Web Handling & Converting Blog

Tucker Pucker

My last two consulting visits involved a common problem; an ugly start of the wound roll. In both cases it is what I labeled a tucker pucker. Specifically, the tail of the web was not dead level and taut and taped to the core in its equilibrium position (as opposed to where the core happened to be). In the first case it was an automated set change on heavy paper that didn’t get the tails dead flat on the cores. In the second case it was winder operators in the textile industry who had always just tossed the tail over the core with no more concern than that it took to get the winding rolls started. The first was a machine tucker, the second was a Manuel (manual) tucker . Two different winding machines, two different materials from totally different industries and two different systemic errors (machine design in the first case and culture in the second); yet the same result; ugly roll edge offsets at the start of the wound roll.

What happens whenever a splice or tail is not perfectly aligned is that the web does not enter the roller perpendicularly. The normal entry law requires the tail then to move toward its equilibrium position in a decaying exponential. However, heavy materials tend to overshoot equilibrium position. Thus the web may shoot past and then comes back again in a decaying oscillation about the equilibrium point.

The moral: don’t pucker the tucker.

Friction Limited Thrusting (in Nips)

Ever wondered why wound roll edges are rough? Ever wondered why bearings on nip rollers don’t seem to last as long as others do? The answer to both may be friction limited thrusting. Few people know that a mutual thrust load is developed when two cylinders are in nip. This could be in the roller-roller nip of a calender, laminator or printer or the roll-roller of a winder. This thrust load is caused because the world is not perfect. Specifically, the diametral profile of the rollers (alignment etc) is not perfect. This causes the two rollers to act like a cone drive which mechanical designers know causes thrust.

The thrust developed is not very dependent on the size of the error; the tiniest of errors can cause similar thrusting as the largest. The thrust is in fact friction limited to the normal (nip) load times the coefficient of friction between the two rolls. Thus, if you had a modest 1 PLI on a 50” wide film winder lay-on roller with a COF of 0.5, you would develop 25# of side load. On a wide paper winder, thrust loads have been measured at #10,000 lbs. Thrust loads on calenders can even be higher because the nip loads are higher on calenders than on almost any other nip in the web industry.

The problem is first with the bearings and frameworks. Every designer knows that almost all bearings (except thrust bearings) used on rollers do not take side loads very well at all. One (new) calender builder did not realize this and sized loads only for the obvious radial loads. The result was a 24 hour bearing life with a bearings that cost several thousand dollars each. Also this side load tends to spread framework that can cause a wide variety of problems.

On winders we have a different issue. Here the side loads can cause the roll edges to dish. This looks something like an offset or telescope. It happens right near the core when side loads first develop and then push the (wound) roll one way and the (nip) roller the other way. Curiously, the more level you get the web profile and machine alignment the worse the problem becomes. If things are way out of whack, the winding set will immediately thrust to one side and stop against the bending frame. If, however, things are nearly balanced, the set may toggle from one side of the play/flexibility band to another causing offsets throughout the wind.

The moral: keep the frame stiff and ball (roller) bearings strong when engaging nips.

Memory, Tarzan and the Science of Spirals

Memories can fade unless they are accessed and reinforced. This is why we can no longer speak French or do the calculus. However, memories of events are different from memories of facts. When events are recalled, they are not only reinforced, they are also modified. Recent research shows that every time we recall an event, the recall is modified just a bit so that eventually it forms a narrative. Humans need a story. A string of observations of facts taking place over time does not construct a story. Unfortunately, to construct a memorable and favorable story we need to fill in the gaps between the facts and, almost certainly; adding to, leaving out or changing the facts.

Everyone knows the memorable first words of Tarzan to Jane; “Me Tarzan, you Jane.” Unfortunately, this is not at all the case. I have begun collecting rare books, such science books, the entire Tom Swift series (something like Hardy Boys for engineers and scientists) and, most recently, Tarzan. Out of curiosity I read Tarzan of the Apes and was absolutely amazed. It was nothing like what I imagined based on the early Johnny Weissmuller TV series, cartoons and other recent fare; it was far better. In fact, the first words of Tarzan with any other human being were written English (he taught himself to read from kids picture books) and to Jane was in French. Tarzan’s first words to Jane Porter were “Je suis Tarzan des singes” or “I am Tarzan of the apes”. (Thank you Babelfish.) Tarzan was just one of dozens of books by the truly interesting Edgar Rice Burroughs.

So, what has this all to do with web handling? I was recently teaching a class where a student objected to my claims that spiral grooving or spiral taping does not spread. He said he had a very reliable colleague who had found otherwise, as is the conventional wisdom. This colleague had supposedly turned a roller around and caused wrinkling. Unfortunately, belief and hearsay have no standing. It matters not whether 50, 500 or 500,000 people believe something, such as the spiral grooving fallacy. It matters not whether a story is credible. All that matters is peer-reviewed science where an idea is published and the public tries to find fault or exception. All published research says the same thing; spiral grooving or taping does not spread.

Science keeps us from twisting facts to make a story. Science keeps us honest. It requires us to check the sources which are most desirably replicated experiments. In a pinch we can use experts. We would truly be desperate if we used belief or testimony

However, science is not a creed or list of facts. It is a self-correcting system. If but a single sound experiment shows that spirals spread, we must modify our theories accordingly. (I hope not for Ron Swanson, myself and dozens of other expert's sake.) If we find but a single rabbit fossil mixed in a bed of dinosaur bones, we must rewrite all of our biology and geology books. (I hope not for the tens of thousands of scientists sake).

Winding Machine Knobs versus Process Knobs

Winding also is a one knob process even though some winders may be equipped with as many as four knobs on the machine that control roll tightness. The TNT’s of winding are Tension, Nip and Torque. The small ‘s’ stands for speed because only some products, primarily film at speeds above 100 FPM and paper at speeds well beyond 1,000 FPM are speed dependent due to air entrainment.

Thus, Tension makes the roll tighter. Nip also makes the roll tighter and, you guessed it, center-surface winder Torque differential makes the roll tighter. Conversely, speed makes some materials looser. Many knobs, one result.

Knowing this allows the process developer to make tradeoffs. If tension is giving her grief, such as via web breaks, she can back off on the tension and add nip in its place. At the end of the day, the roll could be equally tight but without as many web breaks. Conversely, if nip is giving here grief, such as many wound roll defects are prone to, she can back off on the nip and crank up the web tension. Provided that you could move far enough, the wound roll would be the same except with less nip-induced defects. The process developer knows also that if the winder is sped up, one or more of the other knobs must also increase else the wound roll will/must become looser.

Knowing this allows the process developer to avoid silly promises to the boss and customer. For example, having tight and loose defects under the same conditions can not be effectively remedied by changing settings. You can not, for example, loosen the roll with the tension knob to avoid gage bands (a tight defect) and tighten the roll with nip to keep the wound roll rounder (a loose defect).

To the Jedi Knight. Much wisdom is contained in the brief discussions in the posts these last two week. Ponder the significance for your situations.

Calendering Machine Knobs versus Process Knobs

There are many more ‘knobs’ or things that can be adjusted on a machine than there are independent parameters for that same process. Take a calender for example. You may increase the calendering effect (however you measure or define it) by increasing load, increasing cover hardness, decreasing calender roller diameter, decreasing speed, increasing temperature/moisture (for most products) and using multiple nips such as in tandem or stack calenders.

This number of adjustments is, however, beguiling. It makes you think you have much more control than you really do. The reality is that calendering as a process is only a one knob process even though there are many knobs on the machine. All of the knobs listed above essentially do the same thing; increase calendering.

If you know this you can trade one knob for another. For example, if cover life is short you can increase cover hardness and correspondingly decrease nip loading. In many cases the product can end up identical with the only difference is improved cover life. If you know this as product designers and process troubleshooters you can avoid trying to chase mutually opposing goals and think you can achieve them. You would then know better than to, for example, try to increase laminator bond and still maintain bulk (for malleable materials).

IWEB X - Down a bit

The WHRC (Web Handling Research Center) at Oklahoma State University is the birthplace and cradle of the science of web handling. It held its 10th biennial international earlier this month. Though the attendance was down a bit, the quality and quantity of papers presented were as strong as ever.

Neal Michal and John Shelton both gave what may be the first and only substantial papers on accumulators. The consensus was that current designs are primitive and troublesome. The control of alignment and tension is much worse than other web handling components and may be directly responsible for a host of problems such as necking, path control issues and wrinkling.

Wrinkling was, as always, a topic of great interest. Dr Keith Good, the lead researcher at the WHRC, gave a keynote address on the subject. Tim Walker discussed the challenges of semantics (word definitions) and taxonomy (classification).

Other topics included baggy web modeling and measurement (the near consensus is that the web moves to the long side contrary to intuition), a new sensor for tension profile (baggy web measurement), a curl model and a host of papers on tension control and other subjects.

CMM - and out?

In its heyday a decade ago, CMM filled Chicago’s largest convention center with 700 exhibitors and 30,000 attendees. This year at the Rosemont there were only around 100 exhibitors with very few attendees to attend to. Some of the classes had no participants. Sure, everyone has taken a beating (perhaps even some of the Wall Street banking execs who siphoned off our public and private money by selling worthless paper.) However, some question whether CMM is viable if and when the economy comes back even though it sold something quite useful.

Converting Wound Roll Measures

There is only one kind of temperature. Yet there are scores of different measures of temperature: Fahrenheit, Celsius, Kelvin, Rankin, mean free path of a molecule under STP, color of an incandescent body, etc, etc, etc. You can convert any measure to any other measure via a (usually) simple equation.

Exactly analogous is wound roll tightness. There is only one kind of tightness. Yet, tightness can be measured or defined in scores of ways. This includes the TNT settings on our machine, WIT, radial and tangential stresses, roll geometry (such as length in a given diameter), roll density, hardness (by 4 different instruments) and so on and so on. This also includes other units people don’t commonly think of such as air in a roll of film or transparency of a wound roll of film.

You can convert any measure to any other measure with a monotonic curve. Thus, in principle, if you know the transparency of the roll of film at a local area; you know its relative gage, the amount of air there, the roll’s hardness, the stresses inside the roll and so on. Thus, knowing one thing in principle means you know all because winding as a process is a one-knob process.

The rub, however, is that the conversion factors are not simple equations but an empirical curve. Furthermore, the curve is a ‘material property.’ Thus, the conversion from Rhometer roll hardness to Wound-In-Tension and vice versa is different for newsprint as it is for LWC paper as it is for 50 gage polyester. In fact, it is in principle different for nominally identical newsprints (etc) made on different machines.

Still, the web handler who understands this can look at the most convenient measure to infer what is of most interest to him.

Proper Units

We all want to be proper web handlers. We all want to avoid the ‘Tower of Babel’ problem where each machine or plant uses a different language (units) for controls. We want to be able to communicate with each other and with our machine builders. To these ends, we must use proper units.

In the English System*, the proper units for web strength are lb/in or PLI (Pounds per Linear Inch of web width). The proper units for web tension are PLI (not total pounds or dancer pressure in psi). The proper units for nip load are PLI on the product (not cylinder pressure). Finally, the proper units for motor torque are PLI at the roller’s surface (not amps or lb-in). Thus the proper units for the TNT’s of control are all the same: PLI.

Similarly in the metric system the TNT’s are defined as force per unit of web width. In this case we often use kN/m.

*My sincere apologies to the English readers who have not used the ‘English’ system of units for decades. Rather, we should rename these bastardized measures what they really are. Let’s henceforth call them the ‘American/Bangladesh/PapuaNewGuinea system of confusion’ because almost no one else uses them.

Calibrating Nips

Nips are just like any other control; they must be calibrated. Unfortunately, few do. This means that we don’t know where we are at. It also means that we can’t necessarily repeat a successful run on a different machine or even this machine a year from now.

First, we must use proper units as described in the next post. Second, we must define/measure zero as described a couple of posts ago. Third, we must calculate gain. Fourth, we must check gain (more than once due to hysteresis described a couple of posts ago. How to do this depends on the machine. However, it is usually done by pulling against a tension load cell (simple shipping or crane scale may work) strapped to a fixed surface such as a beam or by a compression load cell between the moving and fixed rollers. Finally, this process should be repeated on a regular basis, perhaps annually, for critical processes.