Inside machines: Motion control moves bacon faster
When a New Zealand manufacturer of food slicing machines needed to squeeze more performance from its bacon slicer design, it switched from using a programmable logic controller (PLC) for total machine control to a distributed arrangement that combines a small PLC with a motion controller. Evaluating the trade-offs between central and distributed control is a common practice for machine designers. Whether centralized control architecture will work depends on the application’s requirement for determinism, the need for responding to stimulus in an absolutely predictable and timely manner. Deterministic systems produce repeatable results; the same conditions and timing result in identical responses from machine cycle to machine cycle.
For a very high-speed process to be controlled with determinism, the controller needs to be able to evaluate its inputs and respond appropriately in less than the time allotted by the process. As the need for higher productivity drives a machine to run faster, timing margins that used to be sufficient may no longer allow for completing the necessary control loops in time and the system will fail to function as desired.
That’s the problem that was encountered by Titan Slicer Ltd. of Nelson, New Zealand. The company needed to increase the cutting speed of its Titan 500 bacon slicer (Figure 1) to produce 1600 slices per minute, but the PLC that controlled all machine functions couldn’t keep up. The challenge was not in the cutting of the raw bacon slab, but in controlling how the slab is fed into the slicer. To ensure that the size of the initial slice in the package is uniform with that of the other slices, the bacon slab must be fed into the slicing blade at exactly the right time. Too soon or too late will produce an initial slice that is too thick or too thin, which is exactly the problem the PLC was having.
To resolve the problem, Titan Slicer executives called in an experienced control system integrator, Tui Technology of Rotorua and Whangarei, New Zealand. Tui’s engineer, Malcolm Jones, had worked with motion controllers in the past and came up with a new control system design for the bacon slicing machine using a dedicated motion controller (see Figure 2). A PLC was used in the new design, but for less time-critical functions, such as issuing on/off controls for the blade motor.
The motion controller is programmed to do closed-loop position control using encoder feedback from the servo drive that moves the bacon slab into the cutting blade. The motion controller sends a +/- 10 V analog output to the servo drive that moves the bacon slab. Bi-directional motion control is important so that the bacon can be moved back from the cutting position when the spinning blade is not at the right position for cutting the first slice of a new bacon pack.
Note that the blade is not round, but is instead shaped to slice the bacon as it moves under it (see Figure 3). Different versions of the machine move the bacon slab via belts above and below the blade or via a lower belt and a gripper that grabs the slab’s trailing end. So that the motion controller can track the speed and angular position of the turning blade, the motion controller connects to a second encoder, mounted on the blade motor, as a reference.
The motion controller’s fast cycle time allows the cutting operation to produce repeatable results. At 1600 rpm, there is a 38 ms time window in which to decide whether the bacon slab is to be moved forward to set up for the next cut or backward to clear the blade. The 1 ms loop time in the motion controller running the advanced gear move instruction enables repeatable and accurate slice thickness.
The advanced gear move is just one of a number of gearing functions supported by the motion controller. Gearing is used when one axis (the slave axis) must move incrementally and proportionately in relation to a master, which is typically the position or velocity of another axis. Whereas simple linear relationships can be set up between master and slave (using the motion controller’s gear absolute function), an advanced gear move operation lets the user program a nonlinear relationship between master and slave. For example, Figure 4 shows an advanced gear move operation where the motion of the slave is dictated using a fifth-order polynomial equation. Most of the motion controller’s simpler gearing commands allow the designer to choose the exact gear ratio to be used by specifying a numerator and denominator when the motion step is programmed. This avoids the need for complex software that changes gear ratios on the fly.
“To be clear, achieving a uniform bacon slice width in the middle of the pack wouldn’t be a problem for a PLC-controlled solution,” said Malcolm Jones. The motion controller “pays for itself…in the control of the first slice to meet our critical design goal.”
Besides the machine with single cutting blade, Titan also offers a higher-speed two-blade version, where the blades spin in opposite directions. The reason for the twin orbital drive is to allow very fragile or warm products to be sliced quickly. In this slicer version, each motor drive is given a 0-10 V speed command and separate head encoders provide feedback to indicate the position of each blade. As with the single-blade version, the bacon slab is advanced or pulled back as the blades turn to prevent incomplete slices.
To program and tune the motion, Malcolm Jones used motion control software, which includes a plot manager that displays trend or capture plots for key motion parameters versus time. Figure 4 is a plot produced by the plot manager that shows the operation of the advanced gear move instruction. The motion control software also includes a tuning wizard, which enables quick and easy setting of control loop parameter gains for optimizing motion.
Using the software tools, “I was able to prove that the system had enough timing margin to handle the higher cutting speeds,” said Jones.
The bacon slicing machine is a typical example of a design that requires close attention to cycle timing to ensure the desired results. In developing such a machine, it pays to select components that are optimized for performing closed-loop control with determinism.
– Brad Smith is regional applications specialist, Delta Computer Systems Inc. Edited by Mark T. Hoske, content manager CFE Media, Control Engineering and Plant Engineering, firstname.lastname@example.org.