So I am playing around with the PS-AMC4 right now just trying things out and doing some testing. I am just using 1 axis and doing a simple move of 1 rotation. This works fine when it is on its own, but I am trying to get a feel for the encoder feedback and registration instructions now.

So I loosened the bolts on the encoder so I can turn it by hand a small bit to deliberately introduce error counts into the move. This works as expected and after the move is done the axis position will say 10,000 counts, but the external encoder position will say something like 9,300 counts.

I then subtract the external encoder count from the axis position to get how many pulses I need to move and in what direction. That math result tag is the correction distance of the manual registration instruction. According to the help file, with MREG I can select the master and slave axis to be the same one as long as the master is actually looking at the external encoder as the master position.

When I load the correction distance tag with some counts the axis will move either forward or backward like the counts dictate. However, it moves a small amount like it should and starts to slow down, and then suddenly starts to accelerate again and just goes into runaway. The axis will constantly keep accelerating and never stop, trying to reach the axis max hardware velocity. If I let go of the run button to kill the instructions the axis will stop immediately, so it is definitely the active instruction causing this behavior. I also don't get the runaway problem unless the manual registration instruction is active.

I am not sure what is really causing this behavior. My first thought was that the "apply" subrung to the MREG originally had a NO contact to it, and I thought maybe it was trying to apply new correction distance every scan or something, even though the instruction is edge driven. So I changed the NO contact to a rising edge contact instead just in case, and this did not change the behavior I was seeing at all.

Im not sure if there is some flaw in my ladder logic, or the instruction simply isnt intended to be doing what I think it should be, or if the instruction itself is bugged out and deciding to turn itself into a VMOV after it finishes its correction. Can anyone take a look at my ladder and see if they can spot something wrong? Or if you have experience with the PS-AMC and registration instructions and can give me some pointers?