Flying knife with AB servo's

Gents,

I've to program a new machine which has a flying knife, but I've no idea how start this.

We have an extruder which makes a plastic strip.
This strip must be cut in lenghts between 9 and 18 inch.
The system will get a Kinetix 6500 servo system with 3 axis.
Axis-1 is a servo motor which pulls the plastic strip trough a water bad, and feeds it to the cutter station.
Axis-2 is a linear thruster which has a cutter-head, and has to follow the strip on line-speed, cut the strip, and move back for cutting the next strip.
Axis-3 is a servo motor on the cutter-head to activate the knife.

The encoder of Axis-1 is the reference for the strip lenght.
So Axis-2 has to follow this reference signal, to be on the same speed, but also on the right position to activate Axis-3 and cut the strip.

My problem is, how should I program this?
It's not difficult to start Axis-2 at the lenght signal, and also the speed is not a problem. But Axis 2 has an accelaration time, and during this time the strip moves forward. So Axis-2 should start earlier, or should move quicker to find to cutting-point. and follow the line-speed till the cut is done..
For you're information: the line speed is about 45 mtr. per minute.

The PLC will be a L27-ERM, and communicate over EthernetIP with the Kinetix 6500. Programming is in ladder.

Any idea is welkom, thanks in advance!

Jack

Take a look here:

http://www.ab.com/linked/drives/applicationdetective/cgi-bin/ad.cgi?pg=flying_cutoff&sn=1

I believe you may find something already made or close to it...

It is close to it, yes.
But I see no sollution how I should create the program to find my cut-point on the fly, and this system is for a cycle time of several seconds. Our application should do 0,27 sec. cycle time...

Thanks anyway!

Hi

If you use camming and have Axis 1 as the master then axis 2 as the following axis you will always start at the same position. You can then adjust your cam engage point to get it right . By doing it this way the speed will not be issue as the cam will work the same all the time

Look on Ab for sample camming program

Donnchadh

I was thinking about the camming sollution, but I wondering if axis-1 runs for hours on the same speed, will that be seen as a cam.
Or can I run an endless amount of cams on axis-2 during 1 cam on axis-1?

For your information, axis-1 doesn't make an index for every strip, it runs on a nearly fixt speed with as reference the screw-speed of the extruder.

I found the sample program "FlyingShear" which seems to have usefull information...

I see that you found sample library.
I have designed three of those systems so feel free to ask.
Peter Natchway is the man with all the knowledge.

widelto, have you used the sample program?
Which drives / motor have you used?
Which CPU, communication and why these?
What was the cycle time, and how accurate are these systems?

Thanks

Nightline: No, but I extracted some ideas from that program. My system is a huge one, servo motor (carriage) is 60 HP and saw motor is 150 HP. But the idea is the same.

I use no sercos drive, my drive (carriage) is ABB ACS800, there is an ABB option called servo motion, I did not use it, but by reading the manual seems to be suitable.
I use a controllogix with a 1756-M02AE board.
From PLC to drive I use +/- 10 VDC.
Cut length is 6000 mm, normally is 5999-6002.
Line speed is up to 40 meter/min.
The system is used on a steel pipe mill.

I've selected single-turn / incremental encoders for all motors in ProposalWorks now. Is this the right choice, or is it better to go for absolute encoders?

Incremental encoder is ok, use a switch for home.
Use high count for the encoders 5000 ppr is ok. The higher the count the better the resolution.
Do not forget that encoders must use line driver technology.

In my case I used an encoder splitter, carriage motor encoder is connected to ABB drive and 1756-m02AE board at the same time.

If the line speed is being controlled by the controller then it is best to gear everything to a virtual master that is a perfect device. Gearing to the line encoder is not ideal because it is difficult to compute the derivatives for feed forwards. Also, I have had cases where when the knife cut the material it pulled the material because it was being pushed down by the knife. This causes the line encoder to react. In this case our controller was not controlling the line speed so we had to make a virtual master to filter out the encoder quantizing and reaction to the knife cut. Then everything worked well.

Very good information guys!
I'll check the encoder parameters, and will correct before order the equipment.
Virtual Master...that's a good point...I'll program that.