Announcement

Collapse
No announcement yet.

PLC conversion of an injection molding machine

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


  • PLC conversion of an injection molding machine

    After searching here and across the internet, I find that there is little detailed information about this subject. Here is my questions:

    1) Is the DL205 system a good choice, and what CPU/control card is best?
    2) I have an EA7-T15C color touchscreen. Is that a good choice?
    3) Did see a bit of a forum discussion on Stage programming for this application vs. direct ladder logic, pros and cons? I am comfortable with both.
    4) I am very familiar with the molding process and function of a molding machine. I have detailed electrical and hydraulic prints of the original control system. I also have a detailed OEM timing chart. What I do not have is access to any of the original program (I believe it was a Seimens controller). Any tips from anyone that has converted an injection molding machine to a new PLC controller?

    Thanks
    Les Klar


  • #2
    You said you have a timing chart. How fast does the response time need to be? are we talking milliseconds? or microseconds?
    How many encoders? Or are the positions all analog?
    Circumstances don't determine who we are, they only reveal it.
    Jason Wolthuis

    Comment



    • #3
      1) The 205 is a good platform, I would use the Do-More processor. I did a retrofit on 4 blow mold machines not too long ago. The Do-More nailed the temperature control auto tune the first try, controls +/- 1 degree F and the CPU scan time is much faster than the SLC that I replaced.

      2) The EA7 is a good touch screen (If I were buying new I would use an EA9) I have install a lot of them (over 100) in customers plants. I have had 2 failures in the last 8 years or so 1 was a lightning hit and the other one had a surge on the Ethernet line from the ISP and took out the router, switches and the Ethernet port on the touch screen.

      3) I prefer to write my code in Program blocks, Tasks and Stage. Now with DMD 2.0 you can use subroutines. They all have their place for program control and ease of following the logic. There are videos on the AD web site that will help with how and where to use each.

      4) I know on the Blow mold machines temperature control is a big deal so I put each PID loop (22 zones) in it's own program block. I had the old logic so the conversion was fairly easy. I did add a lot of safety devices to bring the machine up to current code (not PLC related but a required part of the retrofit) and troubleshooting logic to the plc/touch screen to help the operators and maintenance.

      JW

      Comment



      • #4
        LesKlar, keep us informed how it is going. that is a huge undertaking. We have over 60 molding machines. Each machine has many pages on the HMI with lots of setting on each page. Molding machines are very touchy piece of machinery. Would love to know how you are doing. It maybe something we might have to to down the road. And yes Seimens seems to be the plc in most of them.

        Comment



        • #5
          1. The 205 should be fine. I haven't used the Do-More so I'll leave that input to others. I have used the P3000. It would also be a fine choice for this (as well as the P2000 I would assume). I've done several with 405s. I only used that platform because of the 4 loop temperature control card that is available. I was pretty new to this and not having to deal with PID loops along with everything else was a bonus. Once we did one, we just stayed with the same platform.
          2. EA7 is a fine choice, but agree with jwbaker3 that the EA9 is a better choice if for no other reason than support should last longer for it.
          3. I did mine with ladder logic, because at the time Stage programming was a foreign to me. If I had to do it again, I would probably still use ladder. I used a lot of holding circuits in my logic just so the machine would drop out of cycle if certain safety limits were breached. Some hate holding circuits because they are harder to troubleshoot so I may get pushback from others on this comment.
          4. You'll find that the OEM sequence/timing chart will be the most important thing you have. Some OEMs do it better than others.
          5. Use the logic to control functions of the machine (i.e. control bits/relays) not the discrete outputs. Put the inputs and outputs in a separate section of logic, in another subroutine, etc. The general sequence of operations of most machines are the same. Doing this will allow you to reuse most of the logic for a different machine later, even if built by a different OEM.
          6. Get a copy of ANSI/SPI B151.1-1997, Plastics machinery Horizontal injection molding machines Safety requirements for the manufacture, care, and use (or B151.29 for vertical machines). There are some very detailed standards regarding guarding, gate safeties, hydraulic safeties, mechanical (jamb bar) safeties, etc. and they can change depending on the tie bar spacing of the machine. If you're familiar with injection molding machines, the standard is pretty easy to read.
          7. I would install an SPI robot interface even if you don't have a robot on the machine. It's not that much more trouble and can save you a ton of hassle later if you decide to add one. Robot interfaces that aren't integrated well into the controls are a PITA to troubleshoot. The standard for this is SPI AN-116.

          Comment



          • #6
            One other thing I thought of later... For run-off, I used the simplest mold I could find with no ejector rods installed. I used a small cylinder on a table with two end of stroke limits to work out the core sequences. That was by far the hardest part of the logic.

            Comment



            • #7
              Originally posted by LesKlar View Post
              After searching here and across the internet, I find that there is little detailed information about this subject. Here is my questions:

              1) Is the DL205 system a good choice, and what CPU/control card is best?
              2) I have an EA7-T15C color touchscreen. Is that a good choice?
              3) Did see a bit of a forum discussion on Stage programming for this application vs. direct ladder logic, pros and cons? I am comfortable with both.
              4) I am very familiar with the molding process and function of a molding machine. I have detailed electrical and hydraulic prints of the original control system. I also have a detailed OEM timing chart. What I do not have is access to any of the original program (I believe it was a Seimens controller). Any tips from anyone that has converted an injection molding machine to a new PLC controller?

              Thanks
              Les Klar
              what brand molding machine are you converting?

              Comment



              • #8
                Thanks everyone for the tips. I am still putting together the proposal for management, so I will include the cost of a new EA9 15 inch screen.
                To answer a few of the above questions and comments:
                1) This is a Van Dorn 230 ton HT. This uses analog position sensors for the clamp, ejectors, and screw. Uses a proximity switch that counts rotations of the screw. The pressure and flow are handled by a couple of Rexroth amplifiers that require an analog signal to tell the amp the desired set point. The amps handle all the controls for feedback.
                2) I am also planning on updating the safety circuits. Older machines rely on hardwired relays for the safety circuits. I am going to upgrade to modern dual channel safety relays. I also will be adding a few additional (required) safeties. For example, a redundant switch to monitor safety drop bar position.
                3) The machine already has the SPI so this will be integrated into the retrofit. I have the SPI publications for the machine and the robot. I am also going to integrate a process monitor signal interface (RJG). This will allow easy hook-up of sequence signals that most second party monitoring equipment is looking for.
                4) Pressgrove makes a very good point. I always separate the physical I/O into a separate routine or stage. It makes trouble shooting a lot easier.
                5) I have a ladder logic program from a different machine (Mitsubishi PLC). This is not exactly going to transfer as the machine has a different hydraulic structure (uses a variable volume pump with a swash plate), but the basic sequence is the same.

                I expect approval sometime next month, I put this as a month long planning and designing, and then 2 -3 weeks to implement (provided no parts are backordered). This means I should have a functional machine sometime in June. I will try to update this as I go, and I am sure to have questions.

                The one area I have not had much experience in is temperature loops. I did a two channel set-up for a hot stamp machine, but this had a very fast heat/cool time so tuning was a snap. I read on this forum that some of the long heat times (typical of an injection molding machine) can be a beast to tune.

                Thanks
                Les Klar

                Comment



                • #9
                  205 with a Do-more should be a great fit.
                  Circumstances don't determine who we are, they only reveal it.
                  Jason Wolthuis

                  Comment



                  • #10
                    That's awesome. We have the same machine at our shop but its a 400 ton. We are starting to have problems with the S5 (always going into "pc in stop") and we were thinking about doing the same thing. Only thing stopping me from pushing on with it is I do not have enough skill with analog programming from all the sensors and out to the valves. Hope it goes well for you. Hope you post updates and pics/vids of it.

                    Comment



                    • #11
                      Would be very interested to see how your PLC replacement turned out for your Van Dorn HT machine. We have several HT's and wanted to see if it was a succes? please let me know

                      Comment



                      • #12
                        I thought I would give a quick update. I abandoned the DL205 PLC and moved to the P2K. I started the design and initial programming. I set-up the barrel heating part of the program and tested. With a little tuning, this worked very well. The P2K is super easy for PID temperature control. Then use the time proportioning algorithm to cycle the heater contactors. I did cheat on the feed throat temperature control (this is not a critical control point). I used simple ON/OFF control and let this swing +/- 2F. I may convert this to PID, but the water cooling deviates this temperature reading and the scaling we have in our plants also makes this a hard control to center. Also, reverse engineering a few other model molding machines, it appears that is the control method that they are using.
                        My next point in the project is to set-up the analog inputs. The trick I have to investigate is reading a position and controlling a flow (velocity parameter for linear movement). I have to use change in position to derive a velocity signal input for the PID loop and then scale this to a flow control output. I also have to set-up ramping functions for the starting and stopping of the hydraulic movements. It would not be good to suddenly reach the desired position and slam the valve closed.
                        My other thought on the hydraulic control is to use a Rexroth control card and let the card take care of movements. There are two specific cards designed exclusively for clamp control and injection control. I bought them off eBay and they power up and I can communicate using BODAC Rexroth software. The challenge is that these cards require the parameters to be entered to function. Again easy to do through the BODAC software. But I need to do this with the touch screen. These cards can communicate to a PLC using DeviceNEt. Does anyone know of a good Modbus to DeviceNet converter? I would like to use the P2K Modbus as the master and the Rexroth cards as the slaves. This solution would be very easy to implement.

                        Comment



                        • #13
                          Originally posted by LesKlar View Post
                          I abandoned the DL205 PLC and moved to the P2K.
                          Which 205 CPU were you using?

                          Comment



                          • #14
                            Originally posted by LesKlar View Post
                            I thought I would give a quick update. I abandoned the DL205 PLC and moved to the P2K. I started the design and initial programming. I set-up the barrel heating part of the program and tested. With a little tuning, this worked very well. The P2K is super easy for PID temperature control. Then use the time proportioning algorithm to cycle the heater contactors. I did cheat on the feed throat temperature control (this is not a critical control point). I used simple ON/OFF control and let this swing +/- 2F. I may convert this to PID, but the water cooling deviates this temperature reading and the scaling we have in our plants also makes this a hard control to center. Also, reverse engineering a few other model molding machines, it appears that is the control method that they are using.
                            My next point in the project is to set-up the analog inputs. The trick I have to investigate is reading a position and controlling a flow (velocity parameter for linear movement). I have to use change in position to derive a velocity signal input for the PID loop and then scale this to a flow control output. I also have to set-up ramping functions for the starting and stopping of the hydraulic movements. It would not be good to suddenly reach the desired position and slam the valve closed.
                            My other thought on the hydraulic control is to use a Rexroth control card and let the card take care of movements. There are two specific cards designed exclusively for clamp control and injection control. I bought them off eBay and they power up and I can communicate using BODAC Rexroth software. The challenge is that these cards require the parameters to be entered to function. Again easy to do through the BODAC software. But I need to do this with the touch screen. These cards can communicate to a PLC using DeviceNEt. Does anyone know of a good Modbus to DeviceNet converter? I would like to use the P2K Modbus as the master and the Rexroth cards as the slaves. This solution would be very easy to implement.
                            Modbus to Devicenet, take a look at Anybus.com Many third party converters use their chipsets. I've had good results with Modbus to ASI converter I bought a few years back.

                            Comment



                            • #15
                              I've done many IMM's control retrofits with AD plc's. (05, 06, 205, 405)Just depends on how much you want to spend. Van Dorn's are my speciality. Been working on them for over 40 years. I find sometimes retrofit kits developed specifically for IMM's are a cheaper alternative to reinventing the wheel.( Barber-Coleman, SCI, Keba, Sigmatek) My only issue with using the 205 series is temp control. The autotuning sucks as the thermal lag is to great for the loop to tune. That was a while back. Maybe AD revised the temp module. Seems to me your overthinking the speed control. Most Rexroth driver cards have built in ramping which can be turned off and on by a simple discrete command. As far as your velocity control goes, it appears you are trying to go closed loop. The HT machines have proportional valves for clamp and inject (CN and XH) with 1 prop psi valve for entire machine (G) all which are easily commanded by 0-10V analog input signal. Back in the early years of Do-More, it was Think and Do then I tried to work with AD and develop a retrofit IMM controller package. They did not want to play. Have fun.

                              Comment

                              Working...
                              X