Web Handling & Converting Blog

Hydraulics vs Pneumatics

There are some machine components in web manufacturing and converting that could either be powered by pneumatics or hydraulics. The most common is the nip between two rollers. However, you will find other applications such as lifting/lowering/pushing, unwind brakes and even motors. It is true that hydraulics are more common for bigger machinery, such as found in the paper industry. However, their merit extends far beyond mere brute force.

The advantages of pneumatics are primarily just two: clean and cheap. The advantages of hydraulics are so much more extensive that we often prefer them when there is a sensible option. Hydraulics compared to pneumatics are:

1. Strong - 10X stronger
2. Precise of force – more than 10X more precise
3. Precise of midstroke position – 1000X more precise
4. Fast control response – perhaps 1000X faster (however, stroke speeds are similar)
5. Vibration resistant – for example, can cut winder vibration in half
6. Safe – pilot operated check valves on cylinders prevent release of stored energy

Dual Unwind Paths

Dual unwind paths should be standard. By this we mean that you can either unwind from the top of the roll or, by flipping the roll around, from the bottom. Even if you never think that you will never need to invert the web you should be able to do so any way. It gives you the flexibility to do future products not yet anticipated. It gives you the ability to do some troubleshooting. Finally, it is trivial to do in almost all cases if you think ahead, so why not?

Slitter Dust Cleaning

How can I remove slitter dust from my web? Are tacky rollers or vacuum systems better?

My mom and industrial engineering teachings are in complete agreement on a couple of points. (1) It is best not to make messes in the first place. (2) If messes are unavoidable, or s___t happens, clean it up as thoroughly and as soon as possible otherwise its just gonna spread and make even more work.

So in the spirit of (1), please go to slitting school, apply what you have learned and then come back in a year if every effort at reducing mess-making is not good enough.

If I had a Hammer

New beginnings. What would I do ‘if I had a hammer?’ I would hammer down every gage band on every paper, film, foil, rubber, nonwoven and textile manufacturing line. While I was at it, I would hammer down the coating streaks, thick spots in extrusion and even printing ink. I would make webs everywhere as level as needed to satisfy the fussiest customer of all time, the wound roll.

Why? Well, I know that one of the top three problems facing (most of us) in the web industry is wrinkles and related defects. I know that bagginess is a major cause of wrinkling. I also know that gage variation is almost always the cause of bagginess on film and foil and is a major contributor elsewhere. I also know that gage variation causes many other quality issues and converting problems. It may be the largest root cause mechanics of troubles in our industries.

I would NOT start out my level pogrom with fine profile measurements and automatic gage controls. I would start so level on every component that these ‘patches’ to the root problem would not be needed. Indeed, some profile controls are in fact destabilizing and make the problem worse than it would otherwise have been. The reason people don’t realize this is because they paid big bucks for these systems figuring they would be helpful and never turn them off to find out for sure. The paper people who have the most expertise and experience have long known that even in the best of circumstances the operator usually has to take manual control over certain actuators (usually the ends of the machine) because she can do a better job than the computer.

Tension Control Combines Modeling and Measurement

by Clarence Klassen cklassen@klassen.on.ca

Common questions regarding web drives are “Why do we need a model of the tension as well as a tension regulator? or Why can’t the tension regulator alone control tension?”. We need both – the model gets the tension close and the regulator (which measures the actual tension and corrects it) provides the fine control.

Or take ships as an example. Although they are so large and are driven by strong winds, they are steered by a very small rudder wherever the pilot wants to go. James 3:4

The sails are set, the wind is blowing. The captain does not normally muscle the wheel to keep on course. But it’s still nice to know the captain has a wheel and rudder. We don’t presume to sail around the world, but we do need to provide accurate tension control in our web handling process.

There are many aspects of web handling that are well understood. These aspects can be modeled. The model can provide tension that is very close to the tension setpoint. In fact, 50 years ago the model and knobs available to the operator provided the only control of tension.

The model may be in the form of Tension as a result of Torque or Tension as a result of Speed. The tension into torque model is generally more accurate than a tension into speed model. Measurements taken while commissioning set the parameters used by the model. Let’s examine what contributes to a good model of tension as a result of torque.

1. The tension setpoint divided by radius gives a torque value.
2. The static friction in bearings, belts, bushings, D bars can be modeled into torque.
3. The rolling friction in bearings, belts, bushings, D bars, air foils and tables can be modeled into torque.
4. From Newton’s laws for rotating bodies, we can model the torque required to accelerate rollers during a speed change.
5. Process disturbances such as application of coatings and splices can be modeled into torque.

No matter how good the model may be, it will not be perfect. A tension regulator which measures actual tension can provide small course corrections. The output of the tension regulator will be a torque signal. If the model is perfect, the output of the tension regulator will be zero or near zero. If the model calls for too much torque, the tension regulator will integrate negative, reducing the torque to the correct total value. The tension regulator typically needs less than +/-10% authority to vary torque.

The torque signal from the model acts immediately, there is no delay in the calculated torque. In fact, the model can predict the torque required to run a particular product next Tuesday afternoon. While the torque model is an immediate signal, the tension regulator reacts after the fact. There is a lag between a tension disturbance and the response of the tension regulator. This lag can be reduced with a fast responding tension regulator.

The tension regulator does not work well at low speed. Research into improving slow speed tension regulators is ongoing. The model is required to achieve tension during threading and other slow speed operations.

A tension regulator could be designed to work without a model. The following non-trivial difficulties would have to be overcome:

1. We need a fast tension regulator response (>10 radian/second resulting in a 0.1 second lag). The regulator must be fast to handle transitions such as the application of coating.
2. We need a means of ensuring that the tension regulator performs at low speed. Some research in this area involves using tachometers with 2,000,000 Pulses per Revolution (PPR). Typical tachometers have 2,000 PPR. Sometimes low speed operation is not a concern as the line is threaded and splices are made at high speed.
3. Good slip detection and web break detection is required to prevent over-speeding the driven roller.

THE Big Problem With Automated Data Acquisition

THE big problem with automated data acquisition is that it is almost always time-based sampling. However, our internal rejection (usually based on top of roll samples) and customer returns (roll # such and such) are lot based. Until data acquisition is constructed from the ground up to be lot based binning of time based data, our analysis will be slow, clumsy and crippled. (For the database aficionados such as myself this means a relational database instead of a ‘flat filer.’) If you don’t think this is the case, merely answer the question “Are products that are on the low end of acceptable basis weight made last month more likely to cause customer complaints?” and you will see just what a mess the typical systems are. Days of cut and paste before you can answer the simplest question.

Speed or Width

Which is an easier way to increase productivity when designing a new machine: increase speed or increase width?

Easy, speed is easy, at least from a web handling point of view. We can go as fast as your little heart desires with only two limits: air entrainment and vibration. The paper industry has already conquered both and has achieved 12,000 FPM production winding speeds. Thus, if you wanted to double the speed right now on a new machine, web-handlers would merely yawn in most applications.

However, if you wanted to increase web width by a mere 20% from your industry’s previous experience, we would express caution if not raw fear. Roller deflection, vibration, waste and other troubles will all increase far more than proportionally and a countering response is necessary. The most fearful of all width increases is to go from a 1-wide to 2-wide machine that makes side-by-side lanes. I would never promise management anything until it had been done successfully elsewhere.

Wide Width Ranges

Also seen in some web manufacturing but much more commonly in web converting is to ask for a range of widths. While some request only a 20% width range capabilities, a few will be running a 2:1 width range. The tension issues we described above certainly apply. Additionally, we have to tell the drive what width we are running so that a correct tension (tension load per unit width of web) can be run. This makes the operators life slightly simpler as they only have to vary tension for grade (thickness, width etc) but NOT width because the drive/PLC does that compensation.

With properly designed web handling components, the web will run equally well to the front, centered or to the back. This even includes spreaders. (Consider the spreaders used after slitting a web into separate lanes. Most lanes are running ‘off center’ anyway.) Thus, we often keep the back side fixed, especially on winders, to make setups quicker and easier.

A more severe challenge comes from precision processes (as opposed to web handling). A good example to consider would be nipped calender rolls. Here you need to run centered otherwise the resulting deflection is asymmetric and the load would be uneven. Also, calender rolls wear. Thus the rule that you can go narrower without too many issues, but you must consider regrinding the very first time you increase width. If you don’t, the wider product will be running up on the less worn ridge of the roller(s) and carry a very intense nip there. These are just a couple of issues with only one converting process. There are many others that are no less sensitive to product width changes.

Wide Tension Ranges

Everyone wants their machine to cover a wide range of products. Beware of what you ask for, you might not get it. Even a 4:1 gage or weight range is not at all trivial if you give both ends of the product range the 2.5:1 tension range that each would like. (This comes from the rule of thumb of sizing equipment for 10-25% of web strength). These specifications would indicate a 10:1 tension range, a pretty big challenge already. Now consider a center-wind or center-unwind with a 10:1 build-down ratio (such as a 40” shipping roll on a 4” OD core) and you are requiring the drive to be able to handle a 100:1 range of torques.

The challenge is not so much on the high end; pulling hard is easy; just make motors etc bigger. The challenge is always on the low end and it is one of quality. Maintaining steadiness at light tensions and ultra-light torques will challenge the best mechanical and drive designers. You know when you’ve gone too far; hunting of tension during accels/decels and/or especially at light tensions near the core of unwinds and winders.

There are a couple of tricks to make this work better. First, roller bearings and other sources of drag are sized for light products but roller diameter is sized big enough to handle the heavy weights without excessive deflection. Second, load cells are made dual range by having two wrap angles. A heavy wrap for light tensions and a light wrap for heavy tensions. The wrap strategy requires one extra roller and instructions to the operator on how/when to thread each of the two load cell area paths and to tell the drive/PLC which one is being used (so that the proper gain is calculated). Finally, the center wind or unwind motor. We might supply two drive motors, one on each end of the core shaft (if there is room). We might, for example, use a 25 HP for heavy grades but mechanically disengage it and use the 5 HP motor on the other side for light products. The big motor would have to be completely disengaged to avoid parasitic drag of a motor and gearbox back driven by the web.

Name That Defect 07 - MD Tension Upsets due to Idler Roller Drag and Inertia

Description. This schematic depicts the upsets in the MD (down-web) tension profile due to the combined effects of idler roller drag and inertia. Executive Explanation. The middle slightly inclined tension profile (shown here in black) is at steady state, with the steeply rising profile occurring during acceleration (green) and the least upset down-web tension profile for deceleration (red, where bearing drag causing a downweb tension rise is partially offset by a down-web tension decrease due to inertia.) A brief write-up can be found pp 369 of the must-have Roll and Web Defect Terminology book. Tim Walker also discusses aspects of this in his classic paper Practical Application of Idler Roller Performance Measurements and Models, Proceedings of the 5th Int'l Conf. on Web Handling (IWEB) Web Handling Research Center. Here Tim demonstrates how to measure bearing drag to calculate a safety factor for traction (one concern for excessive drag, the other is accumulated tension rise/error as discussed here). A complete treatment of drag and inertia including example problems and calculations to be used for design and troubleshooting of machinery is given on pages 58-61 of my Mechanics of Rollers book. [At this time a brief digression may be useful. Note that all of the detailed sources mentioned here are NOT available on the Internet. It is still the case that the Internet has only a small fraction of the articles and books carrying the best of our web science and technology.]

Remedy. At the design stage, first reduce idler roller count as much as possible. Second, reduce diameter as much as possible while still conforming to maximum allowable deflections (usually Class B for most paper, film and foil components). Third, change the roller shell such as reducing wall thickness or going from steel to aluminum or to carbon. Finally, you can always add a drive section or helper drive, AS LONG AS YOUR DRIVE GUY KNOWS WHAT HE/SHE IS DOING. Adding a drive to fix a drive problem is almost always an act of desperation except in those rare cases where a knowledgeable person corrects the error of a predecessor. There is a saying that “it takes a whole lot more smarts to fix a problem than it did to make it in the first place.” Where is this extra-super smarts going to come from?