Wind / Unwind Application, Torque vs. Speed Control

I'm doing my first wind/unwind project, and I need some input from someone with more experience. ...I'm getting my butt kicked trying to control it.

It's a simple wind/unwind laminating machine. One winder, two unwinders. The drives have firmware for the application. I thought that would be the safe bet, since I was updating the machine with the the same brand drives as it had, with the most recent model and the latest application firmware.

It's the unwind drives that are giving me problems. The old drives had a control loop based on torque, so that is what I expected in the new model. There is a load cell on each unwind roll web. You enter the setpoint for the web tension, and the feedback comes from the load cell. The drive also keeps track of the roll dimension to use in calculations. After setting up the application with the manual (and sometimes wanting more detail about the parameters) I was unable to get the control loop to control the negative torque at all to match the tension setpoint. After several frustrating days I got some help from the local distributor and manufacturer. In turn, they got the applications tech desk involved. I went through everything with them and shared my parameter files. They couldn't get it to work, either. The application firmware was just not controlling the torque at all. The application guys admitted they didn't have experience with that particular firmware and said they needed to consult with specialists later in the day/evening.

After a conference call with a wind/unwind application specialist they got back in touch with me and said we needed to change the control mode to "speed", not torque. I knew that was an option, but it didn't seed as solid as torque. We made the parameter changes and the control loop started to work. Then, it was obvious that I needed to focus on the PID, since the tension control trend was ugly.

After a couple of frustrating days working on the gain and integral, I started to see a pattern. When the rolls were full, I had a decent amount of control... not perfect, but promising. But once the rolls got small, and the RPMs high, (line speed is kept constant), the control fell completely apart. Someone mentioned to me that torque was the better control method for wind/unwind, and it got me wondering if I'm headed toward a dead-end trying to get this machine working.

Tomorrow, I'll head back to try again, and I'm thinking about changing my strategy from proving it can be done, to proving it *can't* be done. I just need some thoughts from someone with more experience.

I always like throwing a picture in, so here's a visual for you. You can see the two loops, paper and poly. The tension feedback for each roll is there, along with the setpoint. You can also see the ramp and the speed of the winder in red. Can't remember what the Magenta line was, though. You can see how the control fell apart towards the end.

Any thoughts are appreciated.

Brian
Need more detail. So is there a nip drive (rolls) between the unwind and wind drives?
And what drives are you using?

Also gain may need to be based on diameter of the roll

Picture of the web path too

ABB, ACS880s, N5000 firmware.

I agree there must be a set of nip rolls to set and control the line speed
these would be the master speed reference
All 3 rolls Rewind and unwind must have load cells or dancers to measure web tension
the rewinder should be setup as torque control to control the tension on the line
the 2 unwinder should be setup as speed control with the load cell feedback to control the unwind tension of it's half of the web.
Remember all 3 winders will have a varying speed as the roll diameter changes
as the roll diameter get's smaller the speed must go up to maintain line speed and it follows that the rewind speed as the diameter get larger the speed must go up along with the required torque to maintain the web tension.
I don't see how you can do what you need without load cells or dancer feedback for each roll.
Let us know how you make out

Line speed reference is given to all three drives over the field bus from the plc. All three motors have encoders on them. The winder has always done a good job of maintaining line speed by using the prox sensor for the roll diameter. This machine was working just fine for over a decade, using torque control.

Yes, there is a relationship between line speed, roll diameter, and RPMs. If you have any 2 variables, you can calculate the Third, which is what the drive does.

Look at the picture, I have pointed out where the load cells are. You just can't see them because they're on the other side of the plate. The load cells are giving the feedback values on that first graph

One small detail is changed from the picture. The winder now is using Dynamic torque control, not scalar mode.

I am going to try and give some help if I can
As I think was said earlier we don’t have all the information there are still a lot of unknowns
In your setup the winder has no tension feedback they are using just the prox sensor to measure the diameter. I really think that the winder prox is really an ultrasonic or laser distance measurement device of some sort. Measuring the diameter of the roll. The diameter of the roll will give you speed requires for the line speed. I prefer to use a speed sensor / encoder to actually measure the speed much simpler to deal with. I have seen systems where they use a torque command to control the output speed Old servo systems were torque control they limit the speed by limiting the torque. On some CNC systems the controller provides only a torque command to the servo motor control The speed is controlled by limiting the torque command to the motor control. So yes they could be controlling the winder speed with a torque control. In my opinion the best way is to use speed control but it gets the job done.
If the winder is running with speed control, you increase torque command to increase the speed. As the roll weight increases the torque required to maintain speed will increase. There is a formula to calculate the speed ratio empty roll to full roll + the additional torque requires to handle the increasing weight of the roll over completing the process with all the additional calculations necessary to do this.

That would mean the unwinder’s must be running in torque mode. For them to do that they would be operating in regeneration mode zero or low speed command. The torque controlled from the feedback from the load cells to maintain the preset tension. The unwinder speed could be calculated from the ration of full roll to empty roll from the diameter but diameter is not really what you care about.

Are you sure that you have VFD – AC motor or DC drive with DC motor or even a drive of some kind on the unwinder.
I have seen many unwinders where they just use just pneumatic or magnetic brakes to control tension. They were more common in older systems. Corrugators, Paper converters, fabric machines.
I hope that helps

Brian

Heres what I would do...

Control the winder in speed control (no torque limit etc). This is the line speed master so the material speed be constant from core to max roll diameter
(measure and test a few rolls with a hand tach)
Material wouldnt even need to be threaded up
Then the winder should be done

Now to the unwinders..

I would limit the torque to both of these motors ..just enough to trim with the load cell and create some back tension (just enough to prevent wrinkles etc)
(can the ABB drives be speed controlled with torque limits?)
If you do this, then the pid doesn't have to be tuned perfectly at this point

See if the system runs at a lower speed through a roll build (min to max rolls)

This machine can be done
But winders are tricky

I will play around with that today. I am considering the winder to be done. It's all about the unwinders now. Thanks. There's a good chance I'll post some more as the day progresses. I have to find the answer.

AutomationTechBrian said:

I'm thinking about changing my strategy from proving it can be done, to proving it *can't* be done.

Click to expand...

GaryS said:

I have seen many unwinders where they just use just pneumatic or magnetic brakes to control tension. They were more common in older systems. Corrugators, Paper converters, fabric machines.
I hope that helps

Click to expand...

If you do scrap the way you are doing it, I would use the brake

I have worked on winders for about 20 years and have seen many styles also built a couple, the break on the unwind side and drive on the wind with a dancer roller is the best way by far.

You can get 4-20 brake caliper and control the brake with the dancer potentiometer, simple and works great.

There are four machines in a row, this is the only one with vfds. There is no way I could change the machine that drastically. I have often wondered if I could use the Allen Bradley compactlogix 5380 to control the PID. I keep thinking about the firmwares inability to use torque for the loop. That seemed to be the basis of the old system.

I would try and contact the vendor again and see if you can get a hold of another "application specialist" and I agree with you, speed is not the correct setting unless you had some tach's on the shafts but with load cells you would need torque.

geniusintraining said:

I would try and contact the vendor again and see if you can get a hold of another "application specialist" and I agree with you, speed is not the correct setting unless you had some tach's on the shafts but with load cells you would need torque.

Click to expand...

Email has been sent. We'll see what happens.

I may end up doing another "almost real-time" event with this later on today. I have the ability, and the last time I tried it it ended up being quite the post (all documented for future reference). I'm getting to the desperate point and my time is running out. .....Time to think out of the box.