Ultra3000, RSLogix Tuning problem

Hi,

I'm hoping someone can assist us with our current dilemma.

We are using an Ultra3000 drive with a MPL-A330P-M motor, with RSLogix 5000 v19, connected via SERCOS. The motor is for a linear axis, moving a part up and down along a screw rail.

We have tuned the drive successfully without a load, and now need to tune this drive under load. The procedure we have been following to do this was left to us by the consultant who originally designed the PLC / drive arrangement:

- ensure the drive is connected, and not enabled.
- switch to Program Mode
- disable Drive Enable Input Checking
- use the Start Tuning button on the Tune tab.
(we hear the drive clicking away as expected)
- accept the values
- renable Drive Enable Input Checking; switch back to Run Mode.

The resulting tuned figures; notably the Load Inertia Ratio on the Output tab match units we have previously sent out in the field. (roughly 7.x under load compared to 2.7x not under load).

However these tuned settings result in the brake being on permanently. We can drive the motor but in its idle state it seems to be driving the brake, almost as if the motor is hunting against the brake.

In the meantime we'll be checking the brake contactors etc, and running the motor using its Tuned values from not under load. (This results in jerky movement)

Is there anything anyone can suggest to alleviate this problem?

Will the consultant come to site and fix up his stuff up? Servo Tuning is best performed by someone who does it on a regular basis.

How good is your Programming?


You do not have the full motor nameplate so I cannot tell if the motor has the inbuilt brake (Assuming you do and it is wired correctly to the drive)

Notes for a Hoist Tune (you have a linear axis moving up and down with a brake)

1. Auto tune only in the up direction
   • This is the direction of greatest load
   • Start at the bottom of the stroke before initiating movement and set the tune movement limit so you do not exceed the stroke limit
2. Damping factor - Default is 0.8 - Under damped response, I usually use 1 ( critically damped) or 1.2 (over damped)
3. On the Axis properties go to the "Fault Action" Tab press the "Custom Stop Action" and set the Brake engage time and Brake Release time delays - Default is 0.0 - I have typical values of 0.1 to 0.2 seconds
4. When Tuning I have a trend running with position, Position error, Velocity, and Torque
   • I also look at the same trend during normal running to check tuning and find "tight spots" which I push back to the mechanical guys (I've found alignment issues sometimes at one end of the stroke)

What tickboxes are ticked on the auto tune tab?

Hi,

and thanks for the quick response. It looks like you've resolved our dilemma.

Our PLC programming is reasonable, but does have something of a blind spot when it comes to drives. The original consultant has stopped consulting and as such is unavailable.

However your suggestions did point us in the right direction.
- set-up trending to assess the drive profile.
- all checkboxes ticked on auto tune.
- we had been doing a Bi-directional Auto Tune, which I think was the root cause of our problem. Changing this, increasing the damping factor to 1.2 and setting Brake Engage times around 0.1s has improved things.

Thanks,

Thomas

tbrightwell said:

Hi,

However your suggestions did point us in the right direction.
- set-up trending to assess the drive profile.
- all checkboxes ticked on auto tune.
- we had been doing a Bi-directional Auto Tune, which I think was the root cause of our problem. Changing this, increasing the damping factor to 1.2 and setting Brake Engage times around 0.1s has improved things.

Thanks,

Thomas

Click to expand...

Your welcome - Glad you got it to work

As a side Note: On the Auto Tune you usually do NOT want all boxes ticked.
On the systems that I usually tune I get suitable performance with Damping set to 1, Velocity and Accel FF turned and Low Pass filter. I only turn on Position integral and velocity integral if the application and resolution requires it. Integral can cause excessive hunting causing mechanical issues.

See Rockwell in Motion-UM001 pdf "Tune a SERCOS axis"

Where tighter positioning is required, Integral gain, and
feedforward constants can be selected. However, when used
incorrectly can cause axis instability.

Click to expand...

Personally I like using the feed-forward constants to set the the motion profile following error close to zero.

The document you pointed at was one we'd somehow been missing all these years. An in-depth read of that proved very informative.

We Auto Tuned without the Integrator settings as you suggested. We also reduced the maximum torque / speed at which we tune (the most we run at is 60% of the rated value)
We were also able to trace a mechanical issue using our greater understanding of things.

Our MAJ and MAM instructions within the code still have a tendency to overshoot the target velocity somewhat. Our parameters for a typical instruction tend to be

Speed (selected value i.e. 50%)
Accel Rate 60%
Decel Rate 60%
Profile S-Curve
Accel Jerk 100%
Decel Jerk 100%
Merge Disabled

Is there anything we can do here to reduce this overshoot?

Thanks again for your help.

tbrightwell said:

The document you pointed at was one we'd somehow been missing all these years. An in-depth read of that proved very informative.

We Auto Tuned without the Integrator settings as you suggested. We also reduced the maximum torque / speed at which we tune (the most we run at is 60% of the rated value)
We were also able to trace a mechanical issue using our greater understanding of things.

Our MAJ and MAM instructions within the code still have a tendency to overshoot the target velocity somewhat. Our parameters for a typical instruction tend to be

Speed (selected value i.e. 50%)
Accel Rate 60%
Decel Rate 60%
Profile S-Curve
Accel Jerk 100%
Decel Jerk 100%
Merge Disabled

Is there anything we can do here to reduce this overshoot?

Thanks again for your help.

Click to expand...

I am unable to determine what is actually going on because there is not enough information - I will give common pointers to what may be going on.

Have you found Motion RM002.pdf ? - Contains How to tune S curve motion and trouble shooting velocity overshoot.

Use the trend - During the Accel is the Torque Flat lining? If so then you have too much Accel for the load or you have limited the output Torque
Especially Check the Decel in a Lower situation - this is where all the grunt is required.
Max Accel at 100% Torque is on the "Output" Tab
Peak torque Available On the "Limits" Tab
Continuous Torque Available on the "limits" Tab
Have a look at the Custom Limits - you get Fwd and Rev limits on Vel Accel and Torque (only had need to use them once so far)
Read up on Jerk in RM002

• Units/s^3 - same value of jerk no matter what the accel or decel rate is
• % of Maximum - Same value of Jerk no matter what the Accel or Decel rate is but can be changed when the Dynamics Tab is changed
• % of Time - the Accel and Decel Profile is the same no matter what the accel or decel value is. Jerk changes as the Accel rate changes.
• Play - Get the emulator or a spare CPU and create a Virtual Axis and some trends and try the MAM with different Velocities and Accel and Jerks and units until you understand what the effects are.

Personally I use real units eg mm/s, mm/s^2 for velocity and accel where ever possible.

• This helps when you use solid works CAD and mechanical designers like it also.
• Another reason is: I like to use CAM tables for motion profiles to and I want to relate the profile units to other Motion commands. eg I can test an axis using MAM with real units and work out a suitable accel and decel rate for the axis (with product) then create the CAM table using the same units.

Notes:

• When you use a 60% Accel Rate that is 60% of the maximum on the dynamics tab - The Dynamics Tab values are set automatically by the auto tune and if you do not adjust them they can be too high. Also if someone performs an Auto tune later your adjusted values are reset. (A Gotcha)
• If you perform a MAS and say NO to the change accel rate then it stops at the Full Dynamic Tab Decel rate.
• There are also conditions that cause the axis to stop at full Torque (I just cannot remember them at the moment - Looked it up MSF and fault conditions)

If you want more help the post the trend - Must have a minimum of Vel, Accel and TORQUE - I also prefer to add the position error - to get it to update must have motion group Auto Tag Update on and in the axis "Servo" tab select it from the real-time attribute.