I run G32 and G29 S1 to enable mesh levelling which is not always need if your bed is in good shape and level I dont run G29 everytime as it stores the heightmap
I run G29 if something has change ...nozzle, effector but only then
Yes you need to make sure the nozzle is clean as it cant account for plastic on the nozzle.
You can run G30 S-1 to do a single probe as a minimum for G29 S1 to be effective
I have tried measuring with Digital calipers the distance between bottom of X rod and top of Y and ended up with L=66.68, R=66.8 and also top of X rod and bottom of Y rod L82.7 R82.77 which is off but not considerably and I don't really trust I can measure very accurately.
What I have noticed though is that the distance between the two X rods is 36.88 on the far left and right but nearer to the X carriage it goes to up to as high as 37.3 so I think the X carriage bearings are not spaced correctly and stretching the rods apart. So that could be it. though if this was affecting the levelling I would have thought it would be a curve along X rather than a slant consistent the whole way.
I forgot to upload my solution to this thread, because luckily I've been able to solve it.
It turned out to be simpler then I first assumed. Basically, the error occurred due to another error in the measurements in calibration. This was then stored in the override.g file. I was suspecting that there might have been an error there, but each time I tried deleting it, it was already stored in the machines RAM.
Here is a step by step guide, explaining how I fixed it:
Temporarily delete/comment out the M500 command in the bed.g file. (It causes the machine to store the new values after calibration)
Delete the config-override.g file and do an emergency stop of the machine.
Restart it, (hopefully there shouldn;t be a config-override.g file) and add the M500 command back in, then perform a delta calibration.
This should have: removed your previously stored values in config-override.g for good --> allowing for new values to replace them, so the printer will home properly.
Also, you don't need the G92 command, it was only a temporary fix
Never mind, i figured it out. The reprap firmware configuration assistant was using M208 to set the minimum z height to 1mm. I changed it to 0mm and now when i run the calibration I'm getting 0.1mm as expected.
To be honest, the configuration assistant was more trouble than help. I'd almost recommend not suggesting it in your documentation.