1.51
#...
·
·
2
likes
|
|---|
|
Thanks Chris for writing this up. There has been a lot of confusion about interpreting tilt in systems and I'm hoping the new features in Hocus Focus will help. I see people asking about how much to shim - and that is one of the benefits of the Hocus Focus routine because it determines the actual geometric curvature and tilt of the 3D surface of best focus in terms of radius (mm) and degrees, respectively. Plus, the orientation of the tilt is given in degrees. This does require that you enter the precise steps per micron of the focuser - but if you do it right that will tell you the angle of tilt you need and how it is oriented. There will be trial and error involved in putting it on the right "side" so it cancels rather than adds to the tilt - but the amount of tilt is measured and you should be able to size the shim as needed. But some of the features haven't been tested much, so there may be refinements needed. The main points I have made about tilt and curvature are 1) You cannot tell them from a single image - you need to study how the stars behave as you go through focus. 2) Sensor tilt does not cause stars to go oblong in corners - but it may enhance the impact of aberrations that are present. 3) Most existing tools cannot distinguish tilt or curvature from other aberrations, so they may do more harm than good in terms of providing insight. Finally - tilt and curvature are geometric properties independent of the sensor size, so they should be expressed as a radius of curvature in mm and as a tilt angle in degrees - rather than percentages of something that don't give insight into, for example, how much to shim. Some time ago I did a simple example of a system with a large amount of tilt to convey that tilt only shows as defocus across the sensor - but to my surprise many people expressed doubt that the example I provided was representative. I hope that the general understanding of tilt has increased so that example now makes sense. If it doesn't make sense to anyone then it would be good to clarify any remaining confusion: https://www.cloudynights.com/topic/804811-what-tilt-in-a-sensor-should-look-like-theory-and-example/ For background, my description of the new Hocus Focus features provided by George Hilios in NINA is here: https://www.cloudynights.com/topic/821151-direct-measurements-of-image-surface-tilt-and-curvature/ Frank |
10.80
#...
·
·
1
like
|
|---|
Steeve Body:Steeve Body: You know what I can't for the life of me find the frames I was comparing to get that 20% difference... but you are most certainly right that this may have been a focus difference... Here is a google drive folder with some data. I have included 2 targets shot during the same session after my attempt at fixing the tilt with a shim. The targets are: the Cat and Paw and the Helix. Helix was shot at rotator angle zero The Cat paw was about 90 degrees to the right compare to the helix. Focus was done at the filter change.. no offsets were used The Helix session has a meridian flip right in the middle of the NB session, the RGB capture (15s exp just for stars) does not have a flip... not sure how useful that is but I've included both with the best RGB frames of the night to minimise seeing variation. The Cat paw was capture only on one side of the meridian for that session.. however I have included 2 comparable frames 15s exp on the green filter before and after the flip (from another session the day before). Apart a slight focus difference between the 2 frame there isn't much difference I think I'm not too sure if you can extrapolate anything out of that data... but hopefully you can! I want to go by elimination process to get my frames as tight as possible. So would you say the order would be 1) Check for focuser sag first 2) Measure tilt variation moving the rotator by 90 degree increment. If tilt moves with rotation and tilt amount does not change much then we can assume the tilt is mostly coming from before the rotator... if not then tilt is mostly coming from after the rotator. If tilt does not change position, the sensor in the camera housing is tilted 3) Install TS optics tilter 4) Move to star region near zenith measure and fix tilt 5) Measure tilt at different camera angle 6) check at different position on both side of the meridian Again thanks so much for your advise and expertise on this! So far I have not had much time with this scope... but it is already looking pretty sharp. My first light with this scope  M16 The Eagle Nebula |
20.43
#...
·
·
1
like
|
|---|
|
Frank, Single frame analysis can certainly be used as a guide. I've demonstrated this in the past, with data over at cloudy nights. You just need to know what you are doing and how to avoid potential pitfalls, which I touched on in a previous post in this thread. Understanding how to use a tool is a key element to being successful with it. Please, lets not let this conversation here devolve. Everyone is in agreement that the focus bracket model is more accurate and easier to use, but as they say, there is more than one way to skin a cat. |
9.03
#...
·
·
1
like
|
|---|
|
Hey Chris, I cannot read through this whole post because it gives me PTSD symptoms due to flashbacks to my flattening of my Stellarvue/Nightcrawler/ASI 6200M system. However, NINA also saved me:  M3 Field Flattening Exercise5: How to micro-manage the ASI 6200 Tilt Plate: NINA Tilt Analysis Now, here were my big take aways after months of work: Realize on the ASI the tilt plate adjustments in and out also count for the backfocus.. you have 1-2mm of adjustment in the plate that must be taken into account. Realize that you must tighten and loosen the ASI tilt plate set screws and the adjustment must be a hair's width. When you move one side in, you have to let the other sides move out... a very delicate dance... I am not going to touch those screws again if I can avoid it, but I doubt these adjustments will last over time and need to be repeated....  JY |
22.65
#...
·
·
2
likes
|
|---|
Frank, Chris, Thanks for that comment. I think that it’s relevant to point out a couple of things that I hope won’t cause the conversation to devolve. It is indeed quite possible to measure sensor tilt using a single frame—to a very high degree of accuracy. The Nina solution looks quite good and your results look outstanding but it’s possible to get similar (and possibly even more) accuracy in the tilt measurement using SkyWave. SkyWave uses a neural network to analyze star-test data to retrieve the wavefront. Unlike the Roddier method (available in PI and used on Hubble), the NN method always converges and it requires only a single image. SkyWave can measure tilt across the field to within a few milliwaves (at each position) using a single frame. Gaston and I presented a paper at SPIE (in 2020) that demonstrated 10 milliwave agreement between SkyWave and PhaseCam interferometer wavefront measurements. Gaston has subsequently shown agreement with a Shack Hartmann sensor under the real sky to within a few milliwaves. The biggest weakness of using through-focus V-curve fitting is that the sensitivity to focus goes to zero at focus, which means that small measurement errors due to seeing fluctuations may translate into significant errors in the curve-fit that determines the focus position. The AI solution used in SkyWave is also sensitive to seeing fluctuations; but in that case, taking multiple frames serves to signal average the focus term itself to reduce uncertainty in the computed focus position. Clearly, in either case, it’s important to make the measurement under the best possible seeing conditions. John PS. I’d be happy to discuss the single frame method in more detail but I don’t want to derail the topic that Chris started so that should happen in another thread. My key points are that: A) single frame methods exist and B) seeing is important. |
11.02
#...
·
·
1
like
|
|---|
| SkyWave is outstanding! While I didn't use it for tilt measurements, I did use it to collimate my 12.5 iDK and the accuracy and ease of use was incredible. If there's a tilt measurement module in the final production version (I was using the beta version of SkyWave at the time) then I'll need to look into investing in that. |
0.00
#...
·
·
2
likes
|
|---|
|
Great Stuff. Keep up the good work. Clear skies!! Steve |
|
20.43
#...
·
·
2
likes
|
|---|
John Hayes:Frank, John, I welcome a discussion on skywave. This topic is open to any objective method for dealing with tilt and/or spacing. The only thing I'd like to avoid is theoretical discourse without data and results to back it up. Those kinds of discussions tend to get into the weeds. Please, feel free to present your experience with skywave. It sounds fascinating. -Chris |
12.64
#...
·
·
1
like
|
|---|
|
Chris Wonderful discussion and information on a DREADED topic, indeed! Two points I would like to have further discussion or comment on, assuming there is anything else to add to them: 1- Why all of a sudden, a sensor title issue at all? 2- The available tools to fix it. Why all of a sudden, a sensor tilt issue at all? It may just be me, but I don't recall this being a big issue, or much of an issue at all, with older cameras like the StarlightXpress cameras I used to use. Is it due to the manufacturers rushing production and having poor QC in place? Is it due to the nature of the larger sensors now being used in these astro cameras? Of course, I am assuming strictly sensor tilt due to its placement in the camera casing. If this is strictly a manufacturer being lazy and not having proper QC steps in place for the prices we pay for these cameras, I think that is a disgrace. The available tools to fix it. There is a rich discussion here on the ways to detect the tilt (I am an avid NINA user and Hocus Focus user) which is fantastic, but I think the topic peters out when the discussion leads to the fixes. You mentioned the ZWO tilt adapter (which I agree is weak in design and high on frustration), the Gerd CTU (which seems really good) and the Octopi (which I was unaware of and seems silly expensive). Are there any other devices on the horizon you are aware of that perhaps make the adjustments less painful? I am guessing not, or you would have mentioned them, but I thought I would ask anyway. If not, I am tempted to try the Gerd CTU on my 80mm with the ASI2600. Fortunately, my ASI6200 on the C11 has little to no tilt, so I dare not touch it up in my observatory. Regards, Bruce |
|
11.02
#...
·
·
2
likes
|
|---|
Bruce Donzanti: In some cases, it's QC issues and in others it's the nature of the beast when working with very large sensors with small pixels. Gone are the days of 5.4-7.5-9 micron pixel cameras being the norm. With these 36x24 chips with 3.76 micron pixels,any deviation, whether caused by the camera itself or not, can impact image quality. On fast systems, with incredibly small CFZ's the tolerance the manufacturer can produce to may be too large. Similarly on f10 systems the CFZ is so wide that the tilt would need to be very pronounced to see any real impact from it. There are a number of camera tilt solutions on the market to choose from. It's really up to the budget of the user and the application they have. For an Epsilon, for example, you have 56.2mm to work with, which is asking a lot when a tilt unit like the Gerd Neumann CTU needs 17.3mm minimum. The Octopi is expensive, no doubt, but consumes very little space and allows tilt and spacing adjustments in concert which is really useful for eliminating the need to constant remove parts to add or remove shims. |
0.90
#...
·
·
1
like
|
|---|
|
My 2600MC on the RASA 8, according to ASTAP, went back from moderate to extreme. Actually, it had improved to severe after a 5mm spacer was installed. This should have got my back focus to very close to specified. I would have expected a dramatic improvement, not to still be parked at "extreme" in my latest subs. I don't see what use all of the graphs and explanations do unless one has the premium octopi to make adjustments. Short of that, I am thinking of trying to move the back focus in 1mm increments, and see if the ASTAP tilt improves. If it improves, I can add a 2nd 1mm. If it gets worse, I can try to subtract 1mm. ? Im so tired of extreme tilt, I might consider selling the RASA and going to an F3 Newtonian. |
|
11.02
#...
·
·
1
like
|
|---|
Paul Wilson: At f2, 1mm is a massive amount of change. Moving in 0.1mm increments with metal shims would be far better. I have the same challenge with the Epsilon, albeit less so since my e160-ED is f3.3. much more forgiving. |
|
20.43
#...
·
|
|---|
Seymore Stars: Thanks Steve! Good to see you! |
|
20.43
#...
·
·
2
likes
|
|---|
Bruce Donzanti: 1) I think it's really a combination of factors. Yes, there does seem to be some QC problems with some companies. Not just with sensor orthogonality, but also the adapters that they sell. Tilt plates, bolt on spacers, OAG's, Filter Wheels, etc... I bought a micrometer about six months ago and started measuring everything. I was shocked to find that even simple things like 5mm thick plates would vary in thickness by 0.1mm. OAG's from both of the major CMOS camera companies have quite a bit of variability. What I really think it comes down to is that larger chips AND tiny pixels are being used. Even the large 16803 chip was fairly forgiving with 9um pixels. Now we are using 3.76um or smaller pixels. These tiny pixels show all the warts in an optical system. Corrected fields that are marketed at 44mm, barely can correct for APS-C with 3.76um pixels. Plus there are a lot more people into astro-imaging than there used to be, so theres more people having issues just by that logic. I used to have strong opinions about sensor tilt, but now I dont. I see just as much tilt potential in every accessory we use, and since the cure is the same whether the sensor is tilted or the OAG is tilted... it really doesnt matter. 2) This entire point of this thread is to give people a process by which to fix tilt. I think understanding tilt and being able to identify it (lets not forget about backspacing) is critical to mitigating it. You dont need any fancy tilt devices to get this job done. I used to use aluminum foil as a shim. It's wonderful. Its only 10 microns thick and you can fold it to create custom thickness wedges. You can insert it between bolted components or in between threaded connectors. You can use the same focus bracket process to fix your stuff with aluminum foil. Just run the routine, learn where your tilt is, and start shimming with aluminum foil. Rinse... repeat. There is also a Baader Tilter. Cheaper than some, but not as precise. Limited application as it is only available in m68. There are the astrodynium tilt shims, which for many people might do a pretty decent job of getting you close. They cost $23 and take up very little backspacing. There is a TS tilt device, that someone mentioned earlier in this thread. Additionally, you really dont need perfect stars. Getting decent stars is about 25% of the effort, and then the remaining 75% is for the incremental improvements to great stars. Personally, I want near perfect stars... but that doesnt need to be everyone's goal. |
|
20.43
#...
·
|
|---|
Paul Wilson: As Bill mentioned, 0.1mm increments are going to be needed for your testing. Give NINA a try, and follow it's advice on whether to increase or decrease spacing. Baader and Blue Fireball sell vrey nice machined shims. You can get a mixed set from 0.1 to 1.0mm in thickness in whatever thread size you use. They just slip in between your adapters. Agena usually has the best inventory. The graphs work just fine with whatever your tilt mitigation method is. Gerd, Octopi, Aluminum Foil, whatever... My recommendation is to fix your backspacing. If you can eliminate backspacing error, a lot of your tilt problems will probably look a lot better. Field curvature and tilt compound and make aberrations worse. Fix your spacing and you might see your tilt improve. |
|
11.02
#...
·
·
1
like
|
|---|
| Great point about the work to get further and further improvement. There is a point where a scope can produce excellent images yet might have some edge star issues. Those minor issues can lead to a lot of hours and dollars to sort out. One needs to decide just how perfect they want their field to be. |
|
9.03
#...
·
|
|---|
|
Hi, IMHO, 0.1 mm is a first approximation. Eventually I had to use automotive spark-plug gauges to get the fine tuning correct. This gets the space accurate and it is then set with the screws. The CFZ for my F 6.9 Stellarvue scope is about 100 microns or 0.1mm so you get in the ball park but with tilt the corners need to be more finely arranged with NINA's help. Note labeling of screws a throw back to my old Tak 180 ED days... CFZ 17 microns!  |
|
1.51
#...
·
|
|---|
Single frame analysis can certainly be used as a guide. I've demonstrated this in the past, with data over at cloudy nights. Well - I'm still concerned that if my example showing tilt still looks odd to you, then there is a disconnect between how you think tilt works, and how it actually does show up in images. If you still have doubts about it then it really should be clarified because my example shows tilt in a very pure form that everyone should recognize. The problem with using different approaches and achieving a final good result is that it doesn't really indicate the tools used actually helped, because as long as you have a clear view of stars across the sensor, you can just make somewhat random adjustments until ultimately it looks good. A key flaw of single image tilt/curvature tools is they can give an impression of what adjustments are needed - but without distinguishing defocus from arbitrary aberrations they can be misleading and cause adjustments that circle around the desired outcome. So I agree that different tools can reach a good result eventually - and that understanding how it all works is important - but at the same time I think that all the time you have spent working on tilt could have been greatly reduced with tools recently made available. There is no need for guesswork when everything can be measured deterministically in a single series of exposures through focus. Frank |
|
20.43
#...
·
·
2
likes
|
|---|
Single frame analysis can certainly be used as a guide. I've demonstrated this in the past, with data over at cloudy nights. I'm sorry frank, despite your condescending tone, I'm not going to bite. I've backed up everything I have written about here and on CN with data, both analysis and stars. To date, despite my asking you over and over and over to provide real world data for your own contributions you have refused to do so. The contentious thread you linked to on CN is not going to cut it. If you would like to continue down this path of discussion, I will politely ask you to start a separate thread. Thank you in advance for understanding. To everyone, I will apologize in advance. Frank and I have a history and I'm sorry if this thread gets derailed. At this time I will not engage any further with him, and kindly ask that others do not get into a debate with him about anything here in this thread. If you would like to debate with him I will ask you to show the same courtesy I have asked of him... please take it to a separate thread. Thank you very much, |
|
20.43
#...
·
·
1
like
|
|---|
| If anyone else would like more information on how to utilize single frame analysis, such as that provided by ASTAP, I am happy to explain in detail how to do this. It is absolutely possible, if you know what to look for and how to use it. There are certainly pitfalls, and I have already explained in a previous post how to avoid them. Since using NINA Hocus Focus is so easy and definitive (and free) I dont see much value in using an inferior tool such as single frame ASTAP, but I am happy to educate more on the topic if people are interested (and willing) to learn. |
|
11.02
#...
·
|
|---|
|
I am not an experienced NINA user, is there something folks should keep in mind about setting up the software to use the Hocus Focus plugin? Here is what my Plugin configuration shows:  Here is what the autofocus options show:  |
|
20.43
#...
·
·
1
like
|
|---|
I am not an experienced NINA user, is there something folks should keep in mind about setting up the software to use the Hocus Focus plugin? Bill, In the Hocus Focus settings disable "Validate HFR Improvement." For the purpose of tilt and spacing analysis, you really dont care if the focus run achieved a good ending focus point or not. It just wastes a little time. For step size, you will just want whatever value gives you about 6 to 8 HFR on the first exposure. Keep in mind that NINA is using HFR while ASTAP is using HFD, so the value I recommended earlier of 20 HFD was for ASTAP. 6 to 8 HFR for the bookends of your curve seems to work really well. Everything else looks good. |
2.11
#...
·
|
|---|
|
This video came out today on The Astro Imaging Channel. Hopefully some find it useful. I know watching this combined with Chris's guides makes the whole exercise clear to me. Just need a clear night now to test and adjust. Telescope Back Focus: The Ultimate Guide [TAIC Short] - YouTube |
2.41
#...
·
·
1
like
|
|---|
|
Great article... this might be helpful : https://www.asgastronomy.com/photon-cage/ It tilts, backfocus, filter slides. I am huge fan of NINA and Hocus Focus... |
|
10.80
#...
·
|
|---|
Josh Jones: This looks like an awesome unit… wonder how $$ |
