by cncdrive » Mon Mar 12, 2018 10:35 pm
What we found is a syncronisation problem of the homing while g-code is running.
It is because of the issue fix with the full circles which I described in the release notes, that if a full circle was coded with first the start point and after some other codes, e.g. text macros which made the controller to syncronise and only then the arc coded then if the starting point of the full circle was not on the grid defined by the 1/steps value then due to the syncronisation of the coordinates this could possibly caused the controller to not cut the full circle, because the resyncronised real coordinates made the full 360° of the circle to around 0°. We fixed this problem with changing how and when the controller is syncronising, but forgot that the homing should be also changed how it syncronises, so now when the home is called from g-code (G28.1) that does not work properly and because it is not syncronising the coordinates the G28.1 could cause issues, like more than one axis moving when homing and axes reversing back to the coordinates where they started and things like that.
Since some CAM programs posting G28.1 into the g-code files and so it might be often used we thought it is better to pull this release because of this issue.
The other thing is what I think Gerry reported is that the probing always give back 0 values in this test release which is due to the upgrade of how the #vars update. It is also noted in the release notes.