Maxcess International -Fife, MAGPOWR, and Tidland present Converting Blog

« Home » March 2006

Filed under: Winding, Unwinding, Roll Defects --- Tim Walker @ 10:22 AM

Inertia compensation is a big deal at unwinds and winders. The torque demand of a unwind or winder comes from three sources: tension x radius, inertia x accel/decel, and torque system losses (from nip, bearings, and couplings).

A question on roll inertia came in recently:

>I'm a novice to web tension control applications but I would like to learn what the proper calculations are that must be made in order to find out the inertia of a center driven unwinder and the torque required to maintain a certain tension.

>For example, let's assume a roll of plastic film weighing 1000 lb. and having a 24" diameter. The Desired tension is 1.0 PLI (Pounds per linear inch). I would like to learn how to make this calculation correctly.

Here's the calculation:
Torque from inertia is I x alpha
I = rotational inertia of the roll
alpha = accel / decel rate in radians per second^2

I = pi x rho x w x (Rout^4-Rin^4) / 2 / g
rho is density in lbs/in^3
g is gravity

[NOTE: Two sharp blog readers pointed out an error in the original post of this equation that was missing the width term. It is now correct, but I want to give out a big thanks for their help.]

I have a spreadsheet where I can calculate all this.
The 1000-lb roll of film at 24" is likely 50" wide on a 6" OD core. This crunches out to an inertia of 200 in-lbs-s^2.

What is alpha?

alpha can be found from line speed and accel time.
If you accelerate this roll to 1000 fpm in 15 seconds, the accel alpha is about 1 radians per second-squared.

So the inertial torque is I x alpha or about 200 in-lbs. At 12-in radius, this is 16 lbs.

If you have a braked unwind and the brake applies 800 in-lbs or 48-lbs of tension at 12-in radius. During the accel, the tension would increase to 60 lbs due to the added inertial torque.

Clear? Let me know.

Advanced control (what really should be standard) has inertia compensation built into the tension control plan to adjust automatically for this natural source of tension upsets.

Filed under: Guiding, Spreading, Wrinkling --- Tim Walker @ 09:14 AM

Laying out the web path of a complex web process is a combination of engineering and art. (Where "art" is more complex engineering, difficult to put numbers to.)

One small tidbit I advise it how to make equipment less wrinkle sensitive:

When your web moves from station to station, say from an unwind stand to a pull roller station, when possible, make this transition in vertical span (or mostly vertical).

Why? Because it is much easy to ensure you have parallelism in the vertical direction. Just pull out your master level and measure it. Checking for parallelism in the horizontal direction is more difficult, requiring either a pi tape, tramming stick, optical transit, or a laser system. All these options are more time consuming and potentially less accurate than a level measurement.

So why make the station-to-station transitions vertically? Within a station, you probably have good allignment. Equipment builders will make some effort to machine sideframes as a set or other tricks so the alignment within a station is generally pretty good. The biggest alignment errors often happen between units or station. If this transition is vertical, alignment is easy. Measure with a leve. Shim to parallel. End of story.

Filed under: Winding, Unwinding, Roll Defects --- Tim Walker @ 03:34 AM

Unwinding rolls are seldom perfect. They are lumpy, eccentric, and out of alignment. If you handleds your web over rollers with these properties, you'd have wrinkles city.

What to do?

The simple solution is to have a relatively long span from the unwinding roll to the first roller. This demagnifies the problem. A 0.1" runout over 50 inches is a smaller upset than the same runout over a 10 inch span.

Filed under: Tensioning, Web Mechanics --- Tim Walker @ 06:29 AM

Here is a recent question and reply on upgrading (or downgrading) a slitter tension control system.

Product: Small rolls of transfer ink on PET. Core diam 1-in., final diam <2-in..

We have an Allen Bradley MicroLogix PLC that we had to install some time ago to replace an old outdated touch screen system. We now have to replace an old outdate PID controller that was used to control the rewind tension or torque signal. I would prefer to use the existing PLC, but the integrator I brought in feels that an update time of 100msec would not be fast enough for rewind tension control. We could try it through software at a much cheaper cost instead of replacing more hardware, but I do not want to go down that path if it wil not work at all. The machine runs thin polyester film, 4.5 microns thick, 1-in cores, outside diameter is about 34mm, and machine speed is 450MPM. Any input articles, rules of thumb, etc. would be greatly appreciated.

(The dialogue posted in the first comment occurred via email, but the inquirer said he'd be happy to have it posted here for the opportunity for others to provide their input.)

Filed under: Guiding, Spreading, Wrinkling --- Tim Walker @ 08:56 AM

A post today asked:

Can you point out any major differences between stepper-controlled offset pivot guides and servo-controlled offset pivot guides?

There is a cost differential that is significant w/ large order. We are desiring to control the web to +/-.005".

First off, I'm not a EE, so I'd love to have a qualified answer to this question.

If a guide is used simply to correct a constant offset, then actuator performance is a big deal, but this is rarely the case. Webs move around and the guide needs to repond to this.

Fife put out a nice article in Converting magazine recently on web guide accuracy. You may ask a supplier, can your guide control to +/-0.005"? (a fairly common spec). The guide manufacturer can't say 'yes.' The guide performance depends on the characteristics of the error it is correct. The range of error is easy to design for, but the rate of error is the tough one. How fast does your web shift off center? For the same position change of shifting laterally 1 inch over 100 inches, for a slow speed process (100 fpm or 20 in/sec), this is a shift of 1 inch over 5 seconds or 0.2 in/s, not too bad. But this same web shift at 1000 fpm is 2 in/s and starts to push the limit of any acuation plan.

I hope we can get some more comments from actuator and motor specialist.

tjw

Filed under: Rollers, Traction --- Tim Walker @ 08:30 AM

One blog visitor recently asked:
Is there a way how we can convert the pressure on the dancer (or nip roller) to PLI?

How to convert dancer pressure (psi) to PLI (lbs/inch of width)?

I often find myself at the side of a converting process, looking at PSI supplied to a dancer roller or nip roller wondering what the created load is. The answer to both of these scenarios is the same.

Pressure x area x number of cylinders = Fc (force out of the cylinders)

Fc (Total cyl. force) x Lcyl / Lroller = Fc,eff (Effective cyl force at the roller)

For pivoting dancers and nips, Lcyl is the length from the pivot point to the cylinder load point, Lroller is the length from the pivot point to the roller's center. For linear dancers and nips, Lcyl/Lroller may be equal to 1, so the Fc = Fc, eff.

For a horizontal nip or dancer, the roller weight is perpendicular to the load on the web, so it is ignored.
N(in PLI) = Fc,eff / width

For vertical nips or dancers, the roller weight needs to be added or subtracted from the Fc,eff.
N(in PLI) = (Fc, eff +/- Wroller)/ width

What is the width?
For dancers, width is web width.

For thick webs in nips, where the web prevents roller-roller contact outside the web width, the width is web width.

For thinner webs and conforming nips, where the roller contact each other outside the web width, the width is the roller contact width. Also, depending on the ratio of web thickness to engagement, you may have a significantly higher pressure at the web due to the engagement difference.

Estimating on and off web nip variations.

Continue reading"Converting PSI to PLI, Nips and Dancers Rollers"

Filed under: Rollers, Traction --- Tim Walker @ 06:49 AM

A roller = cylindrical shell + shaft + bearing + mount to equipment.

End of story.....................it would seem.

What a simple device? If you buy an idler roller for $300 you may wonder why this sophisticated rolling pin cost the same as an iPod. Many new-to-converting design and build shops think the same thing and say 'Why don't we just build our own roller?' If your new equipment supplier tells you they plan to do this, what to do? Very quickly, write a roller specification (or talk them out of it).

Webs, especially high modulus, thin materials are sensitive animals. When your product only stretches 0.1 percent, it's is sensitive to errors in your rollers (or relative alignment of rollers) in this same order of magnitude.

If I wrote a list of 'rollers gone bad', it would go one for some time. Maybe today I'll start that list. Want to help?

Filed under: Tensioning, Web Mechanics --- Tim Walker @ 06:42 AM

Registration isn't important to all converters. So many products are designed to have uniform properties in both MD (machine direction) and CD (crossweb direction).

Many converting process have a need for lateral registration. You need a gross registration to stay on your rollers, but you likely have a more challenging registration to hold a narrow coating or trim slitting margin. We usually don't call this registration, we call it web guiding (or chasing).

Registration usually involved pattern forming processes, whether printing, die cutting, or embossing. Each of these may be a stripe, needing only lateral registration, or full MD-TD registration of a 2-dimensional array of patterns.

For the live blog today, let's take questions in any issue of registration, including equipment, product, or process questions.