Measuring Focuser Backlash and Filter Offsets with a Bahtinov Mask

[@MichaelBorland1]

What really amazes me is that you have enough clear sky time to make such detailed measurements.

[@Aero19612]

Haha. This actually doesn't take that long. Maybe an hour. It's a good thing to do when you get a new focuser: get the backlash and fine steps size. You'll need that info when you set up your autofocus routine (see vid kzbin.info/www/bejne/aYqveGWHmd-aoZo). Thanks for watching, Michael!

[@montygiavelli8125]

Do you see a problem using an accurate dial indicator to measure the backlash ? 1. Move the focuser in until it's past the backlash. 2. Place the indicator up against the camera or filter wheel 3. Move the focuser out in small steps until you see a change in the dial indicator reading The number of steps used should be the backlash number. Thanks for all your videos, Monty

[@Aero19612]

No, not at all. Should work fine. Haha, I've also tried taping a piece of (uncooked) spaghetti to the 10x focus knob and watched the free end move as I sent focus commands. If you use the overshoot method for backlash compensation, you just need an approximate value then put in a large backlash to cover the error. Of course, the dial gage or spaghetti options don't work for an SCT since all the movement is inside the tube. Thanks for watching, Monty!

[@TheCassese]

James, outstanding tutorial. I am going to repeat your work with NINA +bathnov mask. Thank you

[@Aero19612]

Cool, Marcelo! Let me know how it turns out. Thanks for watching.

[@RocketSailing]

Focusing on the focus. Is focusing on another good thing. Merry xmas from Sweden!

[@Aero19612]

Indeed! Hope you had a great Christmas! Sweden sounds like an awesome place to be for Christmas. It was/is 21 C here. Very comfortable, but not very "Christmas-like".

[@patwicker1358]

I agree about getting the backlash as close as you can. On my CNC machine if I'm off by a thousandths it doesn't really show up in the first cut but if I'm cutting a circle it will come out as an ellipse because of the error accumulating.

[@Aero19612]

Hey Pat. A perfect example.

[@1CM69]

Hi James, I just rewatched this video as I was trying to understand a possible issue I had last night while attempting to set my Ha filter's focus offset. I did the whole B Mask procedure to find the number of steps difference to focus in Ha from my in focus position of my Lum filter & came up with a figure. I entered this figure at first in to the ASCOM driver of my filter wheel, made sure that APT had the "Adjust Focus" button enabled & proceeded to start an imaging plan initially using my Lum filter. However, when the plan called for a filter change to my Ha filter, the displayed focus readout position in APT did not increment at all and when the image had been taken, the filename & fits header just showed the value of the focus position of my Lum filter. I then removed the offset from the ASCOM driver & entered it directly in to APT's settings but the setting would not save, everytime I re-opened the settings dialog it displayed an empty offset table. So, to cut a long story short, do you see focus readout change automatically in APT & does this show correctly in the FITs header etc.. when moving between filters that have a focus offset setup? Regards.., Kirk

[@Aero19612]

Wow. I have not seen that behavior before. I enter my numbers in the APT settings area and it works fine. When the filter changes, the focuser offset is added/subtracted to/from the current focuser position. Conflicting information in the ASCOM driver and APT? In your case, though, the filter didn't change so the focuser shouldn't change either. Weird. Let me know what you find out.

[@1CM69]

@@Aero19612 Hi James, I've posted on the APT forum also, so hoping for some insight there. So, just to be absolutely clear, when you change a filter either through a set plan or manually choosing the filter to GOTO in the Filter Wheel GEAR tab, you can actually watch the Focus position readout increase or decrease? Also, I probably didn't make myself entirely clear in my initial post but through my testing I am changing from my Lum filter (referenced as 0, actual position being 26170) to my Ha filter (with set Offset of +65 steps) & although the filterwheel is actioning the filter change the focus position readout in the GEAR tab remains at 26170 & does not increase to 26235 as it should & the resulting Ha Light frame shows FP26170 in the filename & fits header. I'll definitely post back any further info/resolution that I get. Cheers.., Kirk

[@1CM69]

Quick update, Ivo from APT just replied to me that he has found a bug in APT where it no longer reads the offset values from the ASCOM driver, so I will be trying out the APT internal system again tonight.

[@Aero19612]

Ahhh! Very good. I don't use the ASCOM driver for that data entry, so that must be why my version works. Also, now that I'm using my SCT, I found that focuser to be so coarse that I don't use filter offsets for the Celestron focus motor.

[@1CM69]

@@Aero19612 Just to let you know that after removing my focus offsets from the ASCOM driver & entering them directly in to APT, everything worked like a charm last night. Regards Kirk

[@r.g.wilson4567]

This is a great video especially the part about determining backlash. Is the Bahtinov Grabber still available? Please post a link to it. Thanks again

[@Aero19612]

I imagine Bahtinov Grabber is out there somewhere. Try the Google thing. You can also do this with NINA. Thanks for watching!

[@miodragpetkovic6142]

Bravo and outstanding...

[@Aero19612]

Thanks! And thanks for watching, Miodrag!

[@louisrosner7902]

Another great video. You’ve done a lot of work on this backlash issue. I’m also using the Celestron focuser on my SCT. Is there anyway you could post some of your numbers for backlash compensation and fine and coarse settings for the Celestron focuser in APT. That would save a lot of time. Thank you

[@Aero19612]

Ha! Now, I see this comment/question. KZbin has been odd lately--I don't always get comment notifications. I answered the filter offset and backlash questions. The video I reference in the earlier reply also covers fine focus step size as well. That said, I have since had a chance use the focuser during imaging sessions and find that a fine step size of about 5 is OK. When I focus using a Bahtinov mask, I use a step size of 10. Once I get close, I drop down to 5. When I'm really close, I reduce to 3. Hope that helps! Sorry for the delay.

[@jpadron]

Hi James! Really good and informative topics. How do you measure the backslash?

[@Aero19612]

Thanks for watching, Jorge! I measure the backlash in focuser steps as the difference between the focuser position at ideal focus when moving the focuser inward and the focuser position at ideal focus when going the other way (moving the focuser outward). See 13:05 in the video. If there is no backlash (or the backlash is perfectly compensated), the blue lines and the red lines will overlap--something I'll check next time we have some good weather.

[@vaknins]

Hi James, thank you for this detailed video! I must have missed it, but what is the benefit of measuring using a Bahtinov mask vs measuring HFD or FWHM? Is it more accurate?

[@vaknins]

Never mind, you talk about it at 18:34. Thank you! Essentially, because the mask gives a signed error, while HFD and FWHM gives absolute error.

[@Aero19612]

Exactly. Thanks for watching!

[@panagiotisgalanopoulos5364]

Hi, I know that my question is stupid but how do you calculate the intercept and slope? Your videos for focus are just perfect and very informative.

[@Aero19612]

Hi Panagiotis. Excel has built-in functions for those called: INTERCEPT( ) and SLOPE( ). Thanks for watching!

[@panagiotisgalanopoulos5364]

@@Aero19612 thanks i just found everything. But I tried to test it with your results and no matter how many times I divide 192.7 to 0.039 the result is always 4941 and not 4986. Am I doing something wrong?

[@Aero19612]

You're doing fine. This is caused by "round-off error." The 0.039 number is rounded up, probably something like 0.0386 but is displayed as 0.039.