Centre winder drive

[joseph_e2]

I haven't done anything like a winder with carpet but have worked with some that coiled up metal. I would suspect that the torque while winding carpet would vary over a pretty wide range based on the diameter. Not just because of the diameter but because of the "stretchiness" (is that a word?) and compressibility of the carpet. My guess is that you'll have better results by measuring the tension directly, like with loadcells. I've seen that done very successfully with steel and wrought aluminum.
Just a thought...

 

[Puddle]

I would suspect that the torque while winding carpet would vary over a pretty wide range based on the diameter. Not just because of the diameter but because of the "stretchiness" (is that a word?) and compressibility of the carpet. My guess is that you'll have better results by measuring the tension directly, like with loadcells. I've seen that done very successfully with steel and wrought aluminum.
Just a thought...

Sounds like I'm on the right path then.
The carpet does stretch and this in itself is an issue, it can cause creases, can result in a narrow product or just produce general inconsistencies. All of this results in product being lost.

I'm going to mount the load cells as read only before I commission the drive and spend a few days logging the tension so I can get a baseline of how it's currently running and no doubt discover it's double what the material requires.

 

[JRW]

Let us know how it goes.

Thats a lot of gear ratio for decent torque control by the way

 

[Puddle]

Update:

Found a window in production to replace and commission the drive. After some playing around to get the speed consistent, I found that changing the torque limits on p1522 and p1523 didn't work how I expected it to, so I'll have to spend some time with a spare drive on a test bench working that one out. Changed it to purely speed control and the rolls are much better than before, although still getting far too much variance for my liking (+/- 5kg) so there's still some fine tuning to do.
I did consider changing the gearbox, but with the roll diameter I found that an empty roll needed around 65Hz to match speed, but by the time it's reached 700m of material on the roll it's more like 5Hz.

 

[kekrahulik]

Are you measuring or calculating diameter?
Using loadcell feedback?

 

[Puddle]

Measuring through load cells. That was all that was really needed, but the original design didn't have enough analog inputs to take line speed, diameter and tension into account so they did what they could. In theory I could have used the same drive, but thought I'd need to send more data over profinet than I did in the end.

 

[kekrahulik]

Not sure what you are doing, but for speed control winders I calculate the diameter and use that to calculate the speed command. The loadcell feedback is used to trim that speed. So usually the loadcells can only affect +/-5 -10% of the total speed.

 

[Puddle]

Not sure what you are doing, but for speed control winders I calculate the diameter and use that to calculate the speed command. The loadcell feedback is used to trim that speed. So usually the loadcells can only affect +/-5 -10% of the total speed.

That’s pretty much exactly what I’m doing. Calculating circumference of the roll using a distance laser, referencing that to line speed and then trimming it with the load cells run through a PID.
I have it biased more towards the load cells than you’re recommending and thinking about it now, the line speed calculation was pretty good so a smaller ratio should work well. Luckily I made the ratio a configurable value so I think I’ll try dropping it in the morning and see if it helps.

 

[JRW]

and to add...
if the diameter calculator (your laser sensor scaled) is working properly you should be able to put in a 1/2 or 3/4 full roll in and hand tach the roll. The surface speed should be running the line speed (pid load cell trim turned off)

 

[kekrahulik]

I only use a diameter sensor for my initial roll diameter. After that, you can dynamically calculate the diameter if you know the web speed (speed from a pull roll drive) and the winder RPM (from the winder drive). The accuracy of the that calculation is significantly better than you'll get with a sensor.