20 replies to this topic

[philsan300]

Two months ago, I changed telescopes from a C14 to a CDK12.5. I was using ASIAir but have switched to NINA and have really enjoyed the new scope with NINA. It sits on an AP1100. The camera train consists of:
• Hedrick Focuser->68mm wide extension tube->Wonderer Rotator -> Tilt Adapter -> ZWO OAG-L -> ZWO 7 2-inch filter wheel -> ASI6200mm
• I’ve used the spacers to get the correct back focus when the focuser is at ½ its range
• It all sits on a wheely bar that I roll out of my garage into the driveway.

Everything worked very well until last night. I could not get the autofocus routine in NINA (using NINA or Hocus Focus) to work.
The typical focus point is between 14600 and 14700. Below are screenshots of the settings I’ve used for the past two months; they have worked really well. The R-squared values are typically in the  .9 - .97 range.
I also included screenshots of last night’s failed autofocus runs.

Initially, I tried:
• Checking to ensure no wires were caught
• Changing the filters used to focus
• Switching the camera to an ASI2600

I then used the PWI autofocus, and it worked. I noticed the default autofocus step size was 500, so I used this with NINA and Hosuc Focus, and it worked a greater percentage of the time. However, it still failed more times than not (with both Hocus Focus and NINA AF). I did not try re-using PWI autofocus, so I'm not sure if the one time I tried it and it worked was a fluke. I'll retry it tonight.

I don’t know what could have changed from one night to another. All I do each night is roll out the wheely bar about 10 feet and connect power.
I’m running out of ideas, sorry for the long post.

 

Any help would really be appreciated.

Phil

Attached Thumbnails

• 
• 
• 
• 

[VMan]

You're out of my league with that setup. But I too have changed from ASIair to Nina and had to tweak my autofocus settings, on my more modest equipment... so I am curious what more experienced users will comment.

 

There is a setting in Hocus Focus that allows you to collect all the data of a run and rerun it afterwards with different parameters. That's really useful. My memory is poor and my astro computer is not on, but you should be able to find it. There was a star size threshold that I needed to tweak as it was miscalculating the out of focus size due to my particular aberration....

Of course you need to rule out there is not an equipment issue (I also had a temperature dependent slipping belt on my AF, no amount of software settings will fix that )

 

[Shiny]

Nina standard auto focus works great for me but Hocus Focus is completely unreliable and does really strange things. I gave up on Hocus Focus after a considerable amount of investigation and adjustments that didn’t improve anything. I’m using a Primelucelab Esatto focuser.

[WadeH237]

Have you visually checked to ensure that the focus motor is running as expected?

 

I have seen graphs like these with my own setup in 3 situations:

 

• The focuser is slipping.
• The focuser is in manual mode (my focus motor has a "manual" switch that disengages it).
• The cable between the controller and motor is faulty.

 

The third case took me a while to run down, because it would only fail intermittently.

[markalot63]

Hocus Focus doesn't change how the focuser moves, it just changes how the graphs are interpreted, plus options like throwing out bad measurements, etc.  That looks like either focuser slip or backlash not being compensated for.  

 

On a simple R&P focuser i usually set 400 backlash (in one direction, pick in or out but not both).  

 

Can you screenshot your focuser settings in NINA?

[philsan300]

Vman, thanks for the compliment. It was my retirement present and look a few years of savings to purchase.  I'll look at the Hocus Focus re-run options.

Shiny, I may end up coming to the same conclusion on Hocus Pocus.

Wade, the focuser slipping sounds like a good place to start. Do I just look at it visually to see if it slips, or do you know of a more accurate method to determine slippage.

Maraalot, I saw a good video put out by PatriotAstro: https://www.youtube....zvBlO44&t=661s  about Nina Autofocus. He talks a bit about setting the backlash in one directoin or the other, and that helped him. I'm going to play with those setting tonight.

 

Thanks guys, I really appreciate your thoughts and time. I'll post any progress.

Phil

[KGoodwin]

Those pictures look like a focuser that's slipping, but in that case I'm not sure how the PWI focus routine would've worked correctly.  The Headrick focuser is somewhat known for "disassembling itself" under certain conditions.  I'd definitely do a full mechanical investigation of it running it in and out with the hand paddle and checking to ensure it's moving how you'd expect it to.  I personally replaced mine with an Optec Gemini since I didn't think the Headrick was up to par with the rest of the setup.

[philsan300]

Thanks Kyle, the scope and focuser are 4 months old and so still under warranty. I reached out to Planewave support with a description of the problem and pictures to see what they think. 

Phil 

[WadeH237]

Wade, the focuser slipping sounds like a good place to start. Do I just look at it visually to see if it slips, or do you know of a more accurate method to determine slippage.

Here is how I test for slippage.  Actually, it's how I test the mechanical behavior of my focuser in general:

 

I get the system so that it is well focused, manually if necessary.  Once focused, I command the focus software to move a significant distance in either direction.  After the move is complete, I command to move the focuser the same number of counts in the opposite direction.  If the system is working properly, it should be in perfect focus again.

 

If it's not in focus, then there is some kind of problem.  While the movement is happening, I carefully watch and listen to the focus knob to see what it's doing.  If it moves smoothly and sounds normal, but doesn't return to focus, it's probably slipping.  If it has a manual mode that's engaged, you'll hear the motor run, but the focus knob won't move.

 

With the cable problem, I observed the motor moving in fits and starts during the test, but this was after months of occasional failures (like 1 in 5 or 6 full sessions).  Eventually the problem got bad enough that I could see it behave poorly in a test.  At that point, I made up a new cable, and the system has been running flawlessly for several years since.
 

Maraalot, I saw a good video put out by PatriotAstro: https://www.youtube....zvBlO44&t=661s  about Nina Autofocus. He talks a bit about setting the backlash in one directoin or the other, and that helped him. I'm going to play with those setting tonight.

If you have enough backlash, it could cause the same kind of graphs.

 

To set backlash, you need to know how your focuser works with relation to gravity.  For example, a rear drawtube is pulled farther from the scope, moving focus outward.  With an SCT, which moves the primary mirror to focus, gravity pulls the mirror towards the back of the scope, which moves focus inward.

 

Once you have established which direction gravity would move your focus, you want to set up backlash compensation so that when the focus motor moves the drawtube or mirror down, you want to overshoot the target focus position, and then return against gravity to the proper focus position.  It helps hugely to watch the focus motor when backlash compensation is active to make sure that everything is doing what you expect.

 

With regard to the amount of backlash compensation to use, more is better.  A compensation amount that is too high will always reach proper focus.  A compensation amount that is too low will result in potential backlash problems.  It might take some experimentation to find the "Goldilocks Zone".

[rgsalinger]

There's no backlash in that focuser but it can slip. It can't be slipping because then the PW focus routine wouldn't work either. 

 

My experience, though, is that the PWI autofocus routing succeeds with poorer star/shapes (and I think numbers of stars) than HocusFocus does. I've used both and that's something that I found to be correct when I replaced my Hedrick focuser and had to use HF.

 

So, my take on this is that I think you had a bad seeing night. That meant that one routine worked and the other sometimes failed. I would suggest using longer exposures with HF might be just what you need to do as you will certainly get more stars. You can also decrease the stepsize to avoid donut shaped stars which HF doesn't like.  

Edited by rgsalinger, 17 April 2025 - 07:29 PM.

[WadeH237]

So, my take on this is that I think you had a bad seeing night.

Hocus Focus includes tuning for star detection.  If it's having problems finding stars, you could make adjustments there.

 

For what it's worth, my experience with seeing is that I would give up on trying to image before it reached the point that it would fail as shown here.  That assumes that everything is set up properly.

 

Passing clouds, on the other hand, are another way that this might fail.  You can configure Hocus Focus to save images from a failed focus run, so that you can review them the next day.  That would prove or disprove seeing, clouds, etc. as the cause.

[philsan300]

Great news and thanks for all your help! This is what I did and am now getting .95 - .98 R^2  U shaped curves (see attached image). 

1- When Kyle mentioned that the Hedrick Focusers have a reputation for disassembling them selves, I took out my allen wrenches and made sure all the screws (there are a lot) were tightened down. I found 2 that were pretty loose. This lead to a big improvement but still not at .7 R^2

2- I used Wade's suggestion about adjusting the Backlash IN/OUT setting on Nina, using only the OUT and at 700 it started getting beautiful U shaped curves. 

 

Thanks everyone, it's great to have such a helpful community.

Phil

 

Attached Thumbnails

• 

[Greg M]

I would suggest experimenting with your step size and # of steps. It looks like you have it currently set to 4 steps with a step size around 750-800.  Someone above mentioned reducing your step size but I would go the other way as an experiment and change it to 3 steps with a step size of 1100-1200 and see if you don't get a steeper curve, faster.

Also, your backlash looks like it is slightly overshooting, that is why the curve is a little steeper between the first and second point.  Try changing the backlash to 500. If that angle between the first two focus points flattens out too much then try 600.

[rgsalinger]

Glad you solved it.

 

I am surprised that you would have had so much backlash in a new Hedrick focuser. I used mine for about 5 years and never entered anything in the backlash parameter in the PWI3 GUI. I was told by "Kevin" that it was unnecessary and never used it. 

 

I don't understand why one would want a steeper curve. I thought that the reason the displayed curve was shallow was simply because of the aspect ratio. Is there any source for this? I'd love to see if I could improve my focus in some way. 

 

I also found that with my NiteCrawler, I got better HF results (but not by much) using 5 steps instead of 4. I have to play around a bit if we get some clear skies with the idea of three steps. Never know when I've managed to achieve a false positive result!

[philsan300]

Ross and Greg,

Thanks for the suggestions, I'm always looking to improve and will give them both a try.

Ross, I was also surprised by the amount of backlash on the new focuser. 

 

Best regards,

Phil

[KGoodwin]

There's no backlash in that focuser but it can slip. It can't be slipping because then the PW focus routine wouldn't work either. 

 

My experience, though, is that the PWI autofocus routing succeeds with poorer star/shapes (and I think numbers of stars) than HocusFocus does. I've used both and that's something that I found to be correct when I replaced my Hedrick focuser and had to use HF.

 

So, my take on this is that I think you had a bad seeing night. That meant that one routine worked and the other sometimes failed. I would suggest using longer exposures with HF might be just what you need to do as you will certainly get more stars. You can also decrease the stepsize to avoid donut shaped stars which HF doesn't like.  

It would have to have been truly horrific seeing if his settings are appropriate.  I've successfully autofocused with my CDK12.5 with the curve looking perfectly normal even in 4 arcsec+ seeing conditions.  I agree with your suggestions as to how to make it more resilient if he hasn't already done those things.  You don't want to go far off into donut land and 6 sec exposures work well for me for autofocus.

[WadeH237]

I don't understand why one would want a steeper curve. I thought that the reason the displayed curve was shallow was simply because of the aspect ratio. Is there any source for this? I'd love to see if I could improve my focus in some way.

You are correct about this.

 

FocusMax, which was super popular before all-in-one imaging packages became the norm, was built on exactly that premise.  Instead of running a full v-curve each time you focus, you would build an idealized v-curve for your system up front.  At focus time, it would pick a single star and defocus it to a specific FWHM.  To actually set focus, it would use the stored v-curve.

[Arie]

It happened to me too.

I got very strange focus behavior. Just like that after a while.

My focuser is R&P. That does not slip.

When I held the manual focus knob, I could wiggle the focuser slghtly before the motor shaft blocked it. 
The little grub screw that rests on a small flat area on the focus shaft, had worked itself loose a bit. 
After tightening it again, all was ok again. Perfect graphs. R=1 consistently. 
 

[rgsalinger]

The Hedrick focuser is not an R&P or a Crayford. When correctly adjusted and, as I posted, I think that the design prevents backlash. If it doesn't that's a new fact to add to my knowledge base. The PWI3 GUI, though, does have a Backlash Compensation parameter though. 

[philsan300]

Hi Ross,

I initially tried using 100 in the "OUT:" backlash setting and saw an improvement. I kept upping it by 100, and each time it got better. By the time I got to 700 it was almost perfect. Clearly that setting makes a big difference (from less than .7 to almost 1 R^2) , so there must be some backlash at least with my Hedrick.

 

Phil

[rgsalinger]

Well, I guess so. They do have a parameter in the PWI3 software for backlash. I asked about it in 2016 when I got my 12.5CDK and was told not to bother with tinkering with it. So, I didn't.

 

At the time the scope was next to a 24" RC and so I could benchmark the 12.5 against the 24 when auto-focusing. 

 

They both ended with roughly the same FWHM. However, I was using FocusMax4 on the RC.

 

Anyway, success is success.