I just upgraded M2 setup into M4 and trying to calibrate. Because the frame I have was built for M2 , it doesn’t follow the suggested default symmetrical box of M4.
For better precision of M2 is was suggested to have a wider and higher top beam, which I made having the room (the beam on the pic is already lowered from M2 setup, but the workpiece is still not in the center).
Looking at the yaml file, I thought I can set the anchor coordinates regardless of the shape they form. Thought once the machine learns the cables, I should be able to set the workspace with an offset, not necessarily in the center. Not sure if this expectation is wrong or something else, but the machine gets upset from the first move, reports the further travel is unsafe.
Looking at the yaml file, I thought I can set the anchor coordinates regardless of the shape they form. Thought once the machine learns the cables, I should be able to set the workspace with an offset, not necessarily in the center. Not sure if this expectation is wrong or something else, but the machine gets upset from the first move, reports the further travel is unsafe.
pasting those dimensions into http://lang.hm/maslow/maslow4_frame.html shows
that the frame is way too wide, moving the anchors in will give you more working
height.
the old maslow had wider restrictions on the angles of the chains than the new
one (note that the red/green/white areas are not hard limits, they are the
points where the arms hit the uprights and errors will start to creep in.
what is the exact error you are getting and where are you trying to operate.
Getting “MSG:ERR: Unable to move safely, stopping calibration”
if you are sure about your anchor locations, you don’t need to do calibration.
but if you are trying to do calibration, make sure it stays within the green
area of the graph (I don’t think the default calibration size will with your
frame)
if you are sure about your anchor locations, you don’t need to do
calibration.
to clarify, the only purpose of calibration is to find what your anchor
locations are. Once you have that, calibration doesn’t do anything other than to
calculate them again.
Ok, I reduced the frame dimensions and calibration did run, which is good.
9x9 grid ended up with 0.47 fitness, although interim it was 1.15, end result is “Fitness too low. Do not use these calibration values” message, which is not so good.
I see the interim calibration results and they vary quite a bit, I guess I need to run it once or twice more. I don’t mind the time as long as it refines the right coordinates.
Is the frame flexing as you do the calibration? I would suggest adding vertical
supports between the beams (you can setup a camera to watch the end of the beams
and then play it back on fast forward to see the movement more easily)
there is an interstatial .74.x release that includes a frame flex test.
This is probably good enough to use. You can set the threshold for what the machine considers “good enough” in settings and you should be good to go. My guess is that frame having unsupported ends is likely to have a little more flex than the default option, but it should be OK
I assume you mean, “you should be good to go back and calibrate again or recalculate that calibration using an external program.” The old “failed” calibration values will have been discarded.
I think that 0.47 is probably good enough to just move on and start making things. The best option would be to track down where the flex is coming from and make the frame more rigid, but sometimes its nice to have a win and just move on and come back to fix everything fully later
