitely contribute to halos. But intuition is that it would need a LOT of energy to penetrate that weak spot at 710nm and reflect back through the microlens, the filter window, and hit the sensor again.

If it’s purely the off-axis reflection caused by the microlenses, I think the only mitigations are:

Wider bandpass, to tolerate eviction of more of the reflected angles

Larger spacing, as mentioned above, to diffuse the effect over larger area

Actually, a second advantage with larger spacing could be more of the steeper reflection rays terminating on a wall, and not hitting the filter to bounce back at all (John’s example might not account for this? But it might be negligible)

A better AR coating on the sensor window. I don’t know much about those, but I think they target a specific wavelength and are akin to phase cancellation, but I’m really talking out of my ass here. I seem to remember looking into it at one point and concluding that it’s likely a dead end as far as a fix goes for the extremes involved with astro. I think I remember someone else posting that they tried removing the window entirely and seeing no difference, and if it was helping, one would expect the artifacts to get worse?

All this to say, my only slighty-above-layperson level of knowledge does not see a solution here. It’s very interesting that John’s ONAG doesn’t show it though.

I think with the right math you could actually calculate whether or not a potential NIR leak is contributing, based on both the spacing of the artifacts, and the relative difference in energy making it through the filter vs the adjacent signal. There could be something else about the ONAG that is mitigating it too.

jego · May 8, 2026 at 11:30 PM

Here is the six image stack of the O3 filter taken in the same run as the Ha stack. Our next step is to confirm the Ha filter is inserted facing the correct way.

Please show the result of turning the filter around. I predict that you won’t see any difference. The fact that you don’t see the problem with the O3 filter simply says that there is some sort of problem with the coatings on the Ha filter. Either the AR coating is a bit off or NIR light is leaking through.

John