# Winder algorithm

M

#### Marian Stolarski

Hi,
I have a question. Does anyone have a generic winder control algorithm that commands the dc motor control to slow the speed of the shaft down
as the diameter of the winder increases?

Marian

W

#### Willy Smith

Marian,

I did this once with cones of nylon yarn and some winder motors. We had to develop a table; the straight calculation didn't seem to work very well, perhaps because the yarn is compressible. The table had ten points for each yarn size, and the program interpolated in between these points. It took a while to develop the table empirically, but the system did work pretty well. BTW the program was written in ladder logic. Hope this helps a little.

Willy Smith
Numatics
Costa Rica

B

#### Bill Sturm

>I have a question. Does anyone have a generic winder control algorithm
>that commands the dc motor control to slow the speed of the shaft down
>as the diameter of the winder increases?

I wrote one once on a drive that did not support a winder block. I used a PID loop to regulate the speed of the line encoder by varying the speed of the winder. It is a different approach, but it actually worked very well.

The inputs to the PID block were the line encoder speed and the line speed setpoint. The PID output controlled the winder RPM. I think that I used integral only on the loop to avoid any sudden speed changes. The integral gradually adjusted the winder speed, to maintain a constant line speed.

I seem to remember using the integral reset feature when the line stops. The integral
caused the line speed to ramp up from a stop at the integral rate. More integral would produce a faster ramp.

Maybe this will help. Let me know if you need any more details.

Bill Sturm

K

#### Ken Brown

Marian,

What are you winding? Paper / web winding is considerably different than filament winding. Perhaps the easiest way to do this is to use an
ultrasonic sensor to determine real-time diameter. If the product you are winding is robust, run the drive in torque mode with an external velocity loop. Set the velocity target to a value 10 - 20 percent faster than that required and set the torque limit to that required to maintain the tension you desire at the measured radius from the sensor. Again, this has a lot to do with the media you are winding. You will have dramatically different algorithms for winding fiber optic cable at 3000 fpm Vs a steel coiler at 300 fpm. Also, I assume you are doing a center wind (turning the mandrel) vs a surface wind.

Provide a little more information on the application and I am sure you will get many helpful suggestions.

Ken Brown
Applied Motion Systems, Inc.
http://www.kinemation.com

J

#### Jerry Holzer

There are more ways to implement a winder/rewinder in ladders than there are
materials you can run on them. In general, the trick is figgering out the roll diameter; it's all pretty much straightforward from there. Feel free to msg. me privately if you want my \$.02 on PLC-based winding controls.

Jerry Holzer
Curt G. Joa, Inc.

*****DISCLAIMER
The opinions expressed in the above message are not necessarily consistent with those of my employer***

J

#### jmGiraud

Marian,
I have seen that same problem before in the
forum. Recently, I came across the mathematical
solution of it. Sorry I didn't note being sure
I could retrieve easily. Where now???.
From one of the answers, if you have tabulation
pertaining to your project, then I might help.
I suspect the approximation hides something
not easy in it. I'm not bad approximating
functions. <[email protected]>

D

#### David Kane

I will be doing a Control Loop Seminar at the "Electrical Manufacturing & Coil Winding Expo 2000" on Monday October 20 at the Cincinnati Convention Center for AIME.

A friend of mine wrote a 40 page paper on winding that would be of benefit to you. It covers the basics of winding plus many issues that are typically overlooked. It is an excellent paper.

I would not release it without his permission to do so. Anyone interested? Email me. I can ask his permission to get you a copy.

David Kane [email protected]
Kane Engineering Group Inc
and
AIME Association of International Motion Engineers

R

#### Ralphsnyder, Grayg

A winder line that I worked on had 12 motors. One wound up the finished product at the end and the one at the start of the line unwound the feed
stock. Between each of the motors was a tension sensor which would adjust the speed of the upstream motor. Basically, we needed to keep below a certain tension to prevent damaging the product. The winder end motor is where we set the line speed.

Grayg Ralphsnyder

R

#### Richard Norris

Can I suggest you are trying to reinvent the wheel. A eurotherm 590 drive offers excellent support for winding applications. Winder blocks internal mean it is very simple to set up. I would warn of trying to control the speed through a plc as the pid loop time is too slow. One other point is that current limit to control winding does not offer as good a response as using the speed loop.

C

#### Charlie Griswold

As if you have not recieved too much "help" already.

Using a simple mechanical "dancer system" would allow your motor speed to be independent of spool diameter. Proper tension in the material could be
maintained as well. You would use two sensors: 1) start (or speed up) motor 2) stop (or slow down) motor. You sense when the slack in your material is running out (1), and when the slack in your material has reached the most allowable (2).

The system can be as simple as draping the material over a guide rod, hanging under its own weight. Two capacitance, or optic, sensors are used to sense material high/lows.

Slightly more complicated dancer systems use three, or more, idler wheels. One, or more, of the idlers is spring, or gravity loaded to keep tension on the material. Hall effect inductance sensors are then used to register the high/low position of the traveling idler.

D

#### Dave Kane

Oops, Typo. Session at winding expo is October 30
Web site for the expo is www.coilwinding.org

Dave Kane

K

#### Ken Brown

Richard,

When did torque loop (typically 300 Hz in a decent AC Drive - including Eurotherm) ever get less responsive than closing a velocity loop
(typically 10 - 20 Hz in a decent AC drive)?

Ken

R

#### Richard Norris

Although I understood the topic to be dc drives I still maintain that a speed loop gives better control. This can be seen when trying to splice a new roll on an unwind. The torque control is harder to set the speed on a range of roll sizes. The real achievements in winding comes from the tension applied by the dancer and the taper ( in a rewind). The drive in these cases only serves to keep the dancer central therefore maintaining the tension. Another point that can be overlooked is the layon roller (rewind) and its effects.

D

#### David Kane

Speed control will provide better control (directly into the drive)

Torque / current command directly into the drive is generally not the best way to control a winder. Consider the roll inertia. As the roll builds up, the inertia changes substantially.

In a velocity loop the accel/decel is under control. In a torque loop accel/decel is a function of roll diameter, due to inertia change.

If tension control is desired, control the tension is as a function of position, where position/time=velocity. This takes us right back to velocity control.

For this purpose some winders will either have a sensor for roll diameter or calculate diameter by other means.

Now if we are talking nip rolls, that is a different story .

Dave Kane
[email protected]

K

#### Ken Brown

Hi Richard,

My intention was not to criticize the application of a velocity loop for a rewinder as much as it was to clarify what you said vs what you meant.
As far as rewinding aluminum sheet (I believe that is how this thread got started) the best solution has a lot more to do with what type of
drive / controller technology the customer wants to use. If he is free to choose - which he is not, he has already purchased the equipment . .
. I love the scenario where the customer is sold "what he needs" by the vendor and then has to figure out how it really works . . .

I think in the situation where you have a very low inertia core that is built up as the wind proceeds to a high inertia roll . . . an
ultra-sonic radius transducer should set the drive speed (velocity loop) that is trimmed by an idler with force transducer just upstream of the
rewinder. Scale the setpoint for the tension loop based on the real-time radius to get the tapered wind tension to prevent the roll from telescoping. This could all be managed in the drive (we have done this exact application with a Control Techniques UniDrive without the use of an external controller). That said, if I had my druthers . . .I'd use a drive in torque mode and do the app with an inexpensive motion controller using the above approach with additional logic to
switch from torque loop to velocity loop on-the-fly for threading / finishing control enhancements.

As far as AC vs DC . . . probably a topic for another thread, but in my opinion, the only reason for ever using a DC drive nowadays is when you already have the DC motor and don't have the \$ to replace it all at once. AC Vector performance is equal to or better than DC for all
practical purposes.

Best Regards,

Ken Brown
Applied Motion Systems, Inc.
http://www.kinemation.com

D

#### David Kane

I'd like to get more info on the control you mention. Is there a web site or URL?

Thanks
Dave

B

#### Bill Sturm

I will second the versatility of the Eurotherm drives, especially the 590 DC.

The 620 and 605 AC Vectors are powerful also, but they I do not think that they have winder blocks.

They program in a function block fashion, with each block laid out like an integrated circuit. You connect the pins from one block to another. It is like making custom analog drive boards in software.

Their URL is http://www.eurotherm.co.uk/drives/

Bill Sturm

A

#### Al

David
If possible could you please send me a copy of this paper on "winders" you mentioned in your reply to Marian Stolarski's original question.
Regards
Al
[email protected]

M

#### M Spears - Rockwell Automation Drive Eng

This is all well and good if the drive is able to have the same velocity loop bandwidth regardless of inertia. Which is not possible with most drives. Velocity loop gains are set for a given bandwidth at a given inertia. So, unless you have compensation for this, your velocity loop response will be very snappy at small diameter and very sluggish at large diameter. Tension is usually what you are controlling, so, since tension times width is force in lbs, and force times radius of the roll in feet gives torque in lbs-ft; torque is what you want to directly control in most instances where non-zero tension is desired. You will of course, need inertia compensation and a diameter calculator, but most drive vendors with winder experience know how to do all of this very well.

Speed control will provide better control (directly into the drive)

Torque / current command directly into the drive is generally not the best way to control a winder. Consider the roll inertia. As the roll builds up, the inertia changes substantially.

In a velocity loop the accel/decel is under control. In a torque loop accel/decel is a function of roll diameter, due to inertia change.

If tension control is desired, control the tension is as a function of position, where position/time=velocity. This takes us right back to velocity control.

For this purpose some winders will either have a sensor for roll diameter or calculate diameter by other means.

Now if we are talking nip rolls, that is a different story .

Dave Kane
[email protected]