Re: Anyone designed a limit switch or similar to prvent pier col...


Mark Squicquero
 

Joe,

Interesting analysis of the 3D nature of the problem. I hadn't
thought of it in those terms. I keep thinking of my 1200 GTO as a
telescope mount when it fact it is a servo controlled robot that is
carrying a couple of scopes, with all the inherent implications and
capabilities. Hmmmm, if we can just get it to transform into .....

Mark



--- In ap-gto@yahoogroups.com, "Joseph Zeglinski" <J.Zeglinski@...>
wrote:

Mark,

I was thinking of scenarios where you could put the OTA into an
unintended
collision, "while" in the process of backing up.

Lets say the scope had no DEC motion, and the scope was
approaching the
pier, at tracking rate. If you followed you original thinking, and
reversed
both motors, then the OTA might have been fine, except that it is
now
mispositioned during a reverse, where it might now hit the OTA,
from a
different direction. Now the OTA does strike the pier, because of
the DEC
adjustment. Or, if it did hit, during this second unnecessary
swing, the
circuit now reverses the reversal.

Of course, I may have misinterpreted your original intention,
of collision
avoidance during tracking, ONLY, and inaccurately extended the
discussion to
dual motor reversal. At the very least, even quickly stopping all
motion, with
the pier switch approach, will be a large success. Besides, if the
OTA were to
hit the pier during slew, I suspect the momentum would be too great
to avoid a
minor dent, with the circuit. I'm not sure how quickly it could be
emergency
stopped, during a fast slew.

The really bad news is that a pier based pneumatic switch
doesn't cover
all possibilities. When I took my new AP900 for it's first ride in
the rec
room, I let carefully slewed the scope to see how far it would go
before it
hit RA limits. Well, since it was also tracking when I approached
the far RA
extent, the OTA got a scratch when it touched the AP's Pier Adapter
azimuth
adjust knob - no matter how carefully meticulous I was trying to
be. In this
case, the OTA was nowhere near the tripod (pier), but it still
would have
gotten a dent, rather than just a minor scratch, from the "pier
adapter", on
top of the tripod. If the tracking mount strikes the mount parts
itself - gear
box motor housing, or base fork parts, the protection we have been
discussing
will fail - not even a loosely engaged clutch will prevent some
damage from
occurring.

... Then again ... wouldn't it be nice ...

It would be nice to have a software application where you
could "train it"
to identify the free space of OTA operation - train it the way they
train a
robotic arm to work on an assembly line. I believe the trainer
swings the arm
around (manually) in all "task" directions to map out the free
zone. An
example would be the paint sprayer robot on an automotive assembly
line - a
human programmer "shows" the robot where to spray, by literally
taking it by
the hand during it's training session. Then a collision avoidance
profile, or
map, is vectorized and extrapolated from the manually tested
positions. In
operation, after human training, as other applications command the
robot to
slew to a target position, the servo driver software constantly
checks that
the encoders are showing the robot is still within the "safe
envelope", mapped
for the job. This is a 3-D profile, just as would be required for
the
telescope mount. In this case, the ASCOM driver used by the
application (The
Sky, etc.), would make sure that the OTA is within it's trained 3D
envelope.
This would be done for each OTA used on the mount, with perhaps
a "safe band"
added, to allow your repositioning of the OTA to balance
accessories at some
later date, after training.

The good news is that something basically similar already done
in 2D - the
CP3 program allows you to define a "horizon limit" below which the
scope will
not GOTO. Now, second stage, create a piece of software to gather
AZ-ALT
coordinates as you slew the scope (during daylight) up and down
over tree
tops, buildings, and other obstructions on your 360 degree horizon,
and
"vectorize that path". This CP3 extension profile is still 2D, but
at least it
enhances the DEC positioning, to avoid "the trees", instead of
just "the
straight treeline".

Finally, to extend this to 3D (pier/mount avoidance), you would
swing the
OTA vertically (+/- 90 degrees), in Dec, at some increment of Hour
Angle,
indicating to the training program, how far the OTA will be allowed
to slew,
at that specific HA step position. You now have a series of safe
limit
coordinates (two Elevation limits at each HA), which can be joined
into a 3D
safe envelope, for future use. This envelope not only defines the
2D curve of
your visible horizon, it also defines the third level - of safe
limits of OTA
positioning. If you travel to a different site, the horizon curve
limiting
position can be reset to the old "treeline" method, as is done now
in CP3, but
the third level - collision avoidance curve will still be correct,
as long as
you are polar aligned - the same as when it was trained in daylight
back home.
Otherwise, if you go to a somewhat different site, at a
significantly
different latitude, the third level curve could be rotated by the
delta in
geographic latitude, and would still be sufficient to prevent the
OTA from
striking the mount or tripod.

Just a thought ...

Good luck with the basic system.

Joe

Join main@ap-gto.groups.io to automatically receive all group messages.