I made some progress…
Below the video of the homing sequence:
Homing sequence with Grbl:
I configured X as first axis and Y as the second axis to home. It will move X till it reaches the limit sensor (hall sensor), back off x mm (setting $27) retry the switch and then backs of xmm again. this makes it impossible for me to have both sensors in alignment. The X axis sequence takes the magnet out of reach when it backs of the x mm. I tried 0mm but then Grbl seems to have issues- I haven’t found a way yet to customise this behaviour. For now I solved it by adding customised sensor holders and move Y out of alignment. the homing works but the arm is not 100% extended. I don’t think this is an issue, still wondering how to solve this.
It seems the main axis of the X arm (reaching down through the core unit) has play and/or the belt tension is pulling it out of the vertical. This messes with sensors and spacing in general.
But I also haven’t properly attached the stepper with a long screw through both plates; this probably will solve the issue.
it won’t always upload the Grbl firmware. I couldn’t find the reason yet yet but sporadically it will work. I can upload the smaller example codes without a problem. Just Grbl Resetting, moving the cable, changing programer option (shouldn’t have an effect at all) all of these steps have seemed to worked once in a while
It will make a local copy of any manually added library. So when you modify any configs on the original (download) location, your changes have basically no effect as they are not in the local copy ! it took me a day to realise why in my case the homing sequence wouldn’t move anything; it was still trying to move Z axis although I have changed the config (so I thought)
The CNC shield has two pins for the limit switches Gnd / Signal. The hall sensor has three pins Gnd, Vcc, Signal. I don’t have any electronics background, so just tried to attach Gnd + Signal to the limit switch pins (X-) and Vcc to a 5V pin elsewhere , but this didn’t work at all. I have the feeling the X stepper driver died on this experiment. Eric Mikson gave me a hint (Wiring optical endstops · Issue #806 · gnea/grbl · GitHub). this brought me on track, I had to use a PNP transistor as the hall sensor seems to have the opposite behaviour as the optical sensor, but they work now
After trying the hall sensor one of the stepper drivers stopped working, but I didn’t know at that point it was the stepper driver. During my trial and error search I changed the CNC shield. All of a sudden X & Y start to move together. I was busy with the hall sensor and didn’t realise the causality.
Today I found the issue: a bridge on the soldering between X and Y step (pin 3-4 from the left) pins
Next is a rudimentary base and table to test magnets, steel ball, sand…