Electric Automation Forum
Forum » Automation Technologies » how to do load (torque) sharing with Allen Bradley servos?
Topics: how to do load (torque) sharing with Allen Bradley servos? on Automation Technologies
#1
Start by
Mark Kennedy
03-12-2014 01:58 PM

how to do load (torque) sharing with Allen Bradley servos?

Does anyone out there have experience with how to do load (torque) sharing with Allen Bradley servos? My system has 2 Kinetics 6000 servos linked by a belt driving the same load.
03-12-2014 04:24 PM
Top #2
Muralidhar Ekambaram
03-12-2014 04:24 PM
You referred Servo. It would be better to know whether the application demands Motion control or Speed control. Usually when multiple drives are employed for speed control, One will be on Speed mode and the other can be on direct torque mode or with speed follower with torque trim. How ever the exact blocks available along with application details in any system can only decide the apt methodology. The Servos may have been employed for fast response or Moment of Inertia requirements or Fast start / stops or space constarints.
03-12-2014 07:03 PM
Top #3
Mark Kennedy
03-12-2014 07:03 PM
Motion control. Two motors are used because of the difficulty of finding a single motor with enough torque for the application in a potentially flammable environment.
03-12-2014 09:26 PM
Top #4
Nitin Kakkar
03-12-2014 09:26 PM
You need to go for master follower since it's a belt coupled system, you won't be able to run slave in pure torque mode as it will slip. One way is to use window control(not sure if this mode exist in this drive) where slave will be in torque mode but will go to speed control if speed difference from master is big. Other possibility is you can create a PI controller in your plc and reference to slave would be master ref + speed correction to maintain torque same for both motors. PI controller sp would be master torque and MV would be slave torque and max output of controller would be max speed deviation allowed. This control loop need to be executed in fast task for better control.
03-12-2014 11:52 PM
Top #5
Mark Kennedy
03-12-2014 11:52 PM
Clarification: The system is belt coupled, but the belt has teeth (timing belt). There should not be any slip. Perhaps a small amount of stretch.

Right now, I am setting up a torque mode option that can be turned off and using a small torque boost from the slave.

I also have set up a software "gearing" option to gear the slave to the master. I started with a gear ratio of 1:1 and then added logic to adjust that ratio up or down by a very small amount (0.0001) every program scan if the loads on the 2 motors vary by more that 5%.

Both of these options will NOT be enabled at the same time. I am coding different options in advance because this has to be implemented over a weekend shutdown.
03-13-2014 02:36 AM
Top #6
Joshua Irish
03-13-2014 02:36 AM
Interesting solution to the problem.
03-13-2014 05:15 AM
Top #7
Muralidhar Ekambaram
03-13-2014 05:15 AM
As I understand it is a position control application single load shared by two servo motors & drives. Timing belt is considered as positive transmission mechanically. I do agree to the method adopted by Mr. Kennedy - Position control (Gearing) with Torque limit in the master drive. Please also share if possible whether the application is periodic start - forward - stop - reverse - stop kind of motion or any other type of application characteristics. Can the application be more elaborated.
03-13-2014 07:31 AM
Top #8
Mark Kennedy
03-13-2014 07:31 AM
Yes it is position control. Start- forward - stop - reverse - stop - repeat continuously as the machine is running. ~1.5 meters of travel at rates up to 60 cycles per minute.
03-13-2014 10:26 AM
Top #9
Clay Bodenheimer
03-13-2014 10:26 AM
I would be concerned about motor heating with that cycle rate. You may need to also look at cooling for the servos as with starting and stopping that frequently will get a motor hot fast.
03-13-2014 01:17 PM
Top #10
Muralidhar Ekambaram
03-13-2014 01:17 PM
So far I had observed in all my selections that if the point plotted with rms of application torque and average speed of application falls within the envelope defined by Servo motor manufacturer, the application behaves OK. Of course MOI mismatch acceptable is decided by type of Motor and control configuration used - Feed forward/ PID and the Mechanical & Electrical time constants of the Servo motor selected. Good amount of knowledge base material related to selection of system is available with Galileo Controls. Also you may please try to employ Cycloid motion profile in place of Trapezoidal or Triangular profiles even with smoothing S curves at speed transition points . Out of my experience the cycloid profile safe guards the subsequent mechanical transmissions although demands more peak speeds. Typical Servo motor applications can be witnessed in a Synchronous On the Fly Paper sheeters where Double Drag link transmissions & Fly wheels have been replace Dual servo drive systems similar to what you are employing. In fact I was very successful when I replaced the double drag link transmissions and employed Quasi Cycloid profile with Feed forward system on 4 - 5 such installations nearly a decade back employing 28 kw Asynchronous motors with back lash free gear heads. Wish you all the best.
03-13-2014 03:25 PM
Top #11
Pierre Drouin
03-13-2014 03:25 PM
If you are using ControlLogix as servo controller, either Sercos, Ethernet or analogue (ME2AE controller) i would suggest a master-slave relationship. The master axis would be an regular servo axis in closed loop. The slave axis would then be the second motor, but controlled in servo axis in open loop. This servo axis has to be open in terms of servo axis. In fact, a torqu compound will be needed as explained very well by Mark earlier. Regards! Pierre
Reply to Thread