Logix MAM not moving to target successfully

phuz · Apr 6, 2016

In continuing with my Linear Thruster project, I have noticed an anomaly that I can't correct. After running a sequence, we issue a stop command and then use a MAM to move both axes back to zero position. The commanded position is 0mm, but the actual position can show as much as 2mm after the instruction has completed. I feel like I would get better results if I just used a jog at a slow speed and then used a MAS when the actual position reached 0. Any idea why I am getting a differential between commanded and actual position when using an MAM? Any chance this is because I am issuing a MAM on a virtual axis and using MAG instructions to couple the virtual to each physical axis?

kekrahulik · Apr 6, 2016

My first thought was to check if your MAM was absolute or relative. But after seeing the statement about virtual axis and a MAG, I would guess that your real axis is not gearing to the virtual when they are both at the same position or that your MAG is not setup in a 1:1 ratio.

I would look at the relative positions and the MAG gearing

phuz · Apr 6, 2016

MAM is absolute.
MAG shouldn't matter now that I think about it, because when I looked at the physical axis, the commanded position was 0, but actual stopped short.
MAG gearing is 1, but that shouldn't matter because 0 is 0, right?

keshik · Apr 6, 2016

Have you looked at your in position tolerances as well as your tuning? It's possible that it considers 2mm to be close enough to zero so it doesn't try to get back to it.

phuz · Apr 6, 2016

On the position loop for each axis, error limit is 40mm, and lock limit is 1mm.

keshik · Apr 6, 2016

If the lock limit is 1, the commanded position is 0, then it shouldn't be saying it's in position at 2mm. I'd be curious as to what your tuning looks like - it may be that you don't have enough integral output to suck it into its final position.

phuz · Apr 6, 2016

keshik said:
If the lock limit is 1, the commanded position is 0, then it shouldn't be saying it's in position at 2mm. I'd be curious as to what your tuning looks like - it may be that you don't have enough integral output to suck it into its final position.

See attached.

kekrahulik · Apr 6, 2016

What model servo drive?

Are you stopping the MAG or does it remain active? I assume that you virtual axis is getting to zero. If you are remaining geared, I can't believe that the drive would be geared 1:1 to the virtual and not get to zero as well.

Doe sit behave any differently if you change the accel/decel or velocity of the MAM?

phuz · Apr 6, 2016

Kinetix 6500.
I would also assume virtual gets to 0, but did not confirm yet. I also haven't changed velocity nor accel/decel. I didn't think that would have any bearing, but can try it tomorrow.

kekrahulik · Apr 6, 2016

Try doing a MAM on the drive without the MAG active. If it still doesn't get to zero, then you have probably ruled out the MAG as a culprit

keshik · Apr 6, 2016

I would consider making the integrator bandwidth on the position loop a non-zero value. Start with something small and watch the control output when it goes into its final positioning.

phuz · Apr 8, 2016

I played with this more yesterday on site.
Without a load attached, these actuators get into a rather close position to 0. However, with the rigid load attached (which connects the two actuators by means of a solid mechanical linkage), the actuators will rise from a negative position toward their "0" position together. Both physical axes show a commanded position of 0. The one axis stopped at -0.05mm, which was great. The other stopped at -1.43mm, which was not so great. I used MAM instructions on each physical axis, not the virtual axis, so with the lock position limit of 1mm, I would have expected it to stop inside of -1mm, not -1.43.

I am not understanding how the instruction can call itself complete (PC) when it has a commanded position of 0 and fails to reach it.

Peter Nachtwey · Apr 8, 2016

There is a in position parameter. If the error is less than the in position parameter the controller will say it good enough.

keshik · Apr 8, 2016

phuz said:
On the position loop for each axis, error limit is 40mm, and lock limit is 1mm.

Is the lock limit not the "In Position" bit for AB?

phuz · Apr 9, 2016

Peter Nachtwey said:
There is a in position parameter. If the error is less than the in position parameter the controller will say it good enough.

So I should lower the error to less than 1mm as well?

Logix MAM not moving to target successfully

phuz

Member

kekrahulik

Member

phuz

Member

keshik

Lifetime Supporting Member

phuz

Member

keshik

Lifetime Supporting Member

phuz

Member

kekrahulik

Member

phuz

Member

kekrahulik

Member

keshik

Lifetime Supporting Member

phuz

Member

Peter Nachtwey

Member

keshik

Lifetime Supporting Member

phuz

Member

Similar Topics