Determining the value to put in for focuser backlash compensation


#1

Hi guys,

I use autofocus and am interested to try to improve my results. I have a C11 and microtouch focuser which moves the primary mirror as I’m sure most will know. Is there a way to determine the backlash of my focusing system so that I can put in a compensation step size value and should it be an ‘IN’ or ‘OUT’ value?

thanks for any thoughts on this.
Paul


#2

In terms of value, go high… even up to 5% of your range in steps is often OK. In terms of direction, choose whichever direction is opposite of the direction the final AF movement makes. You might need to physically observe the move to the final AF position and then view the effects of backlash comp with respect to that.

Somebody that uses a similar rig might have more specific advice.


#3

thanks Ken, I will experiment. If anyone else has worked this out for a C11
with a microtouch it would be great to know.

I’ll post what works for me.

I tend to get very one sided curves - most AF points on the right and smart
focus typically adds in another 4 or 5 points to my 7 or 9 starting points.
Doesn’t seem to matter how many points I set - I have tried 5, 7 and 9. I
always get a nice right hand side to the curve and seem to always need more
steps - but adding more steps starts the run a long way out of focus with
donut shaped stars. Changing step size doesn’t seem to help.

Anyway, even though I don’t get nice shaped AF graphs, autofocus does work

  • mainly by using the best estimate.

This year at least we have some clear sky so I can experiment…

Paul


#4

Hi Paul,
I also have a C11 (with Starizona Microtouch focuser on stock focus knob). Using the provided instructions I ended up with a 250 step size (Compensation = IN) which worked ok but not perfect. Based on (gratefully appreciated) feedback from others here, I have been increasing this to larger amounts (from 800 to currently 1200). This has improved things - I think the focus point ends up being quite good almost all of the time. I occasionally still get an out-of-line point at the far left or right of the graph, so probably something else going on there.

Here is an earlier thread with more info:

Overall, I’m impressed that it works this well, considering the mirror shift issue of an SCT.

DaveNL


#5

HI Dave - thank you for this, much appreciated. I also get the situation with 1st and second points being adrift. I would never have thought to put in such a high value - I guessed a value of 75 - but it is just a guess… the problem is as you know, no way to see which way the microtouch is turning the focus knob - it’s all hidden away which makes an empirical approach quite difficult. I will give this a try next clear night. At least we’re having some clear nights this year.

Best wishes
Paul


#6

https://starizona.com/acb/MicroTouch-Low-Profile-Stepper-Motor-P3360C745.aspx

The new low-profile motor retains the precision of the previous motor, including the 300 steps per turn of the fine-focus knob (3000 steps per turn of the coarse focus knob). …

On my CPC1100 (with a feathertouch focuser) it seems to take about 1/2 a turn (or maybe even a bit more) of the COURSE knob to remove the backlash and mirror flop when I’m using it visually. So based on that I would set your backlash somewhere around 1500 steps as a starting point.

Hope that helps.

Jared


#7

Great, thanks for the Starizona link Jared, I will try 1500 as a starter and see how that goes.

best wishes,
Paul


#8

I would actually watch the primary focuser knob if possible - and make sure the backlash compensation is at least 1/4 turn of the knob.

Also try to make sure that while the autofocus curve is taken and the focus count is going down numerically, the primary focus knob is turning counter clockwise.

For my EdgeHD11, the main “bowl” of the focus curve corresponds to about a 7-degree turn of the focus knob - so the backlash compensation needs to take at least 6 times the number of steps as the focus curve. So you may need a much larger amount of backlash than many people think.

And if the primary focus knob is turning clockwise instead of counter-clockwise, then the mirror may shift or slump a bit after each focus step - and the curve won’t be very well behaved.

This stuff is more complicated than it needs to be because the curve is always taken from large numbers to small - regardless of how “In” and “Out” are defined. But once you get it right, autofocusing an sct with the primary focus knob should work well.

Frank


Focuser Backlash compensation!
#9

I have 9.25 EdgeHD with feather touch and HSM autofocus motor. The HSM has both manual knob and electronic control so I can see which way the focus is moving. The standard SGP ‘OUT’ turns the feather touch counter-clockwise (i.e. lifts the mirror up). We need to reverse the focuser direction that in the equipment profile manager focuser options to “IN” so the mirror lifts against gravity during the auto-focus routine.
The backlash I have then set to ‘IN’ ‘1500’. In other words when ever the focuser finishes on ‘OUT’ i.e. dropping the mirror down, it will carry on for another 1,500 steps to run out any backlash and then reverse and apply 1,500 steps ‘IN’ to always finish a focus step by lifting the mirror.
The number of steps for the backlash can not be too big, but it can be too small. 1,500 should comfortably run out the backlash.
Dean


#10

interesting stuff and I will check it all out. hanks for all the responses.
Paul


#11

Just remember that “in” and “out” don’t mean anything during autofocus. The routine always moves from high numbers to low ones - regardless of what direction in and out mean. So for best results make sure the primary focus knob is turning counter clockwise as the focus numbers go from high to low.

If there is no way to make the focus knob turn counter clockwise as the numbers go from high to low, then you will be stuck taking the autofocus curve in a clockwise direction - which probably won’t work as well as the other way.

Frank


#12

Frank - I think Dean is referring to the IN/OUT setting in backlash compensation (compensation direction), which is quite important for autofocusing, or any focusing in SGP. Especially for an SCT scope.

This is the setting that can make sure the focus knob is “finishing” each focus (of an SCT) by pushing the mirror against gravity), as described by Dean above. For most SCT’s I think pushing the mirror against gravity requires turning the focus knob counter-clockwise (as you mention). For any given focus adjustment - if necessary - SGP will direct the focuser to first overshoot the focus point so it can “finish” by pushing the mirror upwards in the OTA. This happens for each focus point in the autofocus process). That’s my understanding at least.

DaveNL


#13

Hi-

Yes I know how autofocusing works with sct’s and I have been doing it for years with 3 different sct’s. In all those cases I make sure the autofocus curve is taken while pushing the mirror IN and turning the knob counterclockwise. Backlash compensation is invoked only once at the start of the autofocus procedure - and again at the end when it returns to the new focus position.

But SGP ignores IN and OUT while taking the autofocus curve. It always goes from high numbers to low - which could be either IN or OUT depending on how things are set up.

There are two scenarios to avoid - one of which is not ideal - and the other is very very bad.

The non-ideal one is where the autofocus curve is taken while turning the knob clockwise. If that happens the mirror won’t move in a well defined way and may slump after each step.

The very bad one is where the backlash compensation is set up to be invoked whenever the focus number goes from high to low. In that case backlash compensation will be invoked on every single step in the autofocus curve. That would be slow and terrible.

It’s not clear to me how many people are in either of the two cases above. For anyone focusing an sct with the primary focuser knob I would take steps to understand how it is moving during autofocus.

Frank


#14

Thank you both for clarification. I had not appreciated before that SGP always starts high on the focus steps and then proceeds to lower focus step values during the autofocus routine. IN/OUT is irrelevant ay that point.

My initial point about reversing the direction of the Feathertouch HSM was wrong. The standard movement on my system is that the motor turns counter-clockwise as the focus steps reduce in number (happy days). Had it not, then I would have reversed the direction using the focus-linx control software that acts as the ASCOM hub for my system.

Every thing else is exactly the same as you have also already said.

Dean


#15

Thanks to all for all these helpful comments. My SCT is a non edge c11, bought new circa 2013. It is a hyperstar compatible one. I bought a feathertouch focuser and Microtouch controller from Starizona a few years ago when I visited the U.S.

SGP is really useful and comprehensive software - I use it for all my imaging sessions, so trying to get the autofocus to work effectively with my kit is something I have wanted to do for a while.

As a result of the posts in this thread, I have now got to a definitive position which is completely reproducible - thanks to all who have chipped in.

I find these settings are good
Backlash compensation set to IN and value 800

In settings I use 7 points of 40 steps each.

I obtained reproducible curves on 10 attempts tonight, whereas in the past my curves were a random set of points which I always had to monitor.

curve attached:

thanks again,
Paul


#16

This is a great thread. I am currently trying to figure out how to do exactly this. I used a focus mask and a program called bahtinov grabber where it overlays the spikes and tells you if you are within the CFZ. Then I focused the primary mirror the other direction in steps of 100 until it started to move off of CFZ. That number is 2150 on my 11EdgeHD. I am using a HSM and Focuslynx which unfortunately has a hardware limit of 255 steps for backlash compensation. I see from a post above that SGP has the ability to set this in software so I will have to start working on getting it all set up. Maybe tonight if the skies stay clear.

I also recently purchased (enroute to me) the low profile motor form Starizona to replace my HSM (it hits on the camera). I think the step sizes are bigger with this new motor so maybe the size won’t be so big? I am not sure.

I am excited to get this all working. Frank you convinced me to go this route and I finally made the switch this week. I have had success with Focusmax and I have had the lowest FWHM I’ve ever had. Now I am excited to try SGP to focus and see how that goes.


#17

From the Starizona web site: “The new low-profile motor retains the precision of the previous motor, including the 300 steps per turn of the fine-focus knob (3000 steps per turn of the coarse focus knob).” So its step size is the same as your current motor.


#18

I have seen folks mention to use a Bahtinov mask and just to let you know, for me this isn’t necessary. I just take a 5 sec exposure and look at the stars. If they are sharp pinpoints, I know i can run SGP autofocus and it works.

If they are not sharp points, I pull up control panel and adjust focus by a ‘coarse’ amount and take another 5 sec exposure to see what change has happened. If worse, go the other way, if better perhaps try one more coarse. Takes me about 30 seconds to set up rough focus each time I start an evening’s observing session.

I hope this helps,
Paul


#19

Hey Frank, I also have an 1100 Edge HD with Microtouch. Any chance you can throw me your settings in SGP for the autofocus part of your setup? Would make figuring out this thing a little easier.


#20

Hi-

I am driving the primary focus knob directly with a robofocus motor and rigel controller - combined with my own custom gearing and belt. So my number for backlash won’t translate directly to yours.

But for SCT focus with the primary focus knob I recommend backlash compensation of 1/4 to 1/2 turn of the primary focus knob. If you can watch the knob as your focuser turns it you should be able to get the number of steps needed to make it turn that much - and it may be a very large number and take 30 seconds or so to wind out and wind back. But as long as the focus curves you get are repeatable - there should be no problem with such a large number - except that it will take time.

And of course make sure the autofocus curve is taken while turning the knob ccw - which means backlash compensation happens only when you need to move the focuser clockwise.

Frank


www.mainsequencesoftware.com