I've installed Crayford focuser on my HD 9.25" and never looked back. Before that, the mirror flop was causing the autofocus routine to fail on a regular basis. Great video, thanks for sharing.
I was having the same issues in NINA on my C11 until I changed my auto stretch and black point settings to the recommended values in the NINA documentation for “poor seeing” conditions. After making these changes, I am able to get very consistent focus curves.
VERY interesting. I had not heard that before and did not realize the stretch and black point played into the autofocus process. Thanks for watching, Matt!
Hi James, You are certainly correct about the SCT versus a refractor when it comes to autofocus. My C-11 very rarely will have all the stars somewhere close to the curve in NINA. However, on super clear nights, I have seen some pretty remarkable fit to the curve. I am glad to hear that you too have (had) issues with the Celestron focuser coming loose. I periodically need to tighten the set screws, but if I tighten them too much, the focuser will get stuck.
Hey Pat. That Celestron focuser, or more accurately its interface with the scope, is terrible. The best approach is to leave the mirror alone and move the camera, but those focusers are awfully expensive. I'll swirl around the drain a while and see what happens...
I am using a ZWO EAF on an old (circa '83) Orange tube C8 in NINA with great results, good fit, etc. Your scattered plots remind me of when I was first attempting the setup and had the backlash value in both the IN and OUT fields before I learned it was one or the other. I tried Hocus Focus but did not have a good experience with it so went back to stock. Good luck getting it going!
Hey Bruce, I wrote Hocus Focus. Would you mind taking pictures at various focuser positions as if you were focusing, and sharing them with me (jokogeo) in the Discord channel? This will help me improve star detection for high focal lengths.
Thanks Bruce! I'm glad you're getting good results with your setup. Don't change a thing! As I've been pondering this problem, I have been having evil thoughts of moving to a different focuser. I will try a very large backlash number next time I'm out.
Thanks for your good video's. I use a Celestron SCT 11 with the same focuser as you do. By using a very large backlash compensation (step-size 1500) and making sure the last movement of the primary mirror is "out" (against gravity) I have good and consistent focus results in SGP. I use this setup for over a year with no problems. I remember I had to reinstall the focus motor a few times before it ran smooth without heavy points during a revolution. So yes, all the mirror flop in the Celestron SCT's is not good but it is very well manageable (at least with my sample). ... I now see your follow-up video. That makes my comment obsolete :-)
I hope you don't mean the first comment is obsolete! Haha. Thanks for watching both videos, Patrick! I made several changes, but the backlash setting - I agree with you - is the "biggie."
Many thanks for posting this James. I feel your pain in this regard! Edge HD8 (1422 mm FL) owner here, using NINA. About 2 years in to this autofocus thing. Could not get along with the Celestron SCT focuser gyrating around and locking up! Settings - I am currently using a star HFR, Hyperbolic curve fit, R2 threshold of 0.8, autofocus step size of 500, (Sesto Senso V1 on a FeatherTouch SCT Micro), backlash comp=Overshoot with 1200 steps in the OUT direction. Recently autofocus runs have improved. Getting something that resembles the curve with acceptable HFR's repeatedly. Clear nights in the UK at present, are as rare as hens' teeth!
Thanks Paul! Step 1: try a much larger backlash setting, Step 2: see what Hocus Focus can do for me, Step 3: replace the Celestron focuser. I just wasted a good 3 or 4 clear nights with this. Not as rare as hen's teeth here, but I hate to lose valuable imaging time. Much more painful for those of you at higher latitudes. Thanks for watching!
Appreciate your pursuit on this, James. I'm a dedicated user of Nina auto-focus -- and I use it on a narrow critical focus zone (RASA 12). My curves have been fairly good. I'll be watching your findings closely. Doug
I imaged with only SCT's for years and had a heck of a time with focus. I always found the problem to be either mirror flop/shift or focuser slop. If you don't always focus moving the mirror up after going well past focus it will shift on you. Lock the mirror in place and put a crayford on it.
Interesting - I enhanced the NINA AF routine back when my main rig was an EdgeHD 800. Not quite the same as a C9.25 but close. The main difference is that I was using a Moonlite Focuser, so not moving the primary mirror. That said you have very little HFR change, even with APT. Have you tried maybe doubling the step size? The points being all over the place could also be due to poor star detection. Have you tried turning on star annotation to actually see what was detected as a star in each image? With star annotation, you can then play with the star detection and stretch settings until you get a satisfactory detection of stars...
Excellent suggestions! I have played with step size and going further out on the curve. I think the star detection is worth digging into. Some people have commented that the stretch parameter and black point are used by the autofocus routine. I did not realize that. Also, there's no doubt that leaving the mirror in one spot and moving the camera is the best way to go. Who knows? I may end up going that direction myself. I've got a few more holes to dig on this topic...
.....that´s why I decided to mount and ext. focuser on my C8: "what are my nervs worth"? the ext. focuser is expensive but priceless for me. for the C8 there are only a few options because of the backfocus. I took the MoonLite LightCrawler (smaller version of NightCrawler) with integrated rotator. Very happy since then, no problems, no backlash. 2. My C8 wasnt correctly collimated and I tried several methods; finally I ended up with the Hotech Adv. Laser Collimator. This showed a slightly misaligned optical back..... cheers, Peter
That is the best way to go, especially if you have an EdgeHD with primary mirror locks. That said, now that I've dramatically increased the backlash setting and increased the exposure time, I'm consistently getting good hyperbolic curve fits. So, I'll keep my fingers crossed and my wallet in my pocket for now. Thanks for watching, Peter!
SCTs are such a hassle. I ended up buying a rack and pinion focuser that is designed to fit on the SCT visual back. Now I focus with a bahtinov mask, and then fine tune with the rack and pinion. Bonus points because now it works with every autofocuser when I go down that route.
Hi Clint. That’s a good and effective focuser option for the SCT, especially if you have an SCT with mirror locks. I took the much less expensive route which also comes with additional headaches and tweaking until you get it right. If you see later videos, I finally did get this figured out and am getting good focus results. As for your observation that SCTs are a hassle… PREACH! Totally agree. I always have a bit of dread when switching back to my SCT from using a refractor and look forward to moving back to a refractor after babysitting the SCT. Thanks for watching !
That is EXACTLY what I am facing at this time!!! EXACTLY in my mind, my thoughts about what's the problem... I have a Celestron EDGE HD 8. Can't autofocus with my EAF. Crazy measurements, pixels taken as stars when HFR is measured from NINA... a real mess! I sent my C8 to the shop that sold it to me, cause I am in warranty, to check it. Mirror flipping is my main concern. Something in the mechanics area of the primary mirror. Or its just impossible for NINA or any astro software to measure the HFR of stars that an SCT is producing... These faint smoothed at the edges stars. I really dont know, but at least I dont feel alone now in this head scratching situation!!!
This seems to be a common problem. Mirror flop is a problem with these designs. I will really increasing backlash, increasing exposure time, use an inner crop of 0.8, change auto stretch to 0.1, and black-point clip to -2. And I’ll give Hocus Focus a trial run. There’s got to be a good recipe out there. I’ll keep you posted with my findings!
Well, after one night, so far, so good. I increased my backlash setting by 3x and used 8 sec Lum exposures (instead of 4ish). I didn't change the nominal stretch factor or black-point setting from my nominal (I was seeing good results without having to make that change). I got pretty good curves for each of 12 autofocus runs over the night. Will try again as long as I have clear skies and will summarize test results in a video later.
Hey James. Very interesting video. That made me curious. I read a little in the N.I.N.A. "autofocus" section. And what I guess and see in 4:26 is that your backlash for your SCT is definitely NOT at 20 steps (as said at 1:16) but rather AT LEAST (!) at (4:26 right top graph) 17475-17437=38 so at least double of what you think, since the focus (HFR) does not change in this graph over at least 38 steps. And this flattening of the curve on the right hand side is exactly what you see in the N.I.N.A. "autofocus" documentation online. - All very well explained there. - Additionally you can check the autostretch factor and black point clipping. - Good luck! :)
Hi Tully. It is possible that the colder temps have caused reduced contact between the Celestron focus motor and the SCT shaft thereby allowing more "apparent" backlash than when I first measured the backlash as shown in the graph. Also, I have had to remove and re-attach the focuser several time because of other issues with that focuser. Bottom Line: I need to tighten the set screw to improve the connect with the shaft and increase the backlash setting. You and another commenter mentioned the autostretch and black point settings. I had not realized those were used in the autofocus assessment. Thanks for watching, Tully!
@@Aero19612 Ah, okay..so you are using a Celestron focuser.. Sure, if there is bad contact between focuser and SCT you could get crazy numbers... I already wondered because of the very different focussing graphs in your video.. - Other than that if you try to find your backlash I would really watch the curve ..that's the best indicator I would say. Since I personally have no electronic focuser (yet) it was very interesting for me to learn about that myself. :) Would be great and very interesting to see your progress on that (maybe in a future video?). - Clear skies
The Celestron focuser is not a well-designed piece of equipment and may well be at the source of this issue. In the end, I may be shelling out for an alternative focuser. I'm sure there will be another video...or 10 on this topic. I never seem to run out of problems and hopeful fixes!
I have a thread about focuser slipping on cloudy nights that shows my suggestions on pegasus zero af and adjusting pully size that have worked amazing since I did the mods
Great summary! Had to chuckle when I saw you post this. As a fellow 9.25 owner, I spent many nights (and lost hours) attempting to make this work reliably. It definitely put a damper on my enthusiasm to use this setup more. I may try autofocusing using a separate Crayford focuser, so I can leave the primary mirror alone.
We need a support group. After using refractors over the past couple of years, I always have a sense of dread when I switch to the SCT. Much of that was due to the CGEM. The EQ6-R has erased those concerns, and now I'm down to focus/focuser. Thanks for watching, Andre'!
Hi James I have the Celestron Focuser on my SCT albeit an 8 inch EdgeHD Right now I’m set up for imaging at F10 With N.I.N.A, I’m consistently seeing good symmetrical focus V curves as I’d expect during auto focussing I set the auto focus to take place with an SVBony CLS filter and then let filter focus offsets take care of my SHO narrowband filters and RGB broadband filters I measured my filter focus offsets via Bahtinov mask and using Bahtinov Grabber as a one off exercise during setup/discovery FWIW, my autofocus settings in NINA are: AutoFocus initial offset: 4 AutoFocus step size: 20 Exposure Time: 5s Disable Guiding:ON Focus settle time: 2s AutoFocus Method:Star HFR Curve Fitting. Hyperbolic Backlash compensation: Overshoot Backslash IN/OUT: 0/100 Attempts: 1 Brightest stars: 3 Outer crop: 1 Inner crop: 1 Binning: 1 R threshold: 0.8 Hope you figure it all out BTW, ASTAP has a great tool built in for measuring star shape aberrations and image aberrations/tilt in your subs Clear skies
Thanks for sending your autofocus parameters. We are very close, except you have a much larger backlash. I will change my backlash next time out. I agree: use filter offsets for the SHO and maybe even RGB. I did a similar measurement a while back and found my ZWO filters basically require no offset. Thanks for watching!
Hi James, I have a Celestron 8" Edge SCT and I saw the exact same issue as you report here. In November I moved over to using NINA from SGPro and bought a new CMOS camera. Probably not wise in hindsight to make so many changes at once! But i saw the same focus curve you report here. Like you, I am using the Celestron focuser for 2 years now, but I have only had great experiences with it. I realised that my focus step size was too small, after some experimentation I arrived at a step size of 30 for my f/7 configuration. After this change, I have never had any problems with focusing and have found the entire setup to be extremely reliable. I have not tried f/10 yet in NINA, but looking over my old SGPro configuration the step size for f/10 was 50...although I understand SGPro uses a different fitting algorithm so this may not be directly comparable. Looking at your step sizes here you are at 20, this might be a bit too small for an SCT, especially as you have a longer focal length the optimum focus tolerance will be extremely tight. I hope you fix this soon, let us know how you get on...clear skies!!
Thanks David! I played quite a bit with step size. I think I'm around 30 now. As I mention in the video, I want to see a two- to three-fold increase in HFR over the minimum HFR to get a good curve. I will be getting the scope out tonight and will play with focus some more. I'll increase backlash to 120 (3x higher than what I'm using) and maybe 10-sec exposures (2x longer). Maybe I can try Hocus Focus out too. Thanks for watching! Btw, I'm glad to hear that your Celestron focuser is treating you right. The focuser is fine, maybe even really good. My issues stem mostly from its attachment to the SCT body using the same short M3 screws used to attach the orange cosmetic plate. In my case, one of those threaded holes is damaged. Maybe the longer screws and Loctite will put those issues in the rear view mirror.
You've got my sympathy James! Great Video as usual! Thanks for showing the expected HFR values for several f#'s. Helpful! Good luck with the autofocus. Hope you don't wind up going back to APT...
Haha. I'm glad APT has got both feet in the autofocus pool, but I'm so dug into NINA that it will take a lot of coaxing to get me to go back. That said, I am tempted to try APT just to see if it provides consistent results. Thanks for watching, Dale!
I have a c9.25 sct. Focusing through the telescope focuser was inconsistent for me for years. Eventually bought a used Optec TCF-S focuser. I don't use temperature compensation though. For the longest time the only options available for imaging was Maxim, the SBIG capture software called CCDOps, and later Sky X with its image capture routines. Then came Focusmax. So I bought the TCF and used first the free version of Focusmax and then the paid for version when CCDSoft bought out the software. Personally, manual focus of the SCT works well enough to get close so I can use software to run a V curve. Then Focusmax runs a fine focus. When my QHY 268M comes, I'll be using Nina. I'm hoping I can sort things out by then. I would never trust the accuracy of the focus knob on the sct though.
In a later video, I show what I did to get my focus issues sorted out. It was very reliable and accurate after that. Got my fingers crossed for the upcoming galaxy season!
Mirror flop, Coma, mechanical slop in focuser, shit Celestron focuser, collimation issues. With an Edge HD you can at least lock the primary to mitigate mirror flop. I'd also install a Feather touch micro focuser and motorise that with an EAF or a motorised secondary Crayford. You will still have Coma and this is less of an issue on the newer Edge HD models. I used to have a Meade 16" SCT Classic so I feel your pain
Ditto with my EdgeHD11 running on ASI Air Pro. I am finally getting pretty good results after: Added a Moonlite focuser behind the OTA. Locked the primary mirror in most favorable focus position with Moonlite focuser midway. Slightly PRELOAD PRIMARY MIRROR with the SCT focus knob with the two primary mirror rods LOCKED. The biggest reason that the ASI Air Pro Auto Focus routine was useless was that it begins immediately after moving the EAF. If the EAF has just moved the primary mirror, there is enough slop and mirror resonance as it becomes stationary again (at 2800 mm FL), to blur the image that it is using to determine the star size (focus). Locking the primary mirror and preloading it also drastically helps with mirror flop/movement. Use the EAF on a secondary focuser between the OTA and camera.
Falls in the category of "wish I knew then, what I know now". Locking the mirror and using a focuser that moves the camera seems like the most reliable setup for an SCT. Before I nuke my current arrangement, I want to pull all threads to see if I can get consistent autofocus with this house of cards. Thanks for watching!
Hi James, I use a C11 EDGE HD and I have perfect autofocus in NINA (Perfect curve). You have a 20 steps backlash, ok, you use 40 steps backlash compensation in NINA, that is good to compensate for the focus motor backlash NOT for the mirror flop! Be really aggressive with this value. In your case I should use 100 or even 120 steps. The reason is that the mirror is now really push back against the central tube when it comes back in. The reason the right side of the curve is all over the place is because the mirror is flopping around and is not in its good final position. This 120 steps is not really a problem because they will only be used twice, once at the start of the focus routine and once at the end (and it is important that this last one pushes firmly against the mirror), the other steps in the focus routine are only pushes. And I am not exaggerating, it is a perfect hyperbola I get. (with NINA build in focus routine, not used HOCUS-FOCUS yet). So I would say try it and I hope you get the same results as I do.
Thanks for the input! I think you are right. I plan to do just that: increase the backlash to around 120 as you say and maybe increase the exposure time. I hope to get the scope outside tonight and do some testing. Who knows, maybe even some imaging!
It appears to have worked! I increased the backlash to 150 from 40 and increased the exposure time from 4 sec to 8 sec. I did not change the autostretch or black-point clipping from earlier settings. All of the curves acquired over night look like hyperbolas. Will try again as long as I have clear skies then summarize findings/observations in a follow-up vid.
Hi James, lots of good info here. you are (fast) becoming an Astro SME. The HFR and HFD provide an indication of best focus, so why not scale each column to be a maximum of 10 e.g. NINA column has values from 10 down to 5.70, PixInsight has values from 10 down to 5.48 and for the other columns? Plotting all these values will show how the shape varies with each algorithm. All Focusers have some backlash and some quirks too. My Astro Pegasus's Achilles Heal is definitely the coupling between Focuser and Motor. Get that secure and many anomalies disappear. On Refractor, If the coupling is OK then manually trying to gently move the focus is prevented. On SCT the same type of Motor with a different bracket isn't an issue. Assuming all is coupled correctly the backlash value can be set higher - yours can easily be set to at least double your step size? The question of mirror flop can be checked by either testing at higher/lower Azimuths or at least raising the scope up to the target before focusing? With Autofocus, you start high and come down to get points for a curve, evaluate the curve and set the focus to the calculated best focus. Returning to the best focus point should be easy even allowing for backlash e.g. over compensating. In NINA can you trace the positions sent to the Focus Motor to check how it is being driven? Backlash compensation works differently in different Apps. In my Focuser the backlash is removed differently to SGP Pro's method and probably NINA's. Having more than one method in operation would be confusing at best. It would be tempting for the Motor manufacturer to hide their sloppy mechanisms with built in backlash compensation but that would defeat the controlling Apps like NINA trying to apply their own algorithm. I noticed that the order of taking the images is not in sequence (based on the timestamps). Can NINA save the images during the Autofocus routine or at least repeat them in the same order? To replicate NINA manually you want a series from right to left and then the final image which should be the same/similar to an earlier point if everything worked out OK. Simon
Hey Simon! That's a good idea. The magnitudes of the various metrics are all over the map. At the end of the day, we would each metric to provide the same optimum focus point. Of course, they don't. In this example, APT's HFD is best, APT's FWHM is next, NINA HFR/APT InvPow/Pixinsight FWHM are equally poor, and ASTAP is worst of all. Just one example though. I agree. Backlash in the focuser and, for SCTs, in the mirror moving mechanism may be the biggest problem. If I had it to do all over again, I'd buy an SCT with mirror locks and a focuser that moves the camera instead of the mirror. For now, I'll try to get this mess working "well enough." First thing I'm going to do is dramatically increase the backlash value I've been using. Stay tuned! Correct. The images are not in time/position sequence. I wish there were a setting that would allow NINA to save each picture during an autofocus, but there isn't as far as I know. That would be a good feature for geeks like me! Thanks for watching!
Hi James, i had the same problem with my RC6" and NINA. The curve was much better after a good collimation. You have excludes this option with you tests. With good collimation, i mean all stars in the your Field of View, not only the center star. A SCT Telescope is a little bit limited because you can't collimate the primary mirror but with my RC telescope and full collimation i got nearly good/useable hyperbolic curve. Regards Andreas
Thanks for watching, Andreas! My problem turned out to be excessive backlash in the focuser (see the video I posted about 3-4 weeks after this one). I’ll keep my fingers crossed.
Try an optec leo when your budget provides. Then you can lock down the primary mirror at rough focus and the leo moves the imaging train, so no errors introduced by mirror issues (assuming not too much flop). The leo is really great hardware, you'll appreciate the engineering and build quality of optec products.
Thanks, Jason! I agree, it’s much better to lock the mirror and move the camera. I’ve made some changes to my process with my Celestron move-the-mirror focuser by increasing the backlash setting by a factor of 3. That seems to have really helped as I am now getting consistently good curves. We’ll see how long this lasts!
@@Aero19612 Good to know about setting the Celestron motor focuser to 3x backlash. I've got the celestron motor focuser too, and still use it when the HyperStar is on (because mirror focus is then the only option). So, I've been manually focusing with bahtinov mask when using HyperStar due to this same issue, so I'll have to try the 3x comp and see if that works at f/1.9. As an aside, maybe in the future you can do some episodes about the HyperStar, if you get one and want to go down that rabbit hole.
Thanks, Dennis! I’ve looked at it. It’s a very nice focuser and the most trouble-free way to go with SCTs. I cranked the backlash waaay up and made a few other adjustments and everything seems to be working well (for) now. Fingers crossed! Thanks for watching!
I get seem to get too much backlash on my SW APO to achieve successful autofocus in ShapCap (Pegasus focuser), although I haven't tried generating a V curve in NINA. Maybe I'll get better luck if I give that a try. Thanks for this video and excellent explanation! I love my C8 but I don't know if I have patience to collomate it in the cold. Then again, I'd use the same excuse to avoid using it on a mild and humid summer evening. Well, priority one... I'll give NINA autofocus a chance.
As a general rule, I've had great results with the NINA autofocus routine. Give it a try! I'm really good at generating excuses when it's "too cold" and "too hot." That leaves about 10 days for imaging. Thanks for watching!
Hello James, I am actually quite a large fan of your videos and I believe have watched each of them several times due to the joy of learning many things about the hobby from you. But, in a interesting change of events, I might can actually help you. As not only a NINA and Hocus Focus user myself, but also an Edge HD 9.25 user as well, I can understand the pains of trying to achieve focus in an form of automated fashion. I have toiled for 20-30 hours testing various means of Autofocus on my edge. The problem is one of design though. The graphs you posted are almost identical to ones I saw when using a ZWO EAF on the main focus knob. The issues is mechanical. The way Celestron designed the SCT's focus system seems shortsighted at best, or just cheeping out at the worst. You see, the primary mirror has 3 points of contact on the inside of the scope. Two heavily greased guides, and one guide that is used for moving the mirror. The problem, as I am sure you are quite aware of now, is that that mirror can skew, changing the stars, but also creates an impressive amount of backlash. In colder temps, that grease locks up and makes focusing even more difficult. I believe people with good copies of the scope can use an EAF potentially, but it is far from reliable and do not negate the issues of the dreaded 'mirror flop'. There is a solution. Prima Luce Lab makes the Esatto which is a dedicated focuser for many setups. It has zero backlash (each one is calibrated at the factory and mine has zero measurable backlash) and solved the mirror flop issue. Basically, you just setup the Esatto and all your backspacing to be about 5 mm short of the recommended backspacing for your scope, the as the Esatto moves out the remaining 5mm, it will achieve focus at the same point where the recommended backfocus is. This focuser solves several issues. First, focus curves are just like a refractor, with just a smaller HFR variance. Which my GT81, prime focus gets an HFR for 2.2 but my autofocus routine can measure out to an HFR of almost 20. The problem with SCT's is the prime focus has an already larger HFR of around 4, as you mentioned, so when you defocus the star, it becomes so large that NINA and the other detectors do not think it's a star. With my 9.25, I can get about 3-4x ratio for the extreme points of the autofocus curve which always seems to work fine for me. The next issue it solves is mirror flop. Since you are focusing with the Esatto, you can move the primary focuser so the lead screw is engaged and pushing the mirror again against gravity and lock the mirror locks. This means the mirror does not move (meridian flips do still need an AF run) and it reduces the mirror flop issues. Lastly, AF speed is increased with the Esatto and Nina. Running Hocus Focus and Filter Offsets, I can do an Autofocus routine in less than 2 minutes. Oh, and the filter offsets are a great addition too as since the Esatto moves so reliably you only have to use Lum as a primary filter and then just offset from there. Lastly, since the primary mirror is disengaged for focus, the target does not move as you focus. For planets this is almost a must as you can focus without ever losing the target. For DSO, you do not have to worry about if autofocus is going to move things such that you have to reslew to the target. I'd be more than happy to help here because I have used my Edge on several deep sky projects and have gotten over 100 hours of imaging time on it in the past few months so I understand the system very well. The Essato is a prime piece of kit. Especially given is insanely small step size. It is price for sure, but it would solve every issue you have in this video and more.
Thanks for watching all these stupid videos, Noah! It's a tough job, but someone has to do it. Actually, you don't "have to" so I really appreciate the effort. I totally agree-SCTs bring in a ton of mechanical baggage that good refractors do not. I also think Celestron was only thinking of visual use when they designed their base level SCTs. The EdgeHD versions are better suited to astrophotography, but still have extra mechanical bits that refractors don't have. Thank you for the Esatto recommendation! You make a very convincing case and a new focuser may well be the path forward. I don't have much confidence in this Celestron focuser. It has cost me many nights of good imaging. At some point, I'm going to get totally fed up with it. Off to Google the Esatto...
@@Aero19612 The reason I enjoy the videos is due to the technical nature of them. Not many Astrophotography people put in the level of detail you put in, and I mean that. I take a very academic approach to astro and watch a large amount of videos on this hobby to study and take notes. But that is a wise about about the visual use. All, and I do mean all, of the issues we face with the C series and Edge series would be solved if Celestron would have just locked the primary and put a crayford or some measure of external focuser on the back of the OTA. I do fully believe that the Esatto solves all the issues you face. You should just need two adapters for it and then no more fussing around!
Indeed it would. Just looked it up. The Esatto 2" with the standard SC adapter and the M56/M48 adapter would work. They're not in stock at the moment. Surprise, surprise.
Good video James, you have very variable inconsistent results , did you run while you have cloud running through as have many outlines, so if you have a measured backlash of what you say is 20 then it can't be that.. could it be mechanical issues from the focuser ( you've had plenty of issues with it so far) So the EAF is working perfectly fine on the frac Your collimation doughnut looks ok, you can overlay collimation aid in APT to do the same thing.. but remember your collimation doughnut isn't in the same place on both sides of focus... and as have mentioned before you want to approach focus from one side not reverse and come back as you have a slight backlash error from the mirror drawtube, you need gravity as your friend...
Thanks Derek. Even though I measured the backlash when I first installed the focuser, I have had to remove and re-install it several times. It is possible that the interface between the focuser and the SCT is not as tight as it was originally. Anyway, I'll try longer exposures and a higher backlash setting. If all else fails. I'll go a different route with another focuser.
it's not mirror tilt, I have a secondary mirror focuser I added to my stack - same problems. Algorithms are the problem! Specifically it's the inner circle/astigmatisms of sorts.
I'm running an Edge 8 with the celestron focus motor. My backlash compensation is around 120. Step sizes are similar to yours.. around 30 IIRC. Mine has been working fairly well with SGP (which has a notorious AF routine). It struggles the most on nights with poor seeing.
Hi James. I have never observed focus curve asymmetry with NINA on my C9.25 with the same focuser. The NINA settings I use are nsteps=5, step size=28, overshoot method, 50 steps IN. I would be surprised if mirror flop gave the same consistent asymmetry at different locations in the sky. I would be equally surprised if focuser slip was this consistent. If anyone can get to the bottom of this, I am betting on you. Please let us know what you find. (PS taking multiple shots at each focused position might give you a better trend but the HFD will be higher as I am sure you are aware).
Hey Bob. Our numbers are pretty similar. Interesting how you're getting good results and I'm getting the results I'm seeing. That said, I do occasionally see a nice symmetric curve, but I get more of these asymmetric curves. I'll keep on digging, I guess.
What were your seeing conditions on the night you ran your tests? Was your scope completely thermally equilibrated? With the very fine pixel scale of the SCT, the stars are extremely sensitive to seeing and tube currents. If you run an OAG this is immediately apparent as the guide star wobbles and stretches all over if the seeing is marginal. With my C11, on calm nights the focus curve runs perfectly, with iffy seeing it is quite a challenge. Additionally, the right side of your SCT focus curves appears "flat", which may indicate that all the backlash is not yet out of the system or that you are suffering from mirror tilt with the focuser movement. With overshoot compensation, there is no need to be exact and it may be better to increase the compensation to achieve the full 1/2 counterclockwise focuser turn needed to seat the mirror forward consistently. The EdgeHD series are superior in this respect as you can lock the mirror in place and use a secondary focuser such as the Optec Leo, thus avoiding this issue (also mitigates mirror flop as you meridian flip). I have found the Hocus Focus PSF fitting to yield slightly more robust star identification. Binning to 2x2 can reduce the vulnerability to seeing somewhat. Were your 5 sec measurements a single frame or the average of several frames? Taking multiple exposures at each focus step can help average out the seeing as well (at the tradeoff of being time consuming). The variability in HFR measurements can serve as a measure of seeing quality as can the HFR itself - with a theoretical 12.4 micron Airy disk at f/10, minimum HFRs of 6 in your test on 1/29 suggest marginal seeing, as does the variability from 4 in one run to 6 in another. If you do not have time to leave the scope out for 1-2 hours before imaging, or you have 40 deg F temp drops overnight like i do, insulating your scope with 2 layers of Reflectix may be the best option. Works well on my C11. Hope this is helpful, long focal length imaging is a very different game than short refractors!
I posted a follow-up video to this one showing that a significant increase in backlash solves the problem. I made some other, probably minor, changes as well. I will eventually try Hocus Focus-it looks very good. Thanks for the suggestions!
Hey James, nice video! I wrote Hocus Focus, though I haven't been able to test it much at high focal lengths since I got rid of my SCT some time ago. Would you mind helping me improve star detection for this use case? I believe this will come down to choosing the right detection parameters, and some sample data will really help with this. Thanks!
Definitely worth trying Hocus Focus - and maybe it calls for a Hocus Focus SCT variant. I suffer similar issues and still don't know the cause, but out of focus stars look different on opposite sides of focus on an SCT (or at least on mine) so perhaps that contributes?
Hey George. I'm looking forward to giving Hocus Focus a try. I'd be happy to help. I'm socked in with a weather system for a week. Is there a way to use Hocus Focus in a stand-alone mode to assess the images I discuss in the video?
@@Aero19612 if the images are saved, you could use the file simulator to read them from a folder. It's a bit finnicky to do it this way as you need to load them in the same order an auto focus would. I'm planning to add a "Save" mode for AF, as well as a replay one that'll use the images and focus positions from a prior save and redo star detection and curve fitting with the current settings. Feel free to ask me (jokogeo in the Discord channel) for help if you need!
Hi, I believe I am using a very similar configuration - C9.25 with Celestron SCT focus motor. FWIW I am not seeing these issues, the autofocus curves are looking much like the ones you are showing for your refractor. I did much less analysis than you did and changed only a few NINA parameters based on earlier experience with other programs. I select an ROI between 20%-80%, keep guiding on and have overshoot set to 100 "in". Not sure if any of that is causing the difference in the results.
Thanks for watching. When I leave guiding on, the torque generated by the focuser causes PHD2 to lose the star and then it starts the error beeping, so I stop guiding during focusing. One difference could be that I'm using an OAG so the image scale is much smaller than if using a guide scope. I will try a higher backlash setting and longer exposures.
Hi James . I didn't see you mention any relationship between the "good" focus curves and which direction the focus shaft was turning . I'm told I must always approach focus in a CCW direction due to the mechanics of the SCT . Also , your 9.25" is not an EdgeHD OTA is it , or is it ? If you read the Celestron EdgeHD White Paper on those OTA's you'll find a section dealing (top right page 7) with improved tolerances on the Draw Tube to reduce binding and freeplay when focusing . Maybe you have too much freeplay in the drawtube ? It would be nice to have an OTA to take apart just to see what is actually going on under the covers . Cheers./SRK
to compensate for any mirror slop, perhaps increasing the backlash compensator to say 100? BTW James did really replace a 0.75 pitch screw with a 0.50 pitch screw?
Jeremy: The screws in the C9.25 (may be different in other SCTs) are M3 and all M3 screws have 0.5 mm pitch. Agree: A consistent comment for folks getting good results is to use a much larger backlash setting like 100 or 120. I will definitely try that out if I ever see the sky again...
Hi Scott. I believe I'm always turning the shaft CW during focus, which means (I think) that the mirror is moving back toward the camera. I'm certainly not ruling out scope manufacturing tolerance as a contributor. Of course, I can't do anything about that. There have been several comments where people are using much larger backlash numbers than I am. I will try that and maybe winding the focuser/mirror way out and then back in also addresses some of the mechanical slop you're referring to. We shall see!
Hi James, I have an 8 inch Celestron SCT with a Celestron focus motor, and I seem to get better results when using overshoot on the IN side instead of the OUT. Also when you get disk like stars Nina has a harder time getting the HFR, you might want to use a slightly smaller stepsize or less steps to make sure the stars don't turn into disks. Anyway, hope you figure it out. I'll be interested in the results using Hocus-Focus.
Yeah, you're right about the in/out direction. I use OUT to minimize the focuser motion because of other issues keeping the darn focuser attached to the scope. I've played with various step sizes. Next time I get clear skies, I have a list of things to try, including Hocus Focus. Thanks for watching! Like the X-Files: The Truth is Out There
James, please tell us if what @Matt Nudi said in the comments, worked for you too. That thing with the auto stretch factor and black clipping settings in NINA, that improves autofocus in bad seeing conditions, that more or less we all have.
I had the exact same problem on my 9.25, same shape in the AF curve as your data. I found that longer exposures were needed, and it cleaned up the curve, something about how the intensity changes as a function of defocus for that direction and how HFR is calculated? I spent 2 nights trying everything I could think of, and longer exposures is what did it. (I was using the hocus focus plugin for NINA with backlash overshoot)
I think this may be related to the comment later about stretch factor and black point - related to how it determines the threshold to detect stars - especially with the central obstruction creating the hole. I ended up with 8 s exposures, which made for a long AF run, but that was faster than failing AF runs.
Thanks Brent! I think that's a good idea. Next time out, I'll crank up the backlash setting and the exposure time to try and get more and better-formed stars. I'd be satisfied with a 10-minute-long autofocus if I could get reliable results. As it is, I'm losing nights of imaging. Thanks for watching!
Hi, you really can't rule out collimation because the more defocus you dial up the less sanative it becomes. I find 3 rings is good for the defocus method but you will achieve best collimation on a focused star. Also you need to be careful on the way focusing is done because of mirror slop etc.... Good luck with it.
Thanks for watching, Troy! I'm going to try longer exposures, a larger backlash setting, tighten the focuser/sct set screw, and Hocus Focus. We'll see....
Thank you very much for this video, James. I have de C9.25 EdgeHD with a feather touch focuser and Starizona’s micro touch electronic focuser. Most of the time I use it at f2.3 with hyperstar, but sometimes also at f10. NINA’s auto focus routine has worked very well in both cases. Collimation certainly could be a culprit for the non symmetrical graph you’re having. I don’t know if also the Edge series have different mechanics for moving the primary mirror, since it has the locks. Something to figure out. Could you make Hocus Focus plug-in work? Regards, Alfredo
Hi Alfredo! I installed Hocus Focus, but have not (had) to use it yet. We've got a full Moon, so this is a good time to play with it. Once I got the scope back out, I increased the exposure time to 8 sec (was using about 4) for the Lum filter and increased the backlash by 3x - from 50 steps to 150 steps. I've got 3 nights under my belt and the curves look consistently good! I'll make a follow-up video on this maybe this week (lots of rain moving in...) The ideal setup is the Edge with the mirror locked in place and a focuser that moves the camera. For now, I'm limping along with my Celestron focuser that moves the mirror. I think my much higher backlash setting is compensating for the mirror slop. Thanks for watching!
@@mrkutube James is using the Celestron Focus Motor on his SCT @1:20 he shows the backlash chart - the chart states Celestron Focuser. He also mentions it in a previous video ua-cam.com/video/qXcZ3ZMMMco/v-deo.html
Hi James, is the Problem solve. What was the reason? I have the same Problem with my myfocuser2 and apt. Do you have used a reducer? It possible that the reducer has an effect on the hfr value and on the autofocuser? Greetings from germany
Hi Franz. Yes, I think I have solved the problem. Check out this video: ua-cam.com/video/M7GxDf2mcw0/v-deo.html I'm using NINA, which has "overshoot" backlash compensation. It is possible to do this in APT as well. If you're not, check out this video at 20:40: ua-cam.com/video/rT_cwkVMhno/v-deo.html I'm not currently using the reducer, but I think that would actually help. Also, if you can, perform focus with the Lum filter (if you have a mono camera) in Bin1. Let me know how things work out!
Do you have access to an ASIAir to try its focus routine? How do you get the Bahtinov focus error (pixels) data? I'm not tracking how you get the y axis data?
I don't have an ASIair, but I can and will test APT's autofocus routine. The Bahtinov focus error is reported in terms of pixel error. If you use Bahtinov Grabber, NINA's Bahtinov aid, or APT's Bahtinov tool, it will show you a focus error number. That's what I'm plotting.
I've been using NINA with my Edge8HD and getting perfect fits. However, when I tried with the Celestron focus motor, I got horrible results. Never was able get consistent results. I switched to a ZWO EAF and am having no issues. Nearly perfect every time. The fit parameter is .98 or above every time. Absolutely NOT a fan of the Celestron device. BTW, I am running with the .7 reducer so my FL is about 1400mm.
Great feedback, Steve! I'm glad there's a light (or should I say: hyperbola) at the end of the tunnel. The thought has occurred to me to replace this focuser with the ZWO. I'll dig a little more, then pull the trigger if I can't get what I'm looking for.
Hello James, I can second that. I never got the Celestron focuser for my C9 and went straight for the ZWO. First time I tried it, I had used the same stepsize as with my small refractor. No chance, but once I increased it to 100, the points are nicely following the curve.
Thanks for the feedback and thanks for watching! I'll give this Celestron focuser the rest of this galaxy season and if I can't get good results, I'll find another otion. I was close to buying the ZWO EAF some time back but held off.
Thank you James. I found your videos back last spring when I started with this hobby. I think Ed Ting made we want to buy a C9 but you gave me the confidence of actually getting one. Since (and it is probably true) a lot of people recommended starting with way smaller FL but I thought with support from videos like yours, I might get away with it. You were also the reason I used APT back then and avoided the Celestron focuser. And I found your channel because you have a C9, but I stayed because I like your technical and analytical style. So once again, thank you James.
You could check the number of stars detected, you are all ready using just barely enough data to get any sort of statistical read - at either end of the bell I'm guessing the detections drop down. Because the stars are larger at a long FL they get fainter, and couple that with the bad morphology it probably can't pick out all the stars. You might try focusing on an open cluster or somewhere in the main galactic lane. You are also MUCH more at the mercy of seeing at 2k mm and the only solution there is taking more frames.
There are very few stars in this FoV. I think the real statistical benefit comes from generating multiple points (even if only with one star) at different focuser positions and then using the hyperbolic curve fit to deal with that measurement error. The "problem" with using multiple stars is that all of the stars will have the same poor shape if, say, a gust of wind affects the measurement. I agree, getting the image parameters right to see "enough" stars while not imaging too long and permitting wind, mount, and seeing conditions to dominate. Thanks for watching, Michael!
Hey James - did you raise your issues on the NINA Discord server? There's an autofocus-support channel for just this sort of discussion. The core devs and the author of Hocus Focus are certainly around ...
Tom P: Yes, I did raise the issue on Discord. George (developer of the Hocus Focus plugin) suggested I try Hocus Focus. I had not heard of that plugin before George pointed it out. Thanks for watching!
HFR depends on just about everything. (1) Telescope aperture determines the lowest achievable HFR, (2) Seeing at zenith increases the HFR, (3) Air mass increases the HFR relative to zenith, (4) Wind increases the HFR, (5) Mount performance (uncorrected periodic error and gear noise) increases HFR mostly in the RA direction, (6) Imprecise focus increases HFR, and probably more things that just aren't coming to mind right now.
I just take what I can get. We have very little control over the FWHM - driven by seeing and target altitude. When I stack images, I will leave out images with "high" FWHM, but you can't go too crazy with that approach, or you'll never have enough data to create a final image. So, my approach is to collect 20-30 hours on a target, toss out maybe 10% to 15% of the images with the worst FWHM, then move on to a different target.
@@Aero19612 Ok. I reject anything above 3.5” for 4” refractor and 130mm Epsilon; anything above 3” for C9.25 and 140mm TEC. Every frame with eccentricity >0.5 also thrown away. Rejection for AVX +ZWO focuser was 30-50% frames. For 10 micron + Moonlite focuser 10. Only for very bright targets. Blur and Noise XT totally changed game.
I've installed Crayford focuser on my HD 9.25" and never looked back. Before that, the mirror flop was causing the autofocus routine to fail on a regular basis. Great video, thanks for sharing.
Definitely the way to go. Who knows, I may end up there too soon! Thanks for watching!
Would you mind sharing which Crayford you decided to use? I have the 8 HD and considering this option
@@Dudleydogg I've got a MoonLite focuser.
I was having the same issues in NINA on my C11 until I changed my auto stretch and black point settings to the recommended values in the NINA documentation for “poor seeing” conditions. After making these changes, I am able to get very consistent focus curves.
VERY interesting. I had not heard that before and did not realize the stretch and black point played into the autofocus process. Thanks for watching, Matt!
Hi James,
You are certainly correct about the SCT versus a refractor when it comes to autofocus. My C-11 very rarely will have all the stars somewhere close to the curve in NINA. However, on super clear nights, I have seen some pretty remarkable fit to the curve. I am glad to hear that you too have (had) issues with the Celestron focuser coming loose. I periodically need to tighten the set screws, but if I tighten them too much, the focuser will get stuck.
Hey Pat. That Celestron focuser, or more accurately its interface with the scope, is terrible. The best approach is to leave the mirror alone and move the camera, but those focusers are awfully expensive. I'll swirl around the drain a while and see what happens...
I am using a ZWO EAF on an old (circa '83) Orange tube C8 in NINA with great results, good fit, etc. Your scattered plots remind me of when I was first attempting the setup and had the backlash value in both the IN and OUT fields before I learned it was one or the other. I tried Hocus Focus but did not have a good experience with it so went back to stock. Good luck getting it going!
Hey Bruce, I wrote Hocus Focus. Would you mind taking pictures at various focuser positions as if you were focusing, and sharing them with me (jokogeo) in the Discord channel? This will help me improve star detection for high focal lengths.
Thanks Bruce! I'm glad you're getting good results with your setup. Don't change a thing! As I've been pondering this problem, I have been having evil thoughts of moving to a different focuser. I will try a very large backlash number next time I'm out.
@@georgehilios5079 Sounds good. The wait for clear skies is on (as always).
@@Aero19612 That is true too: I set my backlash *very* large, something like 800 in order to get it to work.
@@BruceMallett 800 seems just huge! That scope must be "interesting" to focus manually.
Great in-depth video as always James. I always learn something when I watch them. Keep up the great work.
Yeah. Here, you learned not to buy an SCT. Haha. Thanks for watching, Tom!
@@Aero19612 No, I learned to continue imaging with my ASIair Pro / EdgeHD 11 / EAF rig until I get better results with NINA and its auto focuser.
It’s a labor of love (maybe. Or just willful obstinance). We just need to keep in mind that there IS an answer. Thanks Tom
Thanks for your good video's. I use a Celestron SCT 11 with the same focuser as you do. By using a very large backlash compensation (step-size 1500) and making sure the last movement of the primary mirror is "out" (against gravity) I have good and consistent focus results in SGP. I use this setup for over a year with no problems. I remember I had to reinstall the focus motor a few times before it ran smooth without heavy points during a revolution. So yes, all the mirror flop in the Celestron SCT's is not good but it is very well manageable (at least with my sample).
... I now see your follow-up video. That makes my comment obsolete :-)
I hope you don't mean the first comment is obsolete! Haha. Thanks for watching both videos, Patrick! I made several changes, but the backlash setting - I agree with you - is the "biggie."
@@Aero19612 Ha, no. My comment should have been: Thanks for your good video's. (period)
Many thanks for posting this James. I feel your pain in this regard! Edge HD8 (1422 mm FL) owner here, using NINA. About 2 years in to this autofocus thing. Could not get along with the Celestron SCT focuser gyrating around and locking up! Settings - I am currently using a star HFR, Hyperbolic curve fit, R2 threshold of 0.8, autofocus step size of 500, (Sesto Senso V1 on a FeatherTouch SCT Micro), backlash comp=Overshoot with 1200 steps in the OUT direction. Recently autofocus runs have improved. Getting something that resembles the curve with acceptable HFR's repeatedly. Clear nights in the UK at present, are as rare as hens' teeth!
Thanks Paul! Step 1: try a much larger backlash setting, Step 2: see what Hocus Focus can do for me, Step 3: replace the Celestron focuser. I just wasted a good 3 or 4 clear nights with this. Not as rare as hen's teeth here, but I hate to lose valuable imaging time. Much more painful for those of you at higher latitudes. Thanks for watching!
Appreciate your pursuit on this, James. I'm a dedicated user of Nina auto-focus -- and I use it on a narrow critical focus zone (RASA 12). My curves have been fairly good. I'll be watching your findings closely.
Doug
Thanks Doug! I'll keep you posted.
I imaged with only SCT's for years and had a heck of a time with focus. I always found the problem to be either mirror flop/shift or focuser slop. If you don't always focus moving the mirror up after going well past focus it will shift on you. Lock the mirror in place and put a crayford on it.
Totally agree. That’s the right way to go. Wish I knew what I was doing when I was doing it! Live, learn, spend…
Interesting - I enhanced the NINA AF routine back when my main rig was an EdgeHD 800. Not quite the same as a C9.25 but close. The main difference is that I was using a Moonlite Focuser, so not moving the primary mirror.
That said you have very little HFR change, even with APT. Have you tried maybe doubling the step size?
The points being all over the place could also be due to poor star detection. Have you tried turning on star annotation to actually see what was detected as a star in each image? With star annotation, you can then play with the star detection and stretch settings until you get a satisfactory detection of stars...
Excellent suggestions! I have played with step size and going further out on the curve. I think the star detection is worth digging into. Some people have commented that the stretch parameter and black point are used by the autofocus routine. I did not realize that. Also, there's no doubt that leaving the mirror in one spot and moving the camera is the best way to go. Who knows? I may end up going that direction myself. I've got a few more holes to dig on this topic...
Hi Cuiv! Hope you doing well and enjoying your break....
.....that´s why I decided to mount and ext. focuser on my C8: "what are my nervs worth"? the ext. focuser is expensive but priceless for me. for the C8 there are only a few options because of the backfocus. I took the MoonLite LightCrawler (smaller version of NightCrawler) with integrated rotator. Very happy since then, no problems, no backlash.
2. My C8 wasnt correctly collimated and I tried several methods; finally I ended up with the Hotech Adv. Laser Collimator. This showed a slightly misaligned optical back.....
cheers, Peter
...correction: mount an external focuser...
That is the best way to go, especially if you have an EdgeHD with primary mirror locks. That said, now that I've dramatically increased the backlash setting and increased the exposure time, I'm consistently getting good hyperbolic curve fits. So, I'll keep my fingers crossed and my wallet in my pocket for now. Thanks for watching, Peter!
SCTs are such a hassle. I ended up buying a rack and pinion focuser that is designed to fit on the SCT visual back. Now I focus with a bahtinov mask, and then fine tune with the rack and pinion. Bonus points because now it works with every autofocuser when I go down that route.
Hi Clint. That’s a good and effective focuser option for the SCT, especially if you have an SCT with mirror locks. I took the much less expensive route which also comes with additional headaches and tweaking until you get it right. If you see later videos, I finally did get this figured out and am getting good focus results. As for your observation that SCTs are a hassle… PREACH! Totally agree. I always have a bit of dread when switching back to my SCT from using a refractor and look forward to moving back to a refractor after babysitting the SCT. Thanks for watching !
That is EXACTLY what I am facing at this time!!! EXACTLY in my mind, my thoughts about what's the problem... I have a Celestron EDGE HD 8. Can't autofocus with my EAF. Crazy measurements, pixels taken as stars when HFR is measured from NINA... a real mess! I sent my C8 to the shop that sold it to me, cause I am in warranty, to check it. Mirror flipping is my main concern. Something in the mechanics area of the primary mirror. Or its just impossible for NINA or any astro software to measure the HFR of stars that an SCT is producing... These faint smoothed at the edges stars. I really dont know, but at least I dont feel alone now in this head scratching situation!!!
This seems to be a common problem. Mirror flop is a problem with these designs. I will really increasing backlash, increasing exposure time, use an inner crop of 0.8, change auto stretch to 0.1, and black-point clip to -2. And I’ll give Hocus Focus a trial run. There’s got to be a good recipe out there. I’ll keep you posted with my findings!
Well, after one night, so far, so good. I increased my backlash setting by 3x and used 8 sec Lum exposures (instead of 4ish). I didn't change the nominal stretch factor or black-point setting from my nominal (I was seeing good results without having to make that change). I got pretty good curves for each of 12 autofocus runs over the night. Will try again as long as I have clear skies and will summarize test results in a video later.
Hey James. Very interesting video. That made me curious. I read a little in the N.I.N.A. "autofocus" section. And what I guess and see in 4:26 is that your backlash for your SCT is definitely NOT at 20 steps (as said at 1:16) but rather AT LEAST (!) at (4:26 right top graph) 17475-17437=38 so at least double of what you think, since the focus (HFR) does not change in this graph over at least 38 steps. And this flattening of the curve on the right hand side is exactly what you see in the N.I.N.A. "autofocus" documentation online. - All very well explained there. - Additionally you can check the autostretch factor and black point clipping. - Good luck! :)
Hi Tully. It is possible that the colder temps have caused reduced contact between the Celestron focus motor and the SCT shaft thereby allowing more "apparent" backlash than when I first measured the backlash as shown in the graph. Also, I have had to remove and re-attach the focuser several time because of other issues with that focuser. Bottom Line: I need to tighten the set screw to improve the connect with the shaft and increase the backlash setting. You and another commenter mentioned the autostretch and black point settings. I had not realized those were used in the autofocus assessment. Thanks for watching, Tully!
@@Aero19612 Ah, okay..so you are using a Celestron focuser.. Sure, if there is bad contact between focuser and SCT you could get crazy numbers... I already wondered because of the very different focussing graphs in your video.. - Other than that if you try to find your backlash I would really watch the curve ..that's the best indicator I would say. Since I personally have no electronic focuser (yet) it was very interesting for me to learn about that myself. :) Would be great and very interesting to see your progress on that (maybe in a future video?). - Clear skies
The Celestron focuser is not a well-designed piece of equipment and may well be at the source of this issue. In the end, I may be shelling out for an alternative focuser.
I'm sure there will be another video...or 10 on this topic. I never seem to run out of problems and hopeful fixes!
I have a thread about focuser slipping on cloudy nights that shows my suggestions on pegasus zero af and adjusting pully size that have worked amazing since I did the mods
Great summary! Had to chuckle when I saw you post this. As a fellow 9.25 owner, I spent many nights (and lost hours) attempting to make this work reliably. It definitely put a damper on my enthusiasm to use this setup more. I may try autofocusing using a separate Crayford focuser, so I can leave the primary mirror alone.
See my reply below.
We need a support group. After using refractors over the past couple of years, I always have a sense of dread when I switch to the SCT. Much of that was due to the CGEM. The EQ6-R has erased those concerns, and now I'm down to focus/focuser. Thanks for watching, Andre'!
Hi James
I have the Celestron Focuser on my SCT albeit an 8 inch EdgeHD
Right now I’m set up for imaging at F10
With N.I.N.A, I’m consistently seeing good symmetrical focus V curves as I’d expect during auto focussing
I set the auto focus to take place with an SVBony CLS filter and then let filter focus offsets take care of my SHO narrowband filters and RGB broadband filters
I measured my filter focus offsets via Bahtinov mask and using Bahtinov Grabber as a one off exercise during setup/discovery
FWIW, my autofocus settings in NINA are:
AutoFocus initial offset: 4
AutoFocus step size: 20
Exposure Time: 5s
Disable Guiding:ON
Focus settle time: 2s
AutoFocus Method:Star HFR
Curve Fitting. Hyperbolic
Backlash compensation: Overshoot
Backslash IN/OUT: 0/100
Attempts: 1
Brightest stars: 3
Outer crop: 1
Inner crop: 1
Binning: 1
R threshold: 0.8
Hope you figure it all out
BTW, ASTAP has a great tool built in for measuring star shape aberrations and image aberrations/tilt in your subs
Clear skies
Thanks for sending your autofocus parameters. We are very close, except you have a much larger backlash. I will change my backlash next time out. I agree: use filter offsets for the SHO and maybe even RGB. I did a similar measurement a while back and found my ZWO filters basically require no offset. Thanks for watching!
Hi James, I have a Celestron 8" Edge SCT and I saw the exact same issue as you report here. In November I moved over to using NINA from SGPro and bought a new CMOS camera. Probably not wise in hindsight to make so many changes at once! But i saw the same focus curve you report here. Like you, I am using the Celestron focuser for 2 years now, but I have only had great experiences with it.
I realised that my focus step size was too small, after some experimentation I arrived at a step size of 30 for my f/7 configuration. After this change, I have never had any problems with focusing and have found the entire setup to be extremely reliable.
I have not tried f/10 yet in NINA, but looking over my old SGPro configuration the step size for f/10 was 50...although I understand SGPro uses a different fitting algorithm so this may not be directly comparable. Looking at your step sizes here you are at 20, this might be a bit too small for an SCT, especially as you have a longer focal length the optimum focus tolerance will be extremely tight.
I hope you fix this soon, let us know how you get on...clear skies!!
Thanks David! I played quite a bit with step size. I think I'm around 30 now. As I mention in the video, I want to see a two- to three-fold increase in HFR over the minimum HFR to get a good curve. I will be getting the scope out tonight and will play with focus some more. I'll increase backlash to 120 (3x higher than what I'm using) and maybe 10-sec exposures (2x longer). Maybe I can try Hocus Focus out too. Thanks for watching! Btw, I'm glad to hear that your Celestron focuser is treating you right. The focuser is fine, maybe even really good. My issues stem mostly from its attachment to the SCT body using the same short M3 screws used to attach the orange cosmetic plate. In my case, one of those threaded holes is damaged. Maybe the longer screws and Loctite will put those issues in the rear view mirror.
You've got my sympathy James! Great Video as usual! Thanks for showing the expected HFR values for several f#'s. Helpful! Good luck with the autofocus. Hope you don't wind up going back to APT...
Haha. I'm glad APT has got both feet in the autofocus pool, but I'm so dug into NINA that it will take a lot of coaxing to get me to go back. That said, I am tempted to try APT just to see if it provides consistent results. Thanks for watching, Dale!
@@Aero19612 I think you should try APT. What ever the results, you will learn something.
Agreed. I will try APT autofocus.
I have a c9.25 sct. Focusing through the telescope focuser was inconsistent for me for years. Eventually bought a used Optec TCF-S focuser. I don't use temperature compensation though. For the longest time the only options available for imaging was Maxim, the SBIG capture software called CCDOps, and later Sky X with its image capture routines. Then came Focusmax. So I bought the TCF and used first the free version of Focusmax and then the paid for version when CCDSoft bought out the software. Personally, manual focus of the SCT works well enough to get close so I can use software to run a V curve. Then Focusmax runs a fine focus. When my QHY 268M comes, I'll be using Nina. I'm hoping I can sort things out by then. I would never trust the accuracy of the focus knob on the sct though.
In a later video, I show what I did to get my focus issues sorted out. It was very reliable and accurate after that. Got my fingers crossed for the upcoming galaxy season!
Mirror flop, Coma, mechanical slop in focuser, shit Celestron focuser, collimation issues. With an Edge HD you can at least lock the primary to mitigate mirror flop. I'd also install a Feather touch micro focuser and motorise that with an EAF or a motorised secondary Crayford. You will still have Coma and this is less of an issue on the newer Edge HD models. I used to have a Meade 16" SCT Classic so I feel your pain
Ditto with my EdgeHD11 running on ASI Air Pro. I am finally getting pretty good results after: Added a Moonlite focuser behind the OTA. Locked the primary mirror in most favorable focus position with Moonlite focuser midway. Slightly PRELOAD PRIMARY MIRROR with the SCT focus knob with the two primary mirror rods LOCKED. The biggest reason that the ASI Air Pro Auto Focus routine was useless was that it begins immediately after moving the EAF. If the EAF has just moved the primary mirror, there is enough slop and mirror resonance as it becomes stationary again (at 2800 mm FL), to blur the image that it is using to determine the star size (focus). Locking the primary mirror and preloading it also drastically helps with mirror flop/movement. Use the EAF on a secondary focuser between the OTA and camera.
@@bobc3144L yep gotta lock that primary mirror
Falls in the category of "wish I knew then, what I know now". Locking the mirror and using a focuser that moves the camera seems like the most reliable setup for an SCT. Before I nuke my current arrangement, I want to pull all threads to see if I can get consistent autofocus with this house of cards. Thanks for watching!
Hi James, I use a C11 EDGE HD and I have perfect autofocus in NINA (Perfect curve). You have a 20 steps backlash, ok, you use 40 steps backlash compensation in NINA, that is good to compensate for the focus motor backlash NOT for the mirror flop! Be really aggressive with this value. In your case I should use 100 or even 120 steps. The reason is that the mirror is now really push back against the central tube when it comes back in. The reason the right side of the curve is all over the place is because the mirror is flopping around and is not in its good final position. This 120 steps is not really a problem because they will only be used twice, once at the start of the focus routine and once at the end (and it is important that this last one pushes firmly against the mirror), the other steps in the focus routine are only pushes. And I am not exaggerating, it is a perfect hyperbola I get. (with NINA build in focus routine, not used HOCUS-FOCUS yet). So I would say try it and I hope you get the same results as I do.
Thanks for the input! I think you are right. I plan to do just that: increase the backlash to around 120 as you say and maybe increase the exposure time. I hope to get the scope outside tonight and do some testing. Who knows, maybe even some imaging!
@@Aero19612 I hope it worked.
It appears to have worked! I increased the backlash to 150 from 40 and increased the exposure time from 4 sec to 8 sec. I did not change the autostretch or black-point clipping from earlier settings. All of the curves acquired over night look like hyperbolas. Will try again as long as I have clear skies then summarize findings/observations in a follow-up vid.
Hi James, lots of good info here. you are (fast) becoming an Astro SME. The HFR and HFD provide an indication of best focus, so why not scale each column to be a maximum of 10 e.g. NINA column has values from 10 down to 5.70, PixInsight has values from 10 down to 5.48 and for the other columns? Plotting all these values will show how the shape varies with each algorithm.
All Focusers have some backlash and some quirks too. My Astro Pegasus's Achilles Heal is definitely the coupling between Focuser and Motor. Get that secure and many anomalies disappear. On Refractor, If the coupling is OK then manually trying to gently move the focus is prevented. On SCT the same type of Motor with a different bracket isn't an issue. Assuming all is coupled correctly the backlash value can be set higher - yours can easily be set to at least double your step size?
The question of mirror flop can be checked by either testing at higher/lower Azimuths or at least raising the scope up to the target before focusing? With Autofocus, you start high and come down to get points for a curve, evaluate the curve and set the focus to the calculated best focus. Returning to the best focus point should be easy even allowing for backlash e.g. over compensating. In NINA can you trace the positions sent to the Focus Motor to check how it is being driven?
Backlash compensation works differently in different Apps. In my Focuser the backlash is removed differently to SGP Pro's method and probably NINA's. Having more than one method in operation would be confusing at best. It would be tempting for the Motor manufacturer to hide their sloppy mechanisms with built in backlash compensation but that would defeat the controlling Apps like NINA trying to apply their own algorithm.
I noticed that the order of taking the images is not in sequence (based on the timestamps). Can NINA save the images during the Autofocus routine or at least repeat them in the same order? To replicate NINA manually you want a series from right to left and then the final image which should be the same/similar to an earlier point if everything worked out OK. Simon
Hey Simon! That's a good idea. The magnitudes of the various metrics are all over the map. At the end of the day, we would each metric to provide the same optimum focus point. Of course, they don't. In this example, APT's HFD is best, APT's FWHM is next, NINA HFR/APT InvPow/Pixinsight FWHM are equally poor, and ASTAP is worst of all. Just one example though.
I agree. Backlash in the focuser and, for SCTs, in the mirror moving mechanism may be the biggest problem. If I had it to do all over again, I'd buy an SCT with mirror locks and a focuser that moves the camera instead of the mirror. For now, I'll try to get this mess working "well enough." First thing I'm going to do is dramatically increase the backlash value I've been using. Stay tuned!
Correct. The images are not in time/position sequence. I wish there were a setting that would allow NINA to save each picture during an autofocus, but there isn't as far as I know. That would be a good feature for geeks like me! Thanks for watching!
Hi James, i had the same problem with my RC6" and NINA. The curve was much better after a good collimation. You have excludes this option with you tests. With good collimation, i mean all stars in the your Field of View, not only the center star. A SCT Telescope is a little bit limited because you can't collimate the primary mirror but with my RC telescope and full collimation i got nearly good/useable hyperbolic curve.
Regards
Andreas
Thanks for watching, Andreas! My problem turned out to be excessive backlash in the focuser (see the video I posted about 3-4 weeks after this one). I’ll keep my fingers crossed.
@@Aero19612 Yes i saw the video. Very informative, keep going.
Regards
Andreas
Try an optec leo when your budget provides. Then you can lock down the primary mirror at rough focus and the leo moves the imaging train, so no errors introduced by mirror issues (assuming not too much flop). The leo is really great hardware, you'll appreciate the engineering and build quality of optec products.
Thanks, Jason! I agree, it’s much better to lock the mirror and move the camera. I’ve made some changes to my process with my Celestron move-the-mirror focuser by increasing the backlash setting by a factor of 3. That seems to have really helped as I am now getting consistently good curves. We’ll see how long this lasts!
@@Aero19612 Good to know about setting the Celestron motor focuser to 3x backlash. I've got the celestron motor focuser too, and still use it when the HyperStar is on (because mirror focus is then the only option). So, I've been manually focusing with bahtinov mask when using HyperStar due to this same issue, so I'll have to try the 3x comp and see if that works at f/1.9.
As an aside, maybe in the future you can do some episodes about the HyperStar, if you get one and want to go down that rabbit hole.
If you can handle the weight, I recommend the the lite crawler, with its motorized rotator is a joy.
Thanks, Dennis! I’ve looked at it. It’s a very nice focuser and the most trouble-free way to go with SCTs. I cranked the backlash waaay up and made a few other adjustments and everything seems to be working well (for) now. Fingers crossed! Thanks for watching!
I get seem to get too much backlash on my SW APO to achieve successful autofocus in ShapCap (Pegasus focuser), although I haven't tried generating a V curve in NINA. Maybe I'll get better luck if I give that a try. Thanks for this video and excellent explanation!
I love my C8 but I don't know if I have patience to collomate it in the cold. Then again, I'd use the same excuse to avoid using it on a mild and humid summer evening. Well, priority one... I'll give NINA autofocus a chance.
As a general rule, I've had great results with the NINA autofocus routine. Give it a try! I'm really good at generating excuses when it's "too cold" and "too hot." That leaves about 10 days for imaging. Thanks for watching!
Hello James,
I am actually quite a large fan of your videos and I believe have watched each of them several times due to the joy of learning many things about the hobby from you. But, in a interesting change of events, I might can actually help you.
As not only a NINA and Hocus Focus user myself, but also an Edge HD 9.25 user as well, I can understand the pains of trying to achieve focus in an form of automated fashion. I have toiled for 20-30 hours testing various means of Autofocus on my edge. The problem is one of design though. The graphs you posted are almost identical to ones I saw when using a ZWO EAF on the main focus knob. The issues is mechanical. The way Celestron designed the SCT's focus system seems shortsighted at best, or just cheeping out at the worst. You see, the primary mirror has 3 points of contact on the inside of the scope. Two heavily greased guides, and one guide that is used for moving the mirror. The problem, as I am sure you are quite aware of now, is that that mirror can skew, changing the stars, but also creates an impressive amount of backlash. In colder temps, that grease locks up and makes focusing even more difficult. I believe people with good copies of the scope can use an EAF potentially, but it is far from reliable and do not negate the issues of the dreaded 'mirror flop'.
There is a solution. Prima Luce Lab makes the Esatto which is a dedicated focuser for many setups. It has zero backlash (each one is calibrated at the factory and mine has zero measurable backlash) and solved the mirror flop issue. Basically, you just setup the Esatto and all your backspacing to be about 5 mm short of the recommended backspacing for your scope, the as the Esatto moves out the remaining 5mm, it will achieve focus at the same point where the recommended backfocus is.
This focuser solves several issues. First, focus curves are just like a refractor, with just a smaller HFR variance. Which my GT81, prime focus gets an HFR for 2.2 but my autofocus routine can measure out to an HFR of almost 20. The problem with SCT's is the prime focus has an already larger HFR of around 4, as you mentioned, so when you defocus the star, it becomes so large that NINA and the other detectors do not think it's a star. With my 9.25, I can get about 3-4x ratio for the extreme points of the autofocus curve which always seems to work fine for me. The next issue it solves is mirror flop. Since you are focusing with the Esatto, you can move the primary focuser so the lead screw is engaged and pushing the mirror again against gravity and lock the mirror locks. This means the mirror does not move (meridian flips do still need an AF run) and it reduces the mirror flop issues. Lastly, AF speed is increased with the Esatto and Nina. Running Hocus Focus and Filter Offsets, I can do an Autofocus routine in less than 2 minutes. Oh, and the filter offsets are a great addition too as since the Esatto moves so reliably you only have to use Lum as a primary filter and then just offset from there. Lastly, since the primary mirror is disengaged for focus, the target does not move as you focus. For planets this is almost a must as you can focus without ever losing the target. For DSO, you do not have to worry about if autofocus is going to move things such that you have to reslew to the target.
I'd be more than happy to help here because I have used my Edge on several deep sky projects and have gotten over 100 hours of imaging time on it in the past few months so I understand the system very well. The Essato is a prime piece of kit. Especially given is insanely small step size. It is price for sure, but it would solve every issue you have in this video and more.
Thanks for watching all these stupid videos, Noah! It's a tough job, but someone has to do it. Actually, you don't "have to" so I really appreciate the effort.
I totally agree-SCTs bring in a ton of mechanical baggage that good refractors do not. I also think Celestron was only thinking of visual use when they designed their base level SCTs. The EdgeHD versions are better suited to astrophotography, but still have extra mechanical bits that refractors don't have. Thank you for the Esatto recommendation! You make a very convincing case and a new focuser may well be the path forward. I don't have much confidence in this Celestron focuser. It has cost me many nights of good imaging. At some point, I'm going to get totally fed up with it. Off to Google the Esatto...
@@Aero19612 The reason I enjoy the videos is due to the technical nature of them. Not many Astrophotography people put in the level of detail you put in, and I mean that. I take a very academic approach to astro and watch a large amount of videos on this hobby to study and take notes.
But that is a wise about about the visual use. All, and I do mean all, of the issues we face with the C series and Edge series would be solved if Celestron would have just locked the primary and put a crayford or some measure of external focuser on the back of the OTA.
I do fully believe that the Esatto solves all the issues you face. You should just need two adapters for it and then no more fussing around!
Indeed it would. Just looked it up. The Esatto 2" with the standard SC adapter and the M56/M48 adapter would work. They're not in stock at the moment. Surprise, surprise.
Good video James, you have very variable inconsistent results , did you run while you have cloud running through as have many outlines, so if you have a measured backlash of what you say is 20 then it can't be that.. could it be mechanical issues from the focuser ( you've had plenty of issues with it so far) So the EAF is working perfectly fine on the frac
Your collimation doughnut looks ok, you can overlay collimation aid in APT to do the same thing.. but remember your collimation doughnut isn't in the same place on both sides of focus... and as have mentioned before you want to approach focus from one side not reverse and come back as you have a slight backlash error from the mirror drawtube, you need gravity as your friend...
Thanks Derek. Even though I measured the backlash when I first installed the focuser, I have had to remove and re-install it several times. It is possible that the interface between the focuser and the SCT is not as tight as it was originally. Anyway, I'll try longer exposures and a higher backlash setting. If all else fails. I'll go a different route with another focuser.
it's not mirror tilt, I have a secondary mirror focuser I added to my stack - same problems. Algorithms are the problem! Specifically it's the inner circle/astigmatisms of sorts.
I'm running an Edge 8 with the celestron focus motor. My backlash compensation is around 120. Step sizes are similar to yours.. around 30 IIRC. Mine has been working fairly well with SGP (which has a notorious AF routine). It struggles the most on nights with poor seeing.
Thanks for the information. Using a much larger backlash setting is certainly worth a try. Thanks!
Hi James. I have never observed focus curve asymmetry with NINA on my C9.25 with the same focuser. The NINA settings I use are nsteps=5, step size=28, overshoot method, 50 steps IN. I would be surprised if mirror flop gave the same consistent asymmetry at different locations in the sky. I would be equally surprised if focuser slip was this consistent. If anyone can get to the bottom of this, I am betting on you. Please let us know what you find. (PS taking multiple shots at each focused position might give you a better trend but the HFD will be higher as I am sure you are aware).
Hey Bob. Our numbers are pretty similar. Interesting how you're getting good results and I'm getting the results I'm seeing. That said, I do occasionally see a nice symmetric curve, but I get more of these asymmetric curves. I'll keep on digging, I guess.
What were your seeing conditions on the night you ran your tests? Was your scope completely thermally equilibrated? With the very fine pixel scale of the SCT, the stars are extremely sensitive to seeing and tube currents. If you run an OAG this is immediately apparent as the guide star wobbles and stretches all over if the seeing is marginal. With my C11, on calm nights the focus curve runs perfectly, with iffy seeing it is quite a challenge. Additionally, the right side of your SCT focus curves appears "flat", which may indicate that all the backlash is not yet out of the system or that you are suffering from mirror tilt with the focuser movement. With overshoot compensation, there is no need to be exact and it may be better to increase the compensation to achieve the full 1/2 counterclockwise focuser turn needed to seat the mirror forward consistently. The EdgeHD series are superior in this respect as you can lock the mirror in place and use a secondary focuser such as the Optec Leo, thus avoiding this issue (also mitigates mirror flop as you meridian flip). I have found the Hocus Focus PSF fitting to yield slightly more robust star identification. Binning to 2x2 can reduce the vulnerability to seeing somewhat. Were your 5 sec measurements a single frame or the average of several frames? Taking multiple exposures at each focus step can help average out the seeing as well (at the tradeoff of being time consuming). The variability in HFR measurements can serve as a measure of seeing quality as can the HFR itself - with a theoretical 12.4 micron Airy disk at f/10, minimum HFRs of 6 in your test on 1/29 suggest marginal seeing, as does the variability from 4 in one run to 6 in another. If you do not have time to leave the scope out for 1-2 hours before imaging, or you have 40 deg F temp drops overnight like i do, insulating your scope with 2 layers of Reflectix may be the best option. Works well on my C11. Hope this is helpful, long focal length imaging is a very different game than short refractors!
I posted a follow-up video to this one showing that a significant increase in backlash solves the problem. I made some other, probably minor, changes as well. I will eventually try Hocus Focus-it looks very good. Thanks for the suggestions!
Hey James, nice video! I wrote Hocus Focus, though I haven't been able to test it much at high focal lengths since I got rid of my SCT some time ago. Would you mind helping me improve star detection for this use case? I believe this will come down to choosing the right detection parameters, and some sample data will really help with this.
Thanks!
Definitely worth trying Hocus Focus - and maybe it calls for a Hocus Focus SCT variant. I suffer similar issues and still don't know the cause, but out of focus stars look different on opposite sides of focus on an SCT (or at least on mine) so perhaps that contributes?
Hey George. I'm looking forward to giving Hocus Focus a try. I'd be happy to help. I'm socked in with a weather system for a week. Is there a way to use Hocus Focus in a stand-alone mode to assess the images I discuss in the video?
@@Aero19612 if the images are saved, you could use the file simulator to read them from a folder. It's a bit finnicky to do it this way as you need to load them in the same order an auto focus would. I'm planning to add a "Save" mode for AF, as well as a replay one that'll use the images and focus positions from a prior save and redo star detection and curve fitting with the current settings.
Feel free to ask me (jokogeo in the Discord channel) for help if you need!
Hi, I believe I am using a very similar configuration - C9.25 with Celestron SCT focus motor. FWIW I am not seeing these issues, the autofocus curves are looking much like the ones you are showing for your refractor. I did much less analysis than you did and changed only a few NINA parameters based on earlier experience with other programs. I select an ROI between 20%-80%, keep guiding on and have overshoot set to 100 "in". Not sure if any of that is causing the difference in the results.
Thanks for watching. When I leave guiding on, the torque generated by the focuser causes PHD2 to lose the star and then it starts the error beeping, so I stop guiding during focusing. One difference could be that I'm using an OAG so the image scale is much smaller than if using a guide scope. I will try a higher backlash setting and longer exposures.
Hi James . I didn't see you mention any relationship between the "good" focus curves and which direction the focus shaft was turning . I'm told I must always approach focus in a CCW direction due to the mechanics of the SCT . Also , your 9.25" is not an EdgeHD OTA is it , or is it ? If you read the Celestron EdgeHD White Paper on those OTA's you'll find a section dealing (top right page 7) with improved tolerances on the Draw Tube to reduce binding and freeplay when focusing . Maybe you have too much freeplay in the drawtube ? It would be nice to have an OTA to take apart just to see what is actually going on under the covers . Cheers./SRK
to compensate for any mirror slop, perhaps increasing the backlash compensator to say 100? BTW James did really replace a 0.75 pitch screw with a 0.50 pitch screw?
Jeremy: The screws in the C9.25 (may be different in other SCTs) are M3 and all M3 screws have 0.5 mm pitch. Agree: A consistent comment for folks getting good results is to use a much larger backlash setting like 100 or 120. I will definitely try that out if I ever see the sky again...
Hi Scott. I believe I'm always turning the shaft CW during focus, which means (I think) that the mirror is moving back toward the camera. I'm certainly not ruling out scope manufacturing tolerance as a contributor. Of course, I can't do anything about that. There have been several comments where people are using much larger backlash numbers than I am. I will try that and maybe winding the focuser/mirror way out and then back in also addresses some of the mechanical slop you're referring to. We shall see!
Ahhh Jermey, I see the problem. I fixed the "0.75 mm" pitch at one location but not the other. Sorry for the confusion. Good question!
Hi James,
I have an 8 inch Celestron SCT with a Celestron focus motor, and I seem to get better results when using overshoot on the IN side instead of the OUT.
Also when you get disk like stars Nina has a harder time getting the HFR, you might want to use a slightly smaller stepsize or less steps to make sure the stars don't turn into disks.
Anyway, hope you figure it out. I'll be interested in the results using Hocus-Focus.
Yeah, you're right about the in/out direction. I use OUT to minimize the focuser motion because of other issues keeping the darn focuser attached to the scope. I've played with various step sizes. Next time I get clear skies, I have a list of things to try, including Hocus Focus. Thanks for watching! Like the X-Files: The Truth is Out There
James, please tell us if what @Matt Nudi said in the comments, worked for you too. That thing with the auto stretch factor and black clipping settings in NINA, that improves autofocus in bad seeing conditions, that more or less we all have.
Right. NINA recommends using 0.1 for autostretch and -2 for black-point clipping for high light pollution. I’ll try those out.
I had the exact same problem on my 9.25, same shape in the AF curve as your data. I found that longer exposures were needed, and it cleaned up the curve, something about how the intensity changes as a function of defocus for that direction and how HFR is calculated? I spent 2 nights trying everything I could think of, and longer exposures is what did it. (I was using the hocus focus plugin for NINA with backlash overshoot)
I think this may be related to the comment later about stretch factor and black point - related to how it determines the threshold to detect stars - especially with the central obstruction creating the hole. I ended up with 8 s exposures, which made for a long AF run, but that was faster than failing AF runs.
Thanks Brent! I think that's a good idea. Next time out, I'll crank up the backlash setting and the exposure time to try and get more and better-formed stars. I'd be satisfied with a 10-minute-long autofocus if I could get reliable results. As it is, I'm losing nights of imaging. Thanks for watching!
Hi, you really can't rule out collimation because the more defocus you dial up the less sanative it becomes. I find 3 rings is good for the defocus method but you will achieve best collimation on a focused star. Also you need to be careful on the way focusing is done because of mirror slop etc.... Good luck with it.
Thanks for watching, Troy! I'm going to try longer exposures, a larger backlash setting, tighten the focuser/sct set screw, and Hocus Focus. We'll see....
Thank you very much for this video, James.
I have de C9.25 EdgeHD with a feather touch focuser and Starizona’s micro touch electronic focuser. Most of the time I use it at f2.3 with hyperstar, but sometimes also at f10. NINA’s auto focus routine has worked very well in both cases. Collimation certainly could be a culprit for the non symmetrical graph you’re having. I don’t know if also the Edge series have different mechanics for moving the primary mirror, since it has the locks. Something to figure out.
Could you make Hocus Focus plug-in work?
Regards,
Alfredo
Hi Alfredo! I installed Hocus Focus, but have not (had) to use it yet. We've got a full Moon, so this is a good time to play with it.
Once I got the scope back out, I increased the exposure time to 8 sec (was using about 4) for the Lum filter and increased the backlash by 3x - from 50 steps to 150 steps. I've got 3 nights under my belt and the curves look consistently good! I'll make a follow-up video on this maybe this week (lots of rain moving in...)
The ideal setup is the Edge with the mirror locked in place and a focuser that moves the camera. For now, I'm limping along with my Celestron focuser that moves the mirror. I think my much higher backlash setting is compensating for the mirror slop. Thanks for watching!
Thank you James. Looking forward to see the result.
@@Aero19612 follow up video please
Hi James, what is the electronic focuser brand?
If it is ZWO EAF suggesting watch the following video:
ua-cam.com/video/E3RDkjfmqWM/v-deo.html
@@mrkutube James is using the Celestron Focus Motor on his SCT @1:20 he shows the backlash chart - the chart states Celestron Focuser. He also mentions it in a previous video ua-cam.com/video/qXcZ3ZMMMco/v-deo.html
Thanks Simon, I missed that.
Hi James, is the Problem solve. What was the reason? I have the same Problem with my myfocuser2 and apt. Do you have used a reducer? It possible that the reducer has an effect on the hfr value and on the autofocuser? Greetings from germany
Hi Franz. Yes, I think I have solved the problem. Check out this video:
ua-cam.com/video/M7GxDf2mcw0/v-deo.html
I'm using NINA, which has "overshoot" backlash compensation. It is possible to do this in APT as well. If you're not, check out this video at 20:40:
ua-cam.com/video/rT_cwkVMhno/v-deo.html
I'm not currently using the reducer, but I think that would actually help. Also, if you can, perform focus with the Lum filter (if you have a mono camera) in Bin1.
Let me know how things work out!
Do you have access to an ASIAir to try its focus routine? How do you get the Bahtinov focus error (pixels) data? I'm not tracking how you get the y axis data?
I don't have an ASIair, but I can and will test APT's autofocus routine. The Bahtinov focus error is reported in terms of pixel error. If you use Bahtinov Grabber, NINA's Bahtinov aid, or APT's Bahtinov tool, it will show you a focus error number. That's what I'm plotting.
I've been using NINA with my Edge8HD and getting perfect fits. However, when I tried with the Celestron focus motor, I got horrible results. Never was able get consistent results. I switched to a ZWO EAF and am having no issues. Nearly perfect every time. The fit parameter is .98 or above every time. Absolutely NOT a fan of the Celestron device. BTW, I am running with the .7 reducer so my FL is about 1400mm.
Great feedback, Steve! I'm glad there's a light (or should I say: hyperbola) at the end of the tunnel. The thought has occurred to me to replace this focuser with the ZWO. I'll dig a little more, then pull the trigger if I can't get what I'm looking for.
Hello James, I can second that. I never got the Celestron focuser for my C9 and went straight for the ZWO. First time I tried it, I had used the same stepsize as with my small refractor. No chance, but once I increased it to 100, the points are nicely following the curve.
Thanks for the feedback and thanks for watching! I'll give this Celestron focuser the rest of this galaxy season and if I can't get good results, I'll find another otion. I was close to buying the ZWO EAF some time back but held off.
Thank you James. I found your videos back last spring when I started with this hobby. I think Ed Ting made we want to buy a C9 but you gave me the confidence of actually getting one. Since (and it is probably true) a lot of people recommended starting with way smaller FL but I thought with support from videos like yours, I might get away with it.
You were also the reason I used APT back then and avoided the Celestron focuser. And I found your channel because you have a C9, but I stayed because I like your technical and analytical style. So once again, thank you James.
You could check the number of stars detected, you are all ready using just barely enough data to get any sort of statistical read - at either end of the bell I'm guessing the detections drop down. Because the stars are larger at a long FL they get fainter, and couple that with the bad morphology it probably can't pick out all the stars. You might try focusing on an open cluster or somewhere in the main galactic lane. You are also MUCH more at the mercy of seeing at 2k mm and the only solution there is taking more frames.
There are very few stars in this FoV. I think the real statistical benefit comes from generating multiple points (even if only with one star) at different focuser positions and then using the hyperbolic curve fit to deal with that measurement error. The "problem" with using multiple stars is that all of the stars will have the same poor shape if, say, a gust of wind affects the measurement. I agree, getting the image parameters right to see "enough" stars while not imaging too long and permitting wind, mount, and seeing conditions to dominate. Thanks for watching, Michael!
Hey James - did you raise your issues on the NINA Discord server? There's an autofocus-support channel for just this sort of discussion. The core devs and the author of Hocus Focus are certainly around ...
Im not James, but I did just post a link to this video there.
Tom P: Yes, I did raise the issue on Discord. George (developer of the Hocus Focus plugin) suggested I try Hocus Focus. I had not heard of that plugin before George pointed it out. Thanks for watching!
Tom D: Thanks!
I thought HFR depends on seeing as well. For example, for f10 4.4+seeing. Am I correct?
HFR depends on just about everything. (1) Telescope aperture determines the lowest achievable HFR, (2) Seeing at zenith increases the HFR, (3) Air mass increases the HFR relative to zenith, (4) Wind increases the HFR, (5) Mount performance (uncorrected periodic error and gear noise) increases HFR mostly in the RA direction, (6) Imprecise focus increases HFR, and probably more things that just aren't coming to mind right now.
@@Aero19612 Thanks. This is what I thought. What FWHM (in arcsec) is acceptable for you, when you use your scopes? Are you coming TSP in May?
I just take what I can get. We have very little control over the FWHM - driven by seeing and target altitude. When I stack images, I will leave out images with "high" FWHM, but you can't go too crazy with that approach, or you'll never have enough data to create a final image. So, my approach is to collect 20-30 hours on a target, toss out maybe 10% to 15% of the images with the worst FWHM, then move on to a different target.
@@Aero19612 Ok. I reject anything above 3.5” for 4” refractor and 130mm Epsilon; anything above 3” for C9.25 and 140mm TEC. Every frame with eccentricity >0.5 also thrown away.
Rejection for AVX +ZWO focuser was 30-50% frames. For 10 micron + Moonlite focuser 10. Only for very bright targets. Blur and Noise XT totally changed game.