Calibration Will Never Converge (Maslow 4.1 Firmware 1.05)

I have a issue with my new built Maslow 4.1 running 1.03 and then 1.05 calibrating to my new frame.
I first ran my system, after building, on a table that can do 2’ x 2’ boards
Worked fine
Cut some parts and they were ok/usable
I cut out the gussets for the wood 8’ x 10’ table from Maslow design
Stepping up to what i ultimately wanted to do 4’x8’ board as the 2’x2’ board/table was just test
I built a sweet table out of plywood 2"x4" board with a 4’x8’x.75" MDF sac board
All areas where a cut can be happening is supported every 2’x2’
Frame is about 200lbs
So shouldnt have any issue with the frame
Frame specs are 6’x10’ (changed design from 8’ to 6’ to make it take up a little bit less space, i still have 14" from the anchor points to any edge of table)
Very similar to Maslow wood design, but substantially beefed up
Running in Horizontal mode
1560mm on extension for belts
I first ran 1050mm x 2050mm 9x9 grid (Max system allows)
1200 on retract 900 on calibrate (i had issue with retract stopping early on some arms sometimes so bumped it to 1200 now dont have that issue)
Never converged after 6 hours or so running on a “high end” laptop
it would never get past the first 4 points
Never rise above .21 on calculation
no error would just keep trying forever
that was on firmware 1.03
Updated firmware 1.05 today
Copied all the settings from Maslow’s updated calibration video on the horizontal run
800 x 800 7x7 1200 1200 with my still 1560 extension
With 1560 extension i am at my anchor points long by about 3/8" if i pull out slack all to one side
In 5 hours or so never converged
Never made it past the first 4 points
Never made it over .21 on the calculations
Started over and did it on high end gaming desktop
Didnt converge in 3 hours
Never made it past the first 4 points
I never get a error it just calculates and retry’s new point forever
It never gets past about .21 in the calc
This is the 2nd or 3rd day i tried to calibrate to this table
Any suggestions
Thanks

Im no expert but can help with some low hanging fruit

Make sure the encoders are working - Retract, extend, and then retract again, you’ll get 4 values in the serial readout, these should all be close to zero if the encoders are working
Check the spools for knots or loops in the belts. Can extend the belts out to 3000+ to clear the loops and then retract again
Ensure Z axis is all the way down.
Ensure horizontal is selected in the config settings, not vertical.
Ensure Maslow is oriented in the correct direction, vacuum outlet should be pointed in the Negative Y axis direction/Pointed to a long side of cutting area, not short side
At the start of the calibration it tests for flex in the frame/anchor points and prints it to the serial readout, I imagine you really dont want these values to be greater than 1mm
Ensure the belt ends can swivel freely on the anchors and dont get jammed

Couple other niche things to check:
Personally with my setup, I couldn’t get a calibration to complete at all without increasing my calibration force, I got completed calibrations with 1300 but best results from 1800
I also had issues where when i was directly connecting to the maslow wifi it would disconnect mid calibration or go unresponsive which caused endless looping. I had to move my wifi modem to be in range of the maslow so i could connect both the maslow and my laptop to the wifi and communicate like that which helped.

2 Likes

Thanks for the reply Deadset
So here are some of my answers before some of todays trials
i believe when i retract i get like a .011 on like 2 of the axis i will check it today
Belts look tight and clean spooled up so don’t think there are knots but i will check
However i do see one belt starting to have issues
One of the steel wire has popped through the opposite side of the belt and twisted a little bit
I believe that is the top axis where on a 2x2 bed during one of the calibrations it had to much excess and got sucked underneath the axis and rapped around the plastic axis til it got stuck
put a nice big cut into the plastic with the belt.
At some point going to need to swap out that belt
That may be causing issues now
For the axis i had the cord facing twords one of the long sides (Y) TL and TR attaching to one end of the Y and BL BR facing the other end of the Y
However looking at recent Maslow walkthrough for 1.05 it is setup where cable is facing twords a X (short side) so i may try that
I will run a calibration again today and record the frame flex
and what i get from the retracting all axis

I will try 1300 and see what i get
I am connected to router then i connect to the IP address
Let me try cable/dust collection where i had before facing a long side and 1300
I will record the data of retract and frame flex
Then will see if it converges
Will give that a few hours
Then if that doesn’t work try 1300 with cable facing a X
Give that a few hours if it doesn’t work i will try 1800 with cable on the Y and 1800 with cable on the X
Thanks will be back with some results

Allright so i have some updates
When i retract after full 1560mm extension i get
TR -0.011 BL 0.247 TL -0.011 BR 0.011
I then trimmed the cable ends a little bit
I then get
TR -0.011 BR 0.097 TL -0.021 and BL 0.097
I keep getting 12mm center point deviation but i don’t have a valid calibration
frame flex on 1300 for calibration is TLBR 2.217 TRBL 2.097
Never converged on this one again
Ran 1800 on calibration tension
stiffness of frame measurement TLBR 3.480 TRBL 0.531
Also did not converge but mabye it was because my anchor points were a little loose when i loosened them up
Anyway i ran another 1800 but i rotated the cable so it is facing the short side and it calibrated in like 20 minutes no restarts.
Mabye one on like the 4th round it had 1 restart
First fitness number was like 1.2 so it just tried a few things then moved on
I redid calibration again except the 800 x 800 grid i did my full 1050 x 2050 gird and instead of 7x7 i did the full 9 x 9
Interesting enough the machine just did all the points then did a final calculation at the end but went through again very quick with 1 retry so interesting that it was able to do it so well by just rotating
However i noticed the next issue which is now why i see the frame is suggested to be 8 x 10
as with my 6 x 10 the angles for the motors are almost 180 degrees in order to max out my Y axis or X axis if i rotate my machine back 90 degrees again
I cant quite tell if they are touching the Z axis bearing holders but if they are then i cant go up and down in the Z at the extremes of my Y axis so i probubly will have to offset my axis outside my frame
If you are to build a frame either make it 8x10 or 7 x 10 to avoid this issue
Any thoughts on my findings for the calibration 90 degrees seems to have fixed it at 1800 and at 1300 on calibration tension
During the last 9x9 grid at full sheet i did have a random screw fall out of somewhere so need to find where that came from

1 Like

I would hand measure your anchor points very carefully and compare with the values in the yaml file.