Hi All!
I have an EAF attached to a GSO Linear Bearing focuser (the one with the steel track) on a Gso Rc6.
I've always used L1P with great success.
Lately I have been looking into employing filter offsets, so, as a preparation for it, I wanted to run a test to make sure that my backlash is properly cleared. I keep driver backlash off, and use the default overscan of 5 x step size, which is 250 steps for me.
I ran 5 Autofocus runs back to back on the R filters, they all succeded with R2=0.98 and the same resulting HFR of 1.15-1.17 which is typical for my seeing and sampling. However, the final step positions seem to be drifting by about 20 steps per run:
Run 1 5208
Run 2 5185
Run 3 5165
Run 4 5164
Run 5 5145
This eaf is mounted so that higher step counts are outwards (0 is fully retracted).
What are these numbers telling at a glance is that a few inward steps are wasted in backlash (is this correct?)
What would be a solution for this? Setting the backlash at even more than 5 x step size?
The error is definitely less than the CFZ because the HFR is very consistent even among filters. However, I fear the drift makes filter offsets a recipe for disaster since they are blind by definition. However it would be nice to use them in order to interpolate filters.
Thanks for any suggestions!

Hi Marco,

Yes it looks like there is a drift in focus as you've observed that will need to be fixed before things like Filter Offsets will work.

It could be backlash or something loose or slipping somewhere (although it seems very regular).

The 5 * step size default no longer exists; you have to explicitly set a backlash value (maybe this is the issue?). The Focuser Backlash section covers the options...
docs.kde.org/trunk5/en/kstars/kstars/too....html#focus-backlash

If you are using AF Overscan for backlash, you could try adding 30 to the current Overscan setting and see how that works (since the drift looks to be 20).

HTH

Thank you! As soon as time and clouds allow, I'll do some testing and I will make sure to report!

Not sure if this will help but I recently setup my backlash to get my filter offsets reliable.

I have 2 scopes I set the EAF backlash with, an SCT and a Refractor.

The refractor was easy with its fine tune focus knob on the opposite side to the EAF.
I drew a straight line with a pen and ruler onto some masking tape, cut it in half and stuck it across the focus / fine focus knobs. This gives fine detail when focusing back and fourth to see if your returning to the exact starting point mechanically.
At this stage it was just a matter of trial and error, moving the focus in and out x steps and check how close the line's met up on the masking tape.
If it wasnt spot on id manually focus a tiny bit to get the lines straight again, write down the focus step number and move in and out once again.
Took about 15 mins but got it as good as possible, onee this was set the offsets and auto focus routine were working great.

As for the SCT it was a similar process but instead of the masking tape and pen I just used a Bahtinov mask as there is only 1 focus knob and the EAF is attached, unfortunately you have to do this method at night under the stars.
Get the focus spot on and then start racking back and fourth using different backlash settings with the EAF as above, manually returning the focus back to perfect after each fail and starting again until it worked.

Hopefully this is useful to you

This is great! Just so I make sure I understand: I have an RC with a refractor-style focuser with fine focus knob like you, so are you saying that you used the relative position of the coarse vs fine knob as a probe of the movement instead of using the drawtube movement?

Sorry its hard to explain exactly, I just took a quick photo to show you.
So these are the steps I preformed from memory.
1. Stick the tape on with the lines.
2. Move the focuser x steps in and then the same number back, watching the tape spin around and come back to center, it wont.
3. Step the EAF in or out to make the lines straight again.
4. Change the backlash number and try again.

Eventually you should get a backlash number which will cause those lines to stay almost perfect whenever you move back and fourth.
Once you think you have it correct, write down the focus position step eg. 15831 and rack it in and out many times randomly then enter the original 15831 and it should come back very very close to the straight lines.
For my Refractor the magic number is 46, if I go 47 or 45 its not as good. So find that magic number and write it down in 3 places!

Always a good idea to verify with a Bahtinov mask when its all configured.

Makes perfect sense, thank you!

Great idea @Max. Do you have some similar arrangement for measuring drawtube slippage as well?

This method worked perfectly and I have interesting findings to report: bench testing, the backlash is fully cleared at about 100 steps, so my overscan setting of 150 is enough to bring back the drawtube and the knobs at a perfect position every time. So, the next 2 candidates for the 20 step per run drift are:
1)Slippage. The bench test is without optical train and with the telescope flat on a table, so it doesn't test for train weight slippage. However, correct me if I'm wrong but a slippage of exactly 20 steps per run seems a bit too regular to me

2)Thermal. When I tested on stars the other day, the tube had been cooling for almost 2 hours, however the ambient temp was dropping about a degree C every half hour, I wonder if I might have been effectively sampling a shrinking OTA.

@Marco
Good to hear !

@AstroMuni
I haven't had to deal with this yet fortunately, my refractor is only a few months old and the focuser still feels tight.
Perhaps tightening very slightly the screws for the focus gear mesh adjustment ?
I dont have any experience with that at all so I cant really recommend anything.