Auto Tuning Tips

In most motion applications, proper servo loop tuning is the key to obtaining an acceptable balance of stability and performance. In order to maximize the use of the auto tuning tool in RSLogix 5000, you need to start with the correct boxes checked and the optimal values entered for your application.

Generally speaking, there are only (3) types of motion applications:

  1. Constant Speed Applications – Conveyors, line shafts or cranks for example
  2. Tracking Applications – Winding/unwinding, flying shear and web control
  3. Point to Point Applications – Pick-and-place, packaging and cut to length

The recommended starting points for each of the (3) types of motion applications are shown below. Always keep in mind that tuning involves trading performance for stability. As a result, be sure to have an emergency stopping capability at hand.

If you are still having trouble with stability after the auto tune, perform the auto tune again,but when the Position Loop Bandwidthscreen pops up, lower (divide) the resulting Bandwidthby the Load Inertia Ratio as a starting point. Press OK and test your results. Repeat this process and keep raising the Bandwidth little by little until you get the desired performance and stability. This is known as the Iterative Tuning Method. Finally, you can add the Output Filter checkbox and re-tune to try to eliminate machine resonance, if necessary.

NOTE:Set the tuning Travel Limit (safe travel distance), Speed(of application) and Torque Limit(leave at 100% if possible) according to your application. Never use bothError Integrators; choose thePositionor Velocityintegrator based on which one is more crucial in your application.

Constant speed application auto tuning starting point:

Tracking (following)application auto tuning starting point:

Point to point application auto tuning starting point:

NOTE:Add the Friction Compensation checkbox if you have trouble achieving endpoint.

Add the Torque Offset checkbox with Bi-Directional tune option for vertical applications.

Manual Tuning Tips

NOTE:It is highly recommended that you try to employ the iterative tuning methoddiscussed on the previous page, when possible. Lowering the position loop bandwidth provides a single parameter value that can change the overall loop performance of your system. This method also keeps the proper relationship between the associated velocity loop bandwidth, the position error integrator and the velocity error integrator and this helps to maintain a stable system.

If you are unable to auto tune your system (possibly due to mechanical travel restrictions), you may need to manually tune. You can begin by manually entering your calculated Load Inertia Ratioon the Output tab of the Axis Properties window (online or offline). This value may come from a motion sizing program, such as Motion Analyzer, or it can be calculated by hand. Entering the inertia ratio sets the Torque/Force Scaling, which is directly related to the system inertia. Once set, do not change this scaling, unless your application dictates a change (i.e. a winder where the inertia is variable).

Block diagram of typical servo drive loops:

Direction of tuning (inner to outer)

Note: The current loop is tuned within the drive on SERCOS, so you have no access to it. Tune the velocity loop first (using a velocity step program) and then the position loop (using a position step program).

Velocity loopmanual tuning procedure:

To properly tune the velocity loop,you will need to create a velocity command step signal for the axis. This will move the motor forward and backwardwith an instant velocity (infinite acceleration) or step. The values for the step amplitude (velocity) and frequency (period time) depend on mechanical load and have to be considered carefully because you can damage the mechanics. Typical values are: velocity step ≈ 5 - 10% of maximum application speed and the frequency ≈ 1 sec.

The required step function can be easily created within RSLogix 5000 by using a Time-CAM (MATC) function.

[ ] / Master / Slave / Type
0 / 0.0 / 0.0 / Linear
1 / 1.0 / 1.0 / Linear
2 / 2.0 / 0.0 / Linear
*

The theoretical maximum Velocity Proportional Gain for Kinetix 6000 equals 1352 [1/s] or 215 Hz. The maximum values for Ultra3000 are 551 [1/s] and 88 Hz. Take this maximum gain value and divide it by the Load Inertia Ratio to get a stable starting point.

Important: Set your Position Loop Gain(s) = 0 and the Velocity Feedforward Gain = 100 (%). This takes the position regulator completely out of the equation so you can focus on velocity.

For velocity loop tuning you will need to monitor (trend) the following signals: the Velocity Command signal; and the Actual Velocity signal. Use these signals to compare the response of the system. Also monitor the Torque Commandsignal in the trend and be certain to stay below your torque limit at all times. Carefully increase the Velocity Proportional Gain in small increments until it begins to “ring” a little. You may actually hear this, but the trend should more likely begin to show an unstable oscillation on the actual velocity or the actual torque. Reduce this Velocity Proportional Gain by 30-50%, as a rule of thumb.

Finally, if you require extremely tight velocity control, you can add Velocity Integral Gain by using the formula Vel I Gain = (0.25)*(0.001)*(Vel P Gain) 2. This should conclude your tuning for the velocity loop.

Position loop manual tuning procedure:

The last loop to tune is the Position Loop. As a practical starting value for the Position Loop Proportional Gain you can start with a value which is 1/3 of the Velocity Loop Proportional Gain.

For position loop tuning you will need to monitor (trend) the following signals: the Velocity Command signal; the Actual Velocity signal;Torque Command signal; and the Position Error signal. Use a move (MAM) command to create a position step with high acceleration and use a 1 second delay before repeating the move. Tuning this loop will eventually be done at application speed, but as a starting point we will use 10% of this value.

Begin carefully increasing the Position Proportional Gain value until the trend begins to show an oscillation on the actual position or the actual torque. Reduce the Position Proportional Gain by 30-50% depending on the application. For point to point applications, you may need to add some Position Integral Gain. For tracking/following applications, adding Velocity Feedforward Gain (ideally 100%) will nearly eliminate position error. Repeat this procedure at full application speed and fine tune again, if needed.

Finally, if you require extremely tight position control, you can add Position Integral Gain by using the formula Pos I Gain = (0.25)*(0.001)*(Pos P Gain) 2.

Logix./SERCOS Auto/Manual Tuning Tipsv1.0210/20/2018