G28 incorrect position feedback

[lutorm]

I just tried to use G28.1 to define a position and then G28 to move to it. It appears there's something funny going on with the position feedback during a G28 move. Here's what happens:

• g0 g53 x10 y-10 z-10.
• g28.1 now stores the position (10,-10,-10) in g53 coords.
• g10 l20 p1 x20 y20 z20. This same position is now (20,20,20) in g54 coords.
• g54
• g0 x21 y21 z21
• g28

The machine now moves to the correct position, (10,-10,-10) in g53. However, it also reports that it is at position (10,-10,-10) in g54 coords! This is displayed as the g54 coords in chilipeppr, and I think it's a G2 issue because the serial port JSON after the move says:

{"sr":{"posx":10.000,"posy":-10.000,"posz":-10.000,"stat":3,"mpox":10.000,"mpoy":-10.000,"mpoz":-10.000}}

I thought this might be how g53 moves are reported, but if I manually do a g53 x10 y-10 z-10, I get the correct feedback:

{"sr":{"posx":20.000,"posy":20.000,"posz":20.000,"stat":3,"mpox":10.000,"mpoy":-10.000,"mpoz":-10.000}}

[justinclift]

Looks like this got missed somehow. It definitely sounds like a bug.

Even though this was reported ~1.5 years ago, we should should definitely check if it's still happening in master and/or edge branches then figure what to do from there.

@lutorm Just to ask... since you're still doing stuff with g2core (eg judging by #323 :wink:), did you notice if this bug was still happening with edge in March? :smile:

[aldenhart]

I will test this one, but I think it's been fixed. Let's leave it open.

[lutorm]

No, I haven't looked lately. I don't have a lot of time for this right now but I can try it next time.

[justinclift]

No worries. Thanks @lutorm. :smile: