Announcement

Collapse
No announcement yet.

Fiber Winding Machine Application

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts


  • Fiber Winding Machine Application

    Hello. Long time AD user. First Time poster.
    We have a 2 axis cylinder winding application that I need some help completing.
    It has GS4 on Mandrel, with encoder to PLC, and a Sure Servo 3KW on Carriage. I have recipes from HMI. Control system is BRX, P#: BX-DM1E-36ED13
    The operator controls the mandrel speed with a speed potentiometer. I feed the speed pot into plc and use analog out to GS4 VFD. I did this to clamp the max speed and control start up speed. One end of the mandrel has a nail strip that catches the fibers. I index the start points on this end. The other end of the mandrel has a cone so the end of the cylinder is less than the length of the part. For the example below, the cylinder stops at 9.75', although the wind material goes wraps around the end of the cone at 10'.

    My issue is the turn around on the ends. I need to pass the end of the cylinder with the carriage, decelerate, stop, accelerate in other direction.
    It needs to control the dwell in degrees of mandrel rotation between when it passes the log point to when it comes back the other direction to the end of the dwell.
    For example: Part is 10' long, when carriage passes 9.75', PLC logs rotation of Mandrel=103degrees, carriage needs to decelerate to 10', stop, accelerate in reverse direction and cross mandrel at 9.75' at 193 degrees. This is what they call a 90 degree dwell. No tensioners involved.

    Any Suggestions on PLC Sequence and Commands would be greatly appreciated. Thank you, Michael


  • #2
    I'm guessing your application is similar to a traverse winder.
    I think may help if you provide a drawing or sketch.

    Normally, for a system like this both Axis are Servos. This is to make sure you can control both speed and position with more detail and precision.
    The way I see is this is:
    1) The GS4 rotates and the carriage move slowly (follows GS4) (depending on GS4 speed), to do the winding. This is the carriage going from 0 to 9.75'
    2) When carriage reaches 9.75' (or close), now the GS4 has to reduce speed, and carriage follows GS4 at a different ratio to get to 148 degree at 10'
    Then process revert

    I think this can be done with AXSCRIPT. Maybe something like this:
    * Follow Master ( Master Encoder): Gear Ratio =R1
    * Wait For Register Match Condition (Encoder)= Position1 (9.75')
    * Follow Master ( Master Encoder): Gear Ratio =R2
    * Wait For Register Match Condition (Encoder)= Position2 ( 10' )
    * Follow Master ( Master Encoder): Gear Ratio =-R2
    * Wait For Register Match Condition (Encoder)= Position1 (9.75')
    * Follow Master ( Master Encoder): Gear Ratio =-R1

    Hope it helps
    Last edited by EGoethe; 05-14-2019, 12:49 PM.

    Comment



    • #3
      Thank you EGorthe for your reply. I totally agree with both being servos. We are currently looking at doing just that.

      Here is a quick sketch of the machine. The Mandrel is slower, up to 12-15 RPM, the Carriage has to move pretty quick. At 6 RPM Mandrel, the Carriage is moving at ~17"/Sec.
      I am using the AXSCRIPT instruction currently. I have a do while loop based on being >2" from end follow Velocity of Mandrel, At <2" from the end switch to S-curve with 90K Jerk command. This gets it to work (on the bench) when the Mandrel is slower. If I switch to Trapezoid move when <2", it jumps to final position quickly. This would not work well for the chain.

      Click image for larger version

Name:	winder_sketch.jpg
Views:	119
Size:	75.0 KB
ID:	122715

      Comment



      • #4
        I donít think is possible without tensioner.
        To make it without, have the carriage slide on a double-threaded screw that makes possible a return on opposite thread. This way the carriage will return ( with a jolt), but the carriage drive wonít need to ever change speed, other than adjust to better follow the drum.
        this is the common method for winding.

        Comment



        • #5
          Thank you Alexandru for your response.
          The system does not have an electric tensioner. It does have tension on the fibers as they go thru a series of feeding eyelets to get to the carriage. In the carriage, they are dipped in resin and then the mandrel pulls the fibers onto the mandrel. The tension is significant, but not measured or controlled.

          Comment



          • #6
            Michael_CE:
            Interesting application,
            So I feel you are saying you are OK until you get to 2".

            At that point, if the quality of the motion is not good, I would reduce the speed of the Mandrel, only while doing the final loop, similar like when driving a car and entering a curve.

            IMHO, the final loop would have to be done segment by segment, with *Follow Master ( Master Encoder): Gear Ratio =R2
            R2 is the gear ratio value, equivalent to the pendant of a line. So for instance initially with value 1, then 0.5, then 0.2, -0.2 (here starts to return in the other direction), -0.5, -1, etc
            So, Reach a point, change the Ratio

            Regards

            Comment



            • #7
              A little more thinking on this one:
              - What is the PPR on the encoder you are using?. I think increasing to a much bigger ratio can help. This will change the ratio you are using, but if small the carriage may not respond fast enough.
              - Definitively doing the the loop may require for you to reduce the Mandrel speed. This is due the fast response of the carriage. It is possible the inertia of your system doesn't help. Again going up in to the encoder ppr may help, as long as you don't exceed the 250K/sec

              Thanks

              Comment



              • #8
                Hi EGoethe,

                Thank you for your ideas. Right now I am bench testing 2 sure servos with 10K ppr. I am trying to get a 90 degree dwell on the end of the part where a conical end is in place.
                The carriage response is the issue. I did slow down the mandrel yesterday briefly, and that does aid in allowing more time for the move. At high mandrel speed, 15rpm, it is running at 25.5"/sec on a 54" part. The carriage has to travel at the same 25.5"/sec. to get a 45deg wind angle on the mandrel. The biggest issue I have right now is the decel on the cone end. The BRX seems to have some motion limitations in the instruction set. It can home, jog, trap move, velocity move, move gear, follow move, cam move, script move. However, I have tried several iterations and if the carriage speed changes, I cannot get a repeatable dwell. Please feel free to call me at 314-456-7567. I am open to all ideas at this point.

                Comment



                • #9
                  I have inserted the EGR(Electronic Gear Ratio) in the Sure Servo Amp. P1-44/45. This allows the servo to move more per each input pulse. The BRX has 250Khz max as you know. The 3/2 EGR, allows increases the servo ouput to 375Khz. The machine max speed is about 15RPM on the Mandrel, so the corresponding 25.5" on the Carriage is the maximum there as well.

                  Comment



                  • #10
                    I just realized you control the servo but in Speed control mode and you are using an analog signal to do that.
                    Unfortunately that is not going to work for you.

                    In order for this to work, you would have to use Position Control mode, and will have to send pulses from the BRX to the Servo.
                    This maybe more complicated for you than you anticipated.

                    Comment



                    • #11
                      Im only skimming this thread but from the OPs last post I believe he is in pulse train position mode already, since that is where P1-44/45 come into play and needing to stay within a 250KHz limit of BRX outputs.

                      Comment



                      • #12
                        All,

                        Thank you all for your responses.
                        Currently(bench testing), the mandrel is now a servo and I feed it a speed based on a value from the potentiometer. The trolley (traversing axis) is also on a servo in position mode. The BRX instruction set will let me due parts and pieces of the motion needed, but not all combined. I had our galil rep out today to potentially switch to that for motion control. I am using the Sure Servo EGR as noted above for the trolley axis. It has to move pretty quick to keep the wind up. I got parts of the application to work, dwell on the end. However, the AXCAM command doesn't allow for a CAM motion of the trolley while match speed of the mandrel, then reversing direction and following speed traversing back.

                        Any insights are appreciated...

                        Comment



                        • #13
                          Here my suggestions:

                          - I feel Motion control requires separate the process different sections and attack every section different. (What is the reason is not right?)
                          - I think the Winding process is OK for you. Then "the loop" is not OK.
                          - There is the possibility the problem in "the loop" is related to Inertia Mismatch, I mean the dynamic response maybe is not good at that speed
                          - I would prefer to control the Mandrel by pulses too, and not by pot. Then I would reduce the speed of the Mandrel when transitioning to loop. Lets say reduce it to a 5%
                          - Maybe unknown but you can switch on the fly from AXSCRIPT to AXCAM on the trolley without stopping motion
                          - I'm not sure, maybe have to split the AXCAM in 2?. This due the reversing... I never tried
                          - If the galil rep can do your sequence that would be great, less work. In past projects I always had bad experiences switching manufacturers. They say the could but they couldn't.

                          Regards

                          Comment

                          Working...
                          X