I've tried the adjuster plate flat on the camera tubs and tried adjusting the focus by loosening the grub crew that holds the lens slightly, moving the lens very carefully into focus and then tightening up again.  Went out of focus!!  Tried pulling and pushing the lens against the flexibility of the ABS spokes and I could move it in and out of focus.  So the focus adjustment needs to be extremely fine.  I think decent focus would only be obtained with considerable luck by manual adjustment. This is all pointing to the need for remote control

Probably the easiest way to do remote controlled focussing is with three instances of the Astroberry Focuser driver or maybe one driver modified for three focusers.  This would then use three A4988 stepper drivers and three 28BYJ-48 5v motors modified for bipolar connections.  Shafts from the stepper motors would then go up to the screw driven focussing.  Another possibility might be to use the Arduino Nano if I don't run out of pins.  Haven't worked out the control arrangements though.

I have springs coming tomorrow so I'll see what I can do with manual 3 point focussing and a spring on each screw.  Assuming the pack contains some springs of suitable size.

The springs have arrived and there is a size that will do nicely   With a fine and clear night forecast I may try that later.

Hmmm...  Impossible!!  I'm wasting my time trying to get focus and orthogonality on the lens manually - I've tried enough now and it just isn't practical.  So it has to be remote (or give up!).  I may think about it but doubt I shall do anything for a while - I'm fed up with it so time for a break, can't do with the hassle.  Time to stop!

I've got enough 28BYJ-48 stepper motors and A4988 driver modules for remote focussing - that isn't the problem!  It's the mounting hardware and control software that is more of a problem and, of course, the coupling from motor to adjuster screws.  Doubt I shall be doing anything hardware wise but may give it thought and maybe try some design work in between clearing up (no.1 priority ATM).

Idea - plate on top of camera sleeve that can be moved for offset.  Then the mechanism for focussing etc. could go on top (possibly).  The first photo attached shows the current manual screw and spring adjuster arrangement whilst the second shows that there would be room for stepper motors on the top of a sliding plate (for offset).  But how to arrange a screw drive from the motor is another matter   I'll have a play with SketchUp...

Here a screenshot of SketchUp showing what I have in mind.  The green part is the camera sleeve on top of which is a plate (orange) with oblong holes and screws to clamp it to the camera sleeve.  Then the lens holder with its spokes is attached to the plate with screws and springs.  Somehow the screws would be driven by the stepper motors.

I can think of two ways to drive the screws - a pair of spur gears, one on the screw and one on the motor shaft or a shaft from the screw downwards with the motors mounted below the camera.  I think I favour the first option.  As long as the top part doesn't touch the bolts holding the plate to the camera sleeve, it should be possible to mount motors gears and screws on the lens holder - yellow part shown above.

A little tricky to mount the motors.  Here is a screenshot of cross-section diagrams showing the gear on screw and then the stepper motor roughly placed.  It may have to be at a different angle in the horizontal plane to allow for the fixings to clear the gear on the screw.  Of course, once one motor position is determined the others are simple 120° around the lens.

I think I may have a workable design.  The small circle on the back of the motor shows the line-up of the shaft and a hole can be arranged in the plate to clear the end of the shaft.

This is the complete triple control.

And this is the resulting adjustment plate.  The motors are space 6mm from the plate.

And this is the intermediate plate that set the offset.  This shows the way up for printing - it will actually go the other way up.  With the threaded bosses downwards.  These take the adjusting screws and will be tapped.  The larger holes allow movement for offset adjustment.  This fastens onto the existing camera tube.

I think I've done it   I think I've designed suitable parts to hold the motors and the lens onto the camera   Oh - got the gears to do

Gears.

A couple of photos - the offset adjuster plate attached to the camera and then the motor plate sat on top of that, in the second.

The motor plate didn't work out so I've redesigned it and now printing it.  The motor mounting stopped the gear rotating!

Just printed the six gears   I think the design seems alright now.  Should have the mechanicals done soon.  Then I need to decide on which stepper drivers to use but it will probably be A4988.  After that I have the software code to sort out.  Nicest would be to program the RPi to drive the stepper motors but I have used simpler arrangements for focussing in the past just using an Arduino and a single control wire.  But no wires at all beats one wire

Well, it's assembled - rather tricky to get the screws in while compressing the springs.

Now I need to drive the motors.  I might lash up an Arduino for testing - but not tonight   It would be useful to provides two modes - all turned together for focussing and separately for differential focus at 120 degrees.  It seems quite possible that once set up the focus may not change.  If this were so I would not need to faff about with driver code.  For testing on the guttering I could run a USB cable to an Arduino at the ASC (if I can find a long enough one).  I might reduce the supply voltage for testing so that I can run the Arduino directly off the main supply (Arduino max is 12v and I normally run at 13.8v) though if I use USB that can provide the Arduino power.

For a simple setup I could use the ULN2003 drivers and run the motor power off the 5.1v supply for the RPi using unmodified 5v stepper motors.

Control could be :-
  0 = A
  1 = B
  2 = C
  3 = All

OR
  0 = All
  1 = A
  2 = B
  3 = C

Makes little difference which stepper driver to use, both cost about a pound (even cheaper from China).  Same with the 28BYJ-48 motors, only a pound each.  The A4988 drivers are easier to connect and program but then the motors need the bipolar mod (just careful use of drill).  This gives me an idea for the control.  Astroberry Board driver for the functions outlined above and the Astroberry Focuser for actual focussing.

The diagram below shows the basic connections for the A4988 driver module.  The Astroberry Focuser driver uses SLEEP, STEP and DIR.  These could be connect to all three driver modules and the ENABLE used to determine which motor(s) to drive.  I'll sort out the logic for driver selection shortly.

Looking now at what control lines I shall need from the RPi if I go for full remote control - thinking it might be best to go straight for it.  I suspect that if I go for just set it up and hope it doesn't change, I'll regret it.

Basically :-

1. Lens Cover - OPEN/CLOSE
2. Dew Heater ON/OFF
3. Camera Cooler - ON/OFF
4. Select Motor A
5. Select Motor B
6. Select Motor C

Hmmm...  Astroberry Board has 4 controls.  OK so 4, 5, & 6. only need 2 bits but overall there are still 5 control bits required.  So this will need either two Astroberry Board drivers or one with the code modified.  This will need looking into.  In fact there are several of my imaging systems that could do with modified or new INDI drivers and I could do with conquering this.  (If I can get my aging head round it!).

The Astroberry Project has be superseded by PiFace and the PiFace Board has 8 relays so the INDI driver for this could be used as is.  Though best to change the labels.