Controlling Your Web from Beginning to End, Part II

In Part II of this two-part series, Bill Hawkins continues his analysis of the devices and designs that serve to keep your web under control.

Last month we took a close look at controlling web tension at the payoff roll, keeping the web taut, dancer rolls, and load cells. That was a good start, but there is more - after all, this series is entitled "Controlling Your Web from Beginning to End." The second and final article in the series will take you through the rest of the process.

Web Guiding from the Unwind
An edge guide is necessary on most unwind stands to keep the web centered on the machine centerline. Usually the edge guide is attached to an adjustable slide to accommodate a full range of web widths for your products.

The adjustable slide is attached to the machine frame. This guide must be set manually for each product width you want to run. The payoff roll, chucks and mandrel, brake or motor, and usually at least one guide roll are moved laterally by an actuator to satisfy the web edge sensor.

Currently, there are several edge sensors on the market that do not need to be adjusted manually for different web widths within certain ranges. These new devices can reduce changeover time significantly when you must go to new web widths. They also reduce waste by eliminating operator error in edge guide setting.

Steering Assemblies and Sensors
Steering rolls are used to keep the web centerline on the machine centerline. They are installed commonly as the first guide roll after a drying oven or any other zone in the thread path where the web may tend to tract away from the machine centerline. Web wander may be due to either induced internal or applied external stresses.

An edge guide is an integral part of the steering mechanism for web control. One of the features that a superior edge sensor has that can be very useful in these positions is an adjustable dead band. The dead band allows for a very small deviation of the edge to take place in the sensor zone without changing the position of the steering roll(s). This sometimes is needed when there is edge flutter. Edge flutter may occur because the edge is not coated, etc., or because of cross air currents in the sensor zone.

An adjustable dead band allows you to optimize the guiding for your product. Many older edge guides used in these positions have a small dead band or none, and they tend to overcontrol. Over control may cause wrinkles to form on the steering roll(s) or the following roll(s).

Often the steering system must be damped severely to prevent wrinkles. Severe damping in one zone allows the web to wander in the next zone. Sometimes the overdamped steering causes web tracking to be unstable throughout the rest of the machine, especially when multiple ovens are used. A suitable dead band will prevent over- or undercontrol of the steering system.

Steering rolls should be textured to promote good friction with the web, and the texturing should not interfere with web tracking. You should avoid any type of surface texturing on steering rolls that have any type of machine-direction (MD) grooves, even spiral-cut ones, that are to be used on light-gauge products. Surfaces of this type tend to wrinkle light products, because the MD tension tends to pull the web into the grooves. A knurled diamond pattern works well on steering rolls. Lengthwise (transverse) micro grooves also are acceptable for this application.

Tension Sensing at the Windup
Usually the web tension is more stable and much easier to sense before the windup position than after the payoff position, because the winding roll normally has less eccentricity than the unwind roll.

The roll on the payoff tends to be larger in diameter than the rewind roll, is mounted on a bladder mandrel, and has been lagged for some period of time.

And, the rewind roll usually is started on a more precision core than the payoff roll, it is chucked with lower eccentricity than the payoff roll, and the web is usually under good tension control before the rewind station.

For these reasons, a load cell roll offers some advantages over a dancer roll as a web tension sensor before the rewind, because less space is required and they are less expensive to purchase and operate.

Two other rolls are required to define a stable wrap angle around the load cell roll for accurate tension sensing. It is important to note that the roller between the sensor roll and the rewind roll must be an idler roll and not a driven roll, or the sensor roll will not measure the tension in the span before the rewind roll.

Actual sensing of the winding tension on the outside wrap of the rewind roll is not possible when windups employ any kind of contact roll on the winding roll surface. This is because the contact roll force results in some stack compression of the wraps on the winding roll. And this stack compression increases the tension in the outside winding wraps.

The amount of stack compression is a function of several variables, including thickness of the entrapped boundary air layer between the wraps; height and population of surface asperity; diameter of contact roll; hardness of contact surface; and amount of contact roll loading force. Thus, tension measurement in the span before the rewind roll must be thought of as control data.

The most successful way to find the optimum control winding tension for any one product is to use the indirect method of determining experimental winding curves based on the starting and ending web span tension before the windup and the finished roll hardness. This trial-and-error method will produce a repeatable winding process for optimum roll quality for any type of web material for any given surface asperity, thickness, and winding speed. (Refer to the "Web Solutions" column in the December 1999 issue of PFFC for a detailed analysis on how to determine these optimum winding curves.)

The load cell rolls must be rugged enough to withstand overload tensions of web wraps without losing their calibration. A large wrap angle (wrap angle >90 deg) around the load cell roll is desired, because the resultant force from the web vectors is stronger and more stable. There are several load cells on the market that meet the full requirements for windup web tension sensing.

Good winding requires negative tension taper capability as the winding roll(s) builds to a larger diameter. For large rolls, the web tension sensor must be accurate through a large tension range.

Multiple Rolls, One Mandrel
Much web production today is slit into narrow width rolls for final processing. Because narrow width production usually involves multiple rolls that are being wound on the same mandrel, there are special issues involved with the slitting and winding that are not involved when winding one roll per mandrel. Some of these issues are listed below:

  • Diameter differences in rolls due to caliper differences in web
  • Nonuniform rewound roll densities
  • Starred rolls
  • Crushed cores
  • Edge thickening, coned edges, and roll dishing
  • Stepped wind on roll face
  • Fuzzy roll sides
  • Shiners or flashers
  • Offset cores
  • Productivity
Web caliper differences across the web width will result in some rolls having a greater diameter on the mandrel as they are being wound. This problem is acerbated when making large-diameter pancake roll production with thin basestock.
  • Usually it is not possible to wind large-diameter pancake rolls on a common mandrel when all cores are locked to the mandrel, because the greater diameter rolls carry higher web tension as the rolls build. The higher winding tension tends to crush the cores on these rolls. These rolls also prevent the smaller diameter rolls from touching the anvil roll when center/surface winding, or touching the contact roll when center winding with lay-on roll. In addition, the rolls will be wound loosely.

    The varying tension problems in the slit ribbons have been addressed in the past by using the "slip core" winding technology. This involves mounting cores and spacers (which are lugged to the mandrel) on the mandrel so that the cores can be loaded by end pressure through a yoke and thrust bearing arrangement at one end of the mandrel.

    Positioning the Cores
    The cores are positioned by fixing an end spacer block to the mandrel. Alignment precision of the cores with the slit ribbons depends greatly on the precision of the cut core lengths and the operator's measurement when fixing the end spacer to the mandrel.

    The mandrel turning speed is operated at 5%-10% higher than the machine speed, so that the cores are always slipping on the friction driving adjacent spacer edge walls. This slip tends to generate substantial debris (core dust) that must be collected to prevent contamination of the ribbons that are being wound.

    Rolls are wound at constant torque when using the center-winding mode when the end pressure is kept constant. Variations of tapered winding tension can be obtained with this system by increasing the end loading pressure as the rolls are winding. Accurate tension taper is not obtained by increasing end loading when using the center/surface winding mode, because the anvil roll speed and contact pressure will override the core slip friction as the controlling factor in winding tension.

    Thus, only when individual lay-on rolls are used for each winding roll in the center-wind with lay-on roll mode, is tension taper possible with the slip core technology. This is because the common surface roll contact governs the turning speed of all the rolls and, therefore, the amount of slip on the mandrel.

    When the mandrel is dressed with cores and spacers, there is measurable clearance between the inside core diameter and the mandrel. This clearance is critical for removal of the finished rolls, because the cores tend to shrink in diameter with the radial pressure of the wound wraps. But the newly mounted cores are not concentric on the mandrel because of this clearance. And, usually there is no installed method of holding the cores concentric when winding starts with this older "slip core" winding technology. Often the first wraps on the new setup are made with varying tensions because of the mandrel-to-core clearance.

    Frequently, the first wraps on a few cores will move off the ribbon thread path until sufficient web tensions are established to keep the ribbons on track. After the starting tensions are stable, buildup on the rewinding rolls generally continues with even (flush) sides until the web caliper differences begin to make a difference in ribbon tensions due to the different rewind roll diameters.

    Sometimes the driving friction of a few rolls is very different from others in the setup. Often the wrap on a lower driving friction roll will begin to wander in its thread path. This wandering creates a fuzzy roll side face that is usually unacceptable roll quality.

    Sometimes web tension will become unsteady from vibration in the mandrel and/or machine. Many times these are short-duration tension changes but result in some of the ribbons shifting to one side or the other before shifting back to the correct thread path.

    These small shifts leave a few wrap edges sticking out of the roll face. These overhangs are sometimes called shiners or flashers. Any abrupt change in winding tension will result in the ribbon tracking in another location on the guide roll surfaces. These tension changes can be seen in the unevenness of the wound roll face. This defect is sometimes called stepped winding.

    New Technology Has Answers
    New differential mandrel technology addresses most of these problems. What this new technology does is separate the core-holding function from the friction drive function.

    Expandable friction pads engage the inside surface of the cores rather than the edges. The cores are positioned on a ruled mandrel with lugged spacers that also are held in place with bladder air pressure. Air pressure can be adjusted to the friction pads while the slitter is running to increase or decrease the winding torque.

    However, unlike the old "slip core" technology, the new mandrels give a much improved repeatability of web tensions at the same air pressure settings and are much easier for the operator to use. Thus, reliable machine settings can be determined that can be used by all the machine operators, and this will increase productivity of the machine.

    Edge thickening due to razor slitting can cause coned edges, sometimes called dished rolls. I refer you to the May 2000 PFFC supplement entitled "Slitting Process Guidelines" for a detailed review of the slitting process for both razor and shear cutting.

    This concludes our two-part look at web control. It is a complex process that requires attention to many factors, but it is one that is crucial to the quality of your finished product.


Subscribe to PFFC's EClips Newsletter