Unrolling without diameter measurement

I have an application where I do unrolling of a bobbin.
It is a central unwinder (axis of the bobbin is driven).
We have no diameter measurement.
The starting diameter of the bobbin can be very small to very big.
There is a balancer after the unwinder.
We setup a linespeed that the unwinder has to follow.
When starting, there is a rather long acceleration time, allowing all getting sychronised well (balancers must stay approximately in the middle).
For now the unwinder gets a speed setpoint proportional to the line speed.
There is also an addition of a PI controller concerning the balancer of the unwinder. The integral factor shall make up for the variation in diameter.
During acceleration a rather big integral factor is used, when at speed the integral factor is reduced.
But we keep having troubles unwinding the large bobins.
Now I was thinking of when accelerating from start, first go to low speed, then waiting until the balancer is stable, and then calculate myself the diameter of the roll based on the actual unwinderspeed and the actual linespeed.
Is this also a procedure that some of you would follow if no actual diameter measurement can be placed?

Can you give us a diagram of your control theory?

Mount a sensor that can detect each rotation of the bobbin. Use that signal to measure the time for one bobbin rotation. That time is how long it takes for an amount of thread equal to the circumference of the bobbin to pass. Multiply that time by the line speed and divide the result by PI to get the bobbin diameter.

Steve
I THINK his problem is that he starts with different size rolls to unwind

I guess a “balancer” is a dancer?

Yes, the machine does not know if large or small diameter bobbin is mount at start.
A balances is a dancer.

Wheres your control diagram?
How do you calculate diameter now? You need one
The dancer should just trim your speed... roughly +-10%

also, is this web or filament?

You could also attach a potentiometer to the dancer arm and use the signal from it to adjust the speed command to the unwind drive.

Mine is set up like Steve Bailey's post only I use an analog prox to determine dancer position since it pulls an arm and does not have that much travel. This is a tear tape unwinder and the spool is more of the "bobbin" type. On the big paper roll splicers we have the Potentiometers determining Dancer position.

Before you start the machine, rotate the unwind until the dancer has moved a defined amount allowing you to perform a reasonable estimate of unwind diameter based on a lookup table or calculation.

It concerns steel wire bobins.
I don't have the exact program now with me, but the speed of the unwinder motor is the summation of (linespeed x constant factor) + (PI controller on difference of requested dancer position : 50% of maximum - actual dancer position).
The integral component should make up for the error that is made because we don't know the actual diameter and can't calculate the actual request speed.

aand74 said:

It concerns steel wire bobins.
I don't have the exact program now with me, but the speed of the unwinder motor is the summation of (linespeed x constant factor) + (PI controller on difference of requested dancer position : 50% of maximum - actual dancer position).
The integral component should make up for the error that is made because we don't know the actual diameter and can't calculate the actual request speed.

Click to expand...

Another way to do it is use the us the dancer position and scale the PID output to something like 0.7 to 1.3 or 0.5 to 1.5. You can then multiple your base speed by that value. Your equation would be LINESPEED x CONSTANT x PID_SCALE_FACTOR. This makes your trim proportional at all speed ranges.


Another thing you may want to look at is to make sure your dancer has enough accumulation to give your PID loop time to react.

Measure actual speed

You need to measure the Actual speed. You measure wire speed from bobbin with an encoder or analog tacho on a support wheel/pulley before the dancer. Substract actual speed from set speed and add the dancer zero deviation. Let the speed difference count for 60..80%, and the dancer position 20..40%. Add this total correction value to line speed, and use set value for bobbin motor drive. With this setup you don't need PID.

Winder speed is linespeed *(Dsync/D) where Winder speed and linepeed are per unit i.e. 0-100%. Dsync is synchronous diameter = the smallest diameter at which 100% linespeed can be achieved i.e. the diameter at which linespeed and winder speed are 100% D is diameter in the same units as Dsync.

Diameter can be calculated as linespeed/winder speed *Dsync. This needs to be filtered and does not give a good result at low speeds.

L D[AR2,P#0.0] gave a good suggestion of how to estimate the diameter to pre-load your diameter filter prior to starting.

Once you have a diameter calculator in your system, the integral value that your PID loop settles at will give you a good indication of the accuracy of your diameter calculation.

Nick

start up the machine, as long as the dancer is low (enough material speed up the machine) if dancer is lowlow reduce unwinding accelaration, otherwise speed it up. it means the unwinder is always between low and lowlow. if dancer is high speed up unwinder, if highhigh reduce machinespeed as winder cant keep up with machine.