No announcement yet.

Motion contol methods

  • Filter
  • Time
  • Show
Clear All
new posts

  • Motion contol methods

    So, I know enough about motion control to be dangerous.
    I'm in the process of a control refit on a machine that draws out lengths of plastic film, and has a hot knife (driven by a separate motor), that cuts and seals, then repeat, in the area of 60-70 times per minute.
    The old control system, and others I have been into, use a speed control method, driving the servo amp with a 0-10 volt signal.
    My question is, If I want a specific, repeatable distance move, at high speed, why would I use speed control over position control mode?
    The BRX PLCs lend themselves to this type of application, and so far, it seems simpler to build the logic to use step / direction pulse trains over creating an analog out to drive the servo amp.
    What am I missing? What is Normal? Is one method faster than the other?
    anything is possible with the right sized hammer.....

  • #2
    They're probably not using the servo in velocity mode as mentioned but controlling the drive with an analog signal with encoder feedback going back to a motion controller. That how almost all old servo systems worked. The drives were set to run in speed or torque mode and fed a Ī10VDC signal. The feedback from the motor either went back to the drive and then to the motion controller or directly to the motion controller. The motion controller would close the loop.

    In your application, I would use position control. It's going to be very repeatable. If you're using a BRX cpu, you can drive a servo amp. with pulse and direction inputs and use a trapezoidal move instruction.

    I typically do all servo motion over EtherCAT these days which is superior to just about everything else. That option is not offered on any of the cpu's at this time.

    Just make sure you're max pulse output is fast enough to move your motor at the speeds you need. More than likely it is.
    Last edited by Adisharr; 10-25-2018, 10:40 PM.


    • #3
      Originally posted by MrMark View Post
      So, I know enough about motion control to be dangerous.

      The old control system, and others I have been into, use a speed control method, driving the servo amp with a 0-10 volt signal.
      My question is, If I want a specific, repeatable distance move, at high speed, why would I use speed control over position control mode?
      Why would you? you probably wouldn't, why did they? I've highlighted the key word, back before microcontrollers were cheaper than a cup of coffee, one saw a lot more analog stuff. I wouldn't be any too surprised if the existing motor used an analog tach. for feed back rather than a digital incremental encoder. (At least a similar system I worked on did, it failed due to (I think) bad brushes in the tach. and i replaced it with an A.D. SureServo , in that application management didn't want to change the control system so it is still in velocity mode) That said, there might be application where one would still want to use that method in a new project (though I doubt yours is one of them) While modern servo systems make it very easy and (usually) precise to control the position of the motor, there could be times, (e.g. due the lack of rigidity in the system) one might prefer to close the loop at a point other than the motor.


      • #4
        Thanks guys for the info. I took a closer look at the existing drive today, it has no inputs for step / direction, the encoder from the motor goes back to the amp, then the amp feeds it back to the controller, which is where we think the problem is.
        Unfortunately, I had to go elsewhere for a new motor and amp, none of the ones offered by AD came close to the torque I need, the existing motor is / was rated at 9 HP, a little over 900 in lbs of torque. I sourced a system from Anaheim Automation that is just over half of that, but, I can change the pulley ratio to gain that back, and still only need 200-300 RPMs out of the motor.
        My next decision is if I'm going to use 2 controllers, one for the motion part, and one for all the ancillary machine functions.
        Thanks again for your input, I feel better about going with position control now.

        anything is possible with the right sized hammer.....


        • #5
          A servo drive doesnít need analogue speed reference. What you describe resembles a regular AC motor drive.

          A servo drive gets configured to execute a motion path. The mode is either torque, velocity (following)- these modes are provided by most AC drives as well, although not in servo configuration. A third mode is position following and the main difference to speed following is that it will stop upon reaching the desired position, while speed follower advances that end position all the time, which determines the drive to ramp speed as to maintain the same lag to end position.

          i donít think the film winding needs any servo drive. The only servo actuation needs to be the knife. The knife servo will retrieve the film length as a shaft encoder count and the knife will always cut to length.

          of course is not that simple, especially if you need to cut at the trigger of a photo eye, but having a knife chasing closely and only ramping to cut exactly is possible.

          theckisters package wrapper machine has a rotary shaft knife which turns to cut only when triggered by plc. This is bad because the shaft accelerates and decelerates too fast, the eventual ratched gear breaks frequently. A simple servodrive turning the knife slowly and onlyramping when cutting is needed would solve the problem of frequent breakdowns.


          • #6
            This is definitely a servo, somewhere around 20 years old. The knife will have a regular 3 phase motor with encoder feedback driving it in constant motion. another encoder will keep tack of the shaft position, and the servo will be locked out from moving for the bottom 20 or so degrees of the knife shaft.
            anything is possible with the right sized hammer.....


            • #7
              in my example with the kisters machine, the rotary knife would have an abrupt motion, way faster than the film, in order to realize the cut. The downside was the sudden start/stop of knife shaft, which in time will ruin the ratchet mechanism. The knife was just a jagged blade along the shaft, obviously sharp when freshly installed (wear out part).

              the length of cut was never dead accurate, but since the film needed to wrap the product, accuracy was not needed beyond not wasting much film.

              i believe your system is similar, although most probably not exactly the same mechanics.

              im trying to figure out the need for analogue speed reference. I believe is for film speed only. If thatís the case, then the two encoders youíve mentioned make sense: one to read the film speed from a shaft, and another one to adjust rotary knife speed- you mention this second encoder located on motor that drives the knife.

              youre asking about position versus speed following mode of programming the servo

              The short and obvious answer is speed reference. Iíll explain the difficulties of position reference and why I believe it wonít work.
              for position reference, you program the ramps, max speed and position count. Then execute motion. Upon reaching position, the motion already ramped the speed down to zero in order to hold position. Since the knife shaft cuts at higher speed than the advance of the film, it results in high inertia which the servo I doubt would be able to handle. This is the main reason I donít think it will work. You still need a kind of clutch to handle the high inertia, but then implementing a servo is not worth.

              problem is when the operator changes the speed of the film on the fly. Say it increases it. Then your trigger to start motion needs done earlier. However, youíre ramping parameters need increased as well, to maintain same cut conditions.
              This will require an operator panel able to send to plc two sets of data: knife trigger and knife speed. Then the operator needs to visually check the film length, until a valid set of parameters is obtained. Repeat the process for a couple more speeds for the film and youíve got parameters for consistent cutting at several speeds. You donít need at any speed of the film, just several.

              now why it wonít work.
              in my kisters example, the blunt of motion of the knife shaft would be supported by a ratcheting mechanism, which would wear out fast.
              in this projection, there is no clutch involved. Therefore the servodrive would overload during acceleration or deceleration ramp due to high inertia of that motion. It will fault at best, or burn at worst. Replacing a servodrive proves more expensive than replacing a clutch.

              why speed following could work better
              matching the speed of the film would be easy task for the knife servo, there no high inertia to overcome. However, a means to lift the knife shaft to cut position and lower it after still need to be implemented.

              you have a nice project, Iím sure youíll find a solution Iíd love to hear about. I wouldnít be surprised if your solution proved my understanding wrong or incomplete. If you need help programming the servo or even the plc, just ask. Good luck!