Cut to Length application using SLC-5/04

Hi...
I have an cut to length application using vector motor and powerflex700VC. Line speed is 40 MPM, cut length is 2000 mm +/- 2 mm and type of knife is rotary knife. Knife circum is 1350 mm. Knife home position is 0 deg and cutting position is 270 deg. How to syncronize the knife speed to line speed? I set the knife speed to line speed and try to cut the material but I got length variation (>30mm).
Any idea or suggestion?
Thanks.
Osias

Are you using a flying shear to profile the knife speed to get the cut length at 2000mm ? How many degrees before and after the cut are you running the knife at web speed ?

I did one of these with a Siemens Masterdrive VC and a T400 with SPS450 software-

Without a motion controller of some kind- you will never get it "just right"

Does the SLC500 have an option board for motion control?

Osias,

With a 0.1% cutting length tolerance, synchronization is very important indeed (hence your question). In order for us to better assist you, I think more information is required.

40MPM means 40 meters per minute, right? If so, that equates to 20 sheets per minute, each one taking 3 seconds to feed at a rate of 667mm/sec. You mentioned that knife home is 0 deg and knife cut is 270 deg. That indicates to me that the knife is not continuously rotating and instead:

- Starts moving from 0
- Accelerates to line speed
- Makes the cut at 270 deg without affecting the material flow
- Stops and waits for the next cycle


If this is true, what triggers the knife?

You also mentioned separate speed controls for the knife and line, so I assume the knife has some sort of dedicated motor and is not directly connected to the machine drive train.

With no other information provided, my understanding thus far is that you need to synchronize two otherwise asynchronous functions (rotary knife and material flow), right?

Please provide more information about your process, such as how the the material fed, how much slippage, method of measuring speed (actual material speed or only the line speed, or not at all), type of knife motor, etc. This will better enable us to help.

You will need some sort of analog speed feedback to the knife, otherwise, you will not get it right, just like JRW indicated.

CeCo3

JRW said:

Does the SLC500 have an option board for motion control?

Click to expand...

There are at least two or three motion controllers that do rotary knives easily connect to a SLC500 over Ethernet. In some cases the motion controller is much more power than the SLC and yet looks like a SLC on RSLinx. The trick is to be sure the cutting action doesn't disturb the master encoder. As all have pointed out there is not enough info to provide a good answer other than symbollic formulas. Even if Oasis has all the math skills to figure out the formulas, I don't see how he could do what a motion controller does in a PLC. In any case, why re-invent the wheel?

If this is true, what triggers the knife?

Click to expand...

The master get so a certain position and triggers the knife but like I said I prefer to move the knife as smoothly and continously as possible so it is more a function of the knife position, velocity and acceleration as a function of the master or line position, velocity and acceleration. Does the chain rule ring a bell?

It may be possible to synchronize the drive to the line by using encoder pulses from the line as a drive input (no motion controller). Your drive will have to be able to tightly control it's motor too.

As the others have said, you will be much better off reading the line speed encoder with a motion controller though. The SLC 5/04 will not be able to control the drive speed directly though. It can interface to the drive for diagnostics, and to a motion controller, but will never be quick enough to meet your specs if it directly controls the drive speed and run logic.

Either way, you'll need good resolution position data from the line you're following in order to set up a flying shear application to those specs.

I tried to do the same using a 1746-HSCE2, installed an encoder on the drive that moves the sheets and used a fast output from the HSCE2 to trigger the knife but it never worked.

When the system was running at low speeds it ran fine but when speed was increased error increased too resulting in a problem, my solution was to use a control logix and a servo control board.

Thanks for reply.

I use SEW duty inverter motor with 2000 PPR encoder that connected to Powerflex700VC (vector control powerflex). I use two 2000 PPR encoder, one use to measure web length, another one use to indicate knife position. Web Conveyor as a master speed with 40 MPM (metres per minute) will feed the knife. Line speed feedback and knife speed feedback are connect to 1746-NI8 (0-10VDC), Knife speed ref from slc analog output. Both encoder will connect to 1746-HSCE2 encoder module.

With 1350 knife circum and the knife distance from 0 deg to 270 deg is (270*1350/360) = 1012.5 mm. Cut length is 2000 mm then knife will start to accelerate if length reached (2000mm-1012.5mm) = 987.5 mm. Knife speed is equal to line speed. Knife will start to cut the material and back to home position. We use dynamic braking resistor to make sure knife stop at its home position, knife always stop at 0 deg.

Web length will calculate from encoder pulse using formula:

Web Length = Encoder_Pulse * Measuring_Wheel_Diameter * 3.14159 / Enc_PPR.

I hope this information will help.

Thanks for idea, suggestion…

Regards,

Osias

You still haven't provided us with the angle at which the knife makes contact and leaves the material being cut. The tangential velocity is omega * radius but the component moving in the horizontal direction is cos(angle from bottom). This trig function must be taken into account the whole time the knife is in contact with the material if you want to cut the material without disturbing it.

What is the time constant of the analog output card you are using? I am almost certain it is too slow for motion control.

Thanks for reply.

The angle that knife make a contact to material is 10 deg (270 to 280 deg).
regards,
Osias

One of our customers just completed a similar application using a rotary shear for cutting sheetmetal. They used our RMC70 motion controller along with a PLC. I don't remember which PLC, but the RMC70 easily communicates with many PLCs, including the SLC via Ethernet or serial RS-232.
Delta has example projects for the RMC70 for doing this type of application. The example takes care of cutting at the exact position while matching the horizontal speed of the knife to the the speed of the material from when the knife enters to when it exits. The cut length can be changed on the fly. Line speed changes are no problem. Contact technical support at Delta Computer Systems, Inc. for more info (www.deltamotion.com).
If your system mechanics are well-designed, meeting the +/-2mm requirement would be no problem.

-Jacob Paso

Osias-

Take the previous posts to heart. I've done a couple of rotary knife application and you aren't going to be able to do this directly out of a SLC. You would be hard pressed to do this in a ControlLogix without a motion card. Neither the processor nor the I/O are up to the task as far as speed is concerned, even considering the relatively low speed and open length tolerance. Also, I can't think of a motion module for the SLC that will allow you to do the gearing and profiling you will want to do. That doesn't mean one doesn't exist. I just don't know of one.

You must have a pretty inefficient geartrain to be able to use such an asymetric velocity profile. Accel from 0 - 270 degrees, constant speed from 270 - 280 degrees, then decel from 280 degrees to 360 (zero). The knives I've worked on have varied only about 5% in the accel and decel torque requirements at max speed. You must be getting a big decel help from the geartrain.

There is no reason to come to a stop complete stop in given the 2000mm case. You can simply decelerate for a while and accelerate back to line speed. The trick is calculating what speed to decel to based on line speed and sheet length.

Part of the problem is you don't appear to be generating a position profile for the cut you want to perform. You are picking discrete points to trigger various motions with no way to adjust if the trigger point is off a little. Granted, it is just math. But, unless you will be doing a whole mess of these knives it doesn't make much financial sense to do this from scratch in a SLC, assuming you can even get the SLC to perform well enough to do this. That is a huge assumption.

I recommend that you talk to someone like Delta Computer Systems about a motion controller for this application. If you subtract off the costs of the PLC I/O you need and the development time you will put in to do this I'm betting the cost of a motion controller will look pretty attractive.
Keith

I am 4758.45 miles from home according to Google Earth.

I was hoping that our tech support would baffle all with magical graphs and mathemagical formulas but they are busy writing manuals. Hopefully I will find time this weekend to detail all to death with formluas and graphs just like we did for Widelto's thread about the flying shear.

Every once in a while I like to go into the field to keep in touch with reality. Ireland is a good place to do so for many reasons. When I have time I will show how real rotary knife or clippers work, This week I have been working long hours and you all know how that goes. Hopefully I will get a chance some time this week end to show how a rotary clipper should work in detail. I think I am working weekends too.

More info.

at 40 m/s I calculate the line speed is 6.666 mm/10 milliseconds which is a typical plc scan time. The PLC scan time is not fast enough let alone being deterministic enough for high performance motion control.

The formula for the slave ( knife ) as a function of the master ( line encoder ) is:

slave position = Radius * atan(master position/height)

I like to use 0 as the cut position or when the knife is straight down. This makes calculations easier.

The Radius is the 1350mm/(2*PI). 214.85917mm.

The Height is the distance between the material being cut and the axis of rotation of the knife. When leaving the cut the equation above applies. When entering the cut the

formula is

slave position = Radius*( 2*PI + atan((master position-CutLength)/height))

At the point of entry and exit one must also calculate the gear ratio and gear ratio rate as a function of the master position. This is done by taking the first and second derivative of the equations above. This is only the first step. One must also apply the chain rule ( calculus ) to take into account the master velocity and acceleration. Then one can calculate the slave velocity and acceleration.

The motion across the top is easy because you have the position, gear ratio and gear ratio rate at the entry point and the position of the master at the point. We have a command that will generate the smoothest motion profile to a point where you just specify the destination, position, gear ratio, and gear ratio rate at the destination and the corresponding position of the master. This means the knife usualy will not stop at the top. The command tries to keep from changing the velocity or acceleration any more than it has to. The slave ( knife ) velocity and acceleration must be calculated to compute the feed forward terms. A PID by itself can never follow a motion profile where the velocities and accelerations are changing rapidly and often. Feed forwards are required in this application.

Note that the line speed is 40mpm.