Backlash compensation: where can I find info?


#1

Where can I find an explanation of what “backlash compensation” does in autofocus and how to properly set it up? Out or in, what is the best method to determine the step size?

There is nothing about that feature in the SGP manual.

Thanks


#2

Sephan, it is quite easy. There are two controls, the direction and the amount. The instructions are in the help file, under control panel/equipment-focusers.
The direction is the direction of the final move. So ‘IN’ means when the focus position has to move out, it goes out a bit further by the backlash amount and then moves IN by the same amount. The amount is set by the steps.
It doesn’t matter if you slightly overdo it, so long as the backlash step count exceeds the actual backlash. For my Lakeside units, with 5 micron steps, I do about 100 steps, or 0.5mm.


#3

Thanks Buzz,
It is still not clear to me.
I understand the compensation and the overshoot to take the backlash out. What I don’t understand is the direction that we need to select. There will be backlash after the focuser turns in any direction (backlash will need to be removed in the opposite direction).
If the focuser goes “out” the backlash that needs to be removed will be “IN”. If the focuser goes “in” the backlash that will need to be taken care of will be “OUT”.
As such, why are we specifying a direction?

Thanks


#4

In the case of backlash control, it has to overcome a deadband. The deadband comes into play when the focuser changes direction. During the focus routine, it typically slowly moves the focuser inwards, creating the V curve and then, having calculated the right position, moves back out to that point. The problem is, with backlash, it moves to the incorrect physical focuser position, though the stepper motor will be right. You effectively remove this deadband by always making the final move in one direction only, typically inwards, against gravity.
Let’s say the focuser has moved from 1000 to 800 during the V-curve and needs to move back out to 900 for the right position. If the backlash is 50 units, it will only get to 850. The trick is to move it further out, by more than 50 units and then return by the same amount, ensuring the mechanism is engaged. In this case, if the backlash setting is more than 50 units, say 70, it will guarantee that the final move of 70 units inwards, actually moves the focus tube. For refractor units, most of us would select IN, as the final movement, as it is against gravity and ensures the mechanism is fully engaged. SCT users, moving mirrors, may do something different.


#5

Thank you for this. The “IN” settings for backlash compensation makes sense now.

What would be the best way to determine the amount of backlash of a system?


#6

Eyeballs! I put the scope on the desk next to me and did manual moves. I would move out say 500 units and see if I could see the scope reverse direction with a move of 50, 100, 150 etc. Each time you try it, you have to move outwards first. In a sense, you cannot overdo it. In my own case, since I also use TSX for modeling (TSX does not appear to have a backlash control for the focuser), I only use the focus module’s firmware to do the backlash compensation. Since I share a module between all my scopes, I set it to be just enough to cope with the worse case, about 100 units.


#7

That’s what I did. I thought maybe there is a more scientific approach.
Thanks


www.mainsequencesoftware.com