Paper slitter/rewinder Decel to predetermined stop footage

I have an application where I want to decel the slitter/rewinder at a calculated preset and stop at a predetermined footage setpoint. The machine is controlled by a ControlLogix processor. I have tried
.5*(V/decel_rate) but need to incorporate jerk_rate also.
I end up way short or overshoot. How about some good advice here.
Decel and jerk are constant.

what is your hardware?

analog output to VSD drive?
Or network speed reference?


the formula that you are trying to use is v^2 = u^2 + 2*Accel*distance

U = current velocity
v = final velocity = 0
Accel is -ve as you are decelerating and it is fixed
distance is distance remaining to stop position

so your velocity reference is u = (2*accel * distance)^0.5

notice no jerk in the equation

This is a servo job with a pulse encoder on the winder core shaft. You can't get there with a VFD especially if the unroll length is to be adjustable.

DickDV said:

This is a servo job with a pulse encoder on the winder core shaft. You can't get there with a VFD especially if the unroll length is to be adjustable.

Click to expand...

servo needed? - it all depends on the accuracy and accel rates required
- 0.01 mm in 1 m and it is important to start stop quickly yes
+/-500 mm in 10,000 m and each roll takes 15 minutes to create - no

Pulse encoder on the back of a any rotating axis (driven by motor or servo) is still just a counter

The servo allows you to have high accel and decel rates with accurate stopping per Rev
Eg rockwell M02AE can control a VSD motor with an encoder on the back of it and get better than 1 % positioning accuracy per rev
servo I expect 0.01% accuracy

this is all before any gearing which which can improve the accuracy

D = (V² / (2*a)) + ((V * a)/(2 * J))

where:
D = Stop distance in units
V = run speed in units/sec
a = acceleration in units/sec²
J = jerk in units/sec³

There are a couple of important caveats to this. If your speed is low enough that you never reach constant acceleration this equation won't work. If you are actively accelerating you can't count on the distance value from this equation. But if you are running a constant speed and the device implementing the ramping actually ramps according to the configuration values this will be pretty accurate. Also, your scan time and device response time needs to be fast enough that the drive begins to decel very shortly after you tell it to. In the cases I have used this in the past I can stop within a foot from 2000 FPM. If that is accurate enough for you this should work fine.

Keith

MichaelG, I think you've just proved my point. You've put a positioning controller ahead of a VFD and used the VFD as an amplifier. That's a servo system in my view except the dynamic performance is quite low.

Whenever your are counting pulses to get position or length, that's a servo system in my view.

There are VFD's out there with positioning front ends. ABB used to make a model ACP drive available only in Europe that was built that way. To me, that's a servo too.

I understood you to want to get length without counting pulses somewhere. I don't think that will work unless the accuracy is way low.

All
I am using a ControlLogix processor and talking to 3 Siemens
D.C. drives over Profibus. The unwind is a variable brake under dancer control
Thanks for all the feedback thus far. I will apply
these formulas on Monday and see what happens.

Originally posted by DickDV:

That's a servo system in my view except the dynamic performance is quite low.

Click to expand...

So I'm not the only one, eh? 20 years ago an automation market segment was allowed to commandeer a term that is not exclusive to it nor specifically indicative to what it means today. Now we all get to live with it. Some day, just for funsies, everyone should look up the definition of servosystem. It would be enlightening.

We've both done enough of these systems. I think we are pretty safely playing the percentages to assume that stopping within a couple feet is probably OK for what he is doing. Additionally I would expect he has a web measurement encoder somewhere or is using pull roll or winding drum position to determine lineal run (as opposed to summing speed). Given his initial equation he would have had to be pretty wildly inconsistent.

Keith

DickDV said:

MichaelG, I think you've just proved my point. You've put a positioning controller ahead of a VFD and used the VFD as an amplifier. That's a servo system in my view except the dynamic performance is quite low.

Whenever your are counting pulses to get position or length, that's a servo system in my view.

There are VFD's out there with positioning front ends. ABB used to make a model ACP drive available only in Europe that was built that way. To me, that's a servo too.

I understood you to want to get length without counting pulses somewhere. I don't think that will work unless the accuracy is way low.

Click to expand...

+ 1

lol - it is all in the definitions

my previous experience difficult to control the length using this method,

result will be overshot or short due load/ramp down.

although using clutch brake,

other method using 2 step speed high and low if using VFD/DC Drive.

to slow or high paper will break.

drive synchronize also important.