Fell of the top of the board then belts snarled inside the spools

[wouldchuck]

Made a new frame, smaller than the others. roughly 4300 by 2300. May try to make it taller and less wide to get a better cutting area. Calibration took about an hour? and with a calibration grid set at 400 by 400 5x5 it finally worked but the calibration grid top row was about 10cm off of the top of the plywood. I saved it by holding it up with my hand. Is there a way to set the calibration grid to be centered on the original position? Here is the log in case that helps make sense. I tried to visually center it and then reset the xy home to that point. I had read in some of the forums that it might be possible to manually describe the frame size as an estimate in order to speed up calibration? I don’t see that on the config dialogue. Where would that be? Happy that i can finally drive my machine around using the buttons. Yay. Perhaps I’ll try some practice movements tomorrow.

Questions in order:

Machine fell off the top of the board, how to stop this?
Is it possible to speed up calibration?

Is it possible to center the grid on the starting position?
Where does one enter the estimated frame dimension?
What is a good test file to run without cutting?

How do people put a marker in the chuck for drawing or testing?

Thank you very much.

Maslow-serial(11).log (69.1 KB)

[wouldchuck]

New problem. I was driving it around the board for fun and when it got to the corner and the shortest belt had about four inches of slack. I tried to pull it out of the spool and things looked ok but then when I drove it back to the center the two middle belts became fouled on the spools.
Pictures and logs. No idea what is happening. Just got it working and was driving it around just a little bit.

Maslow-serial(12).log (70.9 KB)

[dlang]

wouldchuck wrote:

Made a new frame, smaller than the others. roughly 4300 by 2300. May try to
make it taller and less wide to get a better cutting area. Calibration took
about an hour? and with a calibration grid set at 400 by 400 5x5 it finally
worked but the calibration grid top row was about 10cm off of the top of the
plywood. I saved it by holding it up with my hand. Is there a way to set the
calibration grid to be centered on the original position? Here is the log in
case that helps make sense. I tried to visually center it and then reset the
xy home to that point. I had read in some of the forums that it might be
possible to manually describe the frame size as an estimate in order to speed
up calibration? I don¢t see that on the config dialogue. Where would that
be? Happy that i can finally drive my machine around using the buttons. Yay.
Perhaps I¢ll try some practice movements tomorrow.

the calibrate routine starts with all belts very close to equal lengths, it
currently need to calibrate around tht point. once it’s calibrated, you can
shift your workpiece as you want, but for the calibration, you need a surfact
the size of your calibration grid + ~150mm one each edge (+300mm total) to keep
the sled from falling off

Questions in order:

Machine fell off the top of the board, how to stop this?

see above (for now)

Is it possible to speed up calibration?

not at the moment [1], but you don’t have to do calibration unless you change
the frame

Is it possible to center the grid on the starting position?

no

Where does one enter the estimated frame dimension?

that is not needed/used any longer

What is a good test file to run without cutting?

just jog around with the arrows

what size is your work area? I have a 1’ and a 4x8’ test grid at
Index of /maslow you can use

How do people put a marker in the chuck for drawing or testing?

a stubby sharpie can just be jammed in there

Ian_a has a thread posted recently where he has a 3d printed part that holds a
mechanical pencil lead

other pen holders have been posted.

David Lang

[1] I am doing a lot of work in this area right now. Currently it’s too easy for
people to specify too small a grid or too large a grid (goes ‘out of the green’)
Bar and I have been working on separate projects to try and automate this.

I am also replying to this message while I wait for copilot on github to
implement my most recent change to a process that would virtually eliminate the
wait for it to do the calculation (where it shows the fitness scores) and I have
floated an approach that would handle non-standard layouts like yours (see
Discovery on moving calibration grid for
some of the discussion)

please follow along on this and when I have new versions to test, it would be
especially helpful for someone with an odd frame/placement to test it.

David Lang

[dlang]

wouldchuck wrote:

New problem. I was driving it around the board for fun and when it got to the corner and the shortest belt had about four inches of slack. I tried to pull it out of the spool and things looked ok but then when I drove it back to the center the two middle belts became fouled on the spools.
Pictures and logs. No idea what is happening. Just got it working and was driving it around just a little bit.

the fact that slack developed indicates that your calibrtion did not go well.

the good news is I do not see any sign from your pictures that the belt has gone
through the gears

what you are going to need to do is one of two things (neither is very easy)

the safer way is to disassemble the arms and get the belt straight and
reassemble the arms (if you do so, check how smoothly things move, lube them
while you are in there)

the ‘faster’ way is to take advantage that there are some loops in the belt (and
therefor a tiny bit of slack) move the slack around until you can get all the
belts aligned, even if loose. Then you should be able to run the motor.

there are commands that you can send (via the fluidnc tab where you can type in
gcode, via USB from a terminal program, or via telnet to the maslow

TLO will feed the TL belt out just a little bit (unless things go wrong and it
spools out fast, forcing you to pull the plug)
TLI will retract the belt a little bit
ALL is the same as the retract-all button and will spool in all the belts.

when you run the motors, you need to make sure that parts of the belt that are
sticking up do not go through the gears. This is what makes this way less safe.

When I went through theis a couple days ago for the first time, it took me and
hour to coax the belts into a clean setup, extend them all the way via TLO type
commands. the 10th or 12th time it happened to me this weekend (when you are
doing development work on the firmware, bad things can happen) it only took me
about 10 min for a worse snarl.

to keep your sanity as if you take the second approach, keep reminding yourself
that the other option is to tear down the machine to get the arms out then
disassemble the arms :-/

David Lang

[dlang]

one thing that I’m finding is that when the belts get in trouble like this
(spooled out a lot) pulling on them can be a really bad thing to do.

unless the belts are folded back on themselves, or so jammed up that they are
already out of line, just hitting retract all will probably solve the issue as
it pulls the belts in from the center, and by the time the outer layers start
moving, the other belts are out of their way and they can spool up correctly.

David Lang