I was running a test of some G code with no material, no bit, and the router turned off. The maslow panicked and e-stopped, dropping a few inches on the frame and releasing extra belt from the bottom two spools.
How, from this starting point, would I go about being able to use the machine without rerunning calibration?
I don’t see a path to removing the slack in my lower 2 belts without completely starting calibration from scratch, and I’d really like to not have to remove the Maslow from the frame and run calibration every time it has a failure.
I can’t take slack till I’ve extended the belts, I can’t extend them till I’ve retracted them, and there appears to be no other options for managing the tension of the belts after running calibration based on all of the obvious options giving me errors that suggest I have to run calibration from step 0 to do anything for this at all.
So whenever the maslow 4 gets to this state, you must power down. Then to get it off the frame again, you would start it up, hit the “RELEASE TENSION” button here:
You can then take it down, retract-all, extend-all, hang it again and you are back in business. calibration, once done will save into the maslow.yaml file (which you should save somewhere, don’t ask me how I know) which defines all the parameters the maslow needs. Once you have a successful calibration, you don’t need to run it again unless you move the anchors or switch to another frame, etc.
I’m not asking for help for how to get it off the frame and recalibrate. I’ve run calibration at least 10 times since assembly and have had to remove it from the frame each time. That’s easy, I even have a jig for mounting my sled to the center of my spoil board to make it super easy to do.
I also don’t need to give slack to get it off the frame, the belts are already slack and I mentioned that above. The belts being slack is the whole problem I’m trying to solve.
Tbh, running a full calibration is huge overkill for a failure like this and it should be considered unacceptable if it’s the only option for recovery.
I am trying to not have to run calibration just because the belts have slack after an e stop. I already ran calibration and it was accurate when I tested with a square and circle path in G code.
Bar has said that we should only have to recalibrate if we put it on a different frame and suggested that all we have to do to swap it between frames it has been calibrated on is swap the yaml file. If I cant recover from this state without being forced to run calibration from step 0, if we can’t apply slack/tension without running calibration from step 0, then those things are not true.
I’ll give it a go and see if it let’s me load G code files after that. I expected to be stuck having to run calibration fully if I started the process at all after having those buttons being still locked to only working when ran in order after successfully calibrating.
I just don’t see why I am unable to just tell it to apply tension when in this state and skip that part of the dance (taking it off the frame fully). Just makes me glad I never put the wing nuts on. Probably never will and will just sub in quick clamps to hold the anchors in place, if I’m liable to have to remove the sled from the frame with every failure or any time a belt goes slack.
Exactly right. I think as the machine moves (but at least before E-stop), the maslow FW should store current belt lengths in non-volatile storage, so it can know on reboot where it was. Then there would not be a need to take it off, retract/extend, put it back. This is something we have discussed on other answers and I think its “on the list”.
I’ll give it a go and see if it let’s me load G code files after that. I
expected to be stuck having to run calibration fully if I started the process
at all after having those buttons being still locked to only working when ran
in order after successfully calibrating.
Calibration is to figure out where the anchors are, once they are in the file
you should not have to run it again.
I just don’t see why I am unable to just tell it to apply tension when in this
state and skip that part of the dance (taking it off the frame fully). Just
makes me glad I never put the wing nuts on. Probably never will and will just
sub in quick clamps to hold the anchors in place, if I’m liable to have to
remove the sled from the frame with every failure or any time a belt goes
slack.
The problem is that when the machine is turned off, it doesn’t know exactly
where the belts are, so when you apply tension, it can’t be 100% sure what the
bent lengths are. That requires the retract/extend to figure out (it’s the
equivalent of homing a more traditional machine)
Thee is talk of adding a button saying ‘the machine didn’t move since being shut
down’, which should work for a software stop, but not for cutting power (when
you cut power the system doesn’t have time to save it’s current location, and it
can’t save it’s location continually or it would not be able to do anything
else)
