Unwind Control

[fooltech 0]

I have to unwind material from a drum at a rate variable from 5 to 20metres per minute I have as 0-10v input and can sense the diameter of the roll (0-10v) I have an inverter which can control the rotational speed of the roll but I am looking for ideas for controlling input to inverter because obviously it will be a non linear reponse ie higher rate of change of speed as roll gets smaller Will be using SLC5/03 Any suggestions would be greatfully received.

[mjrx 1]

Even though it's non-linear, the shape of the curve should be constant, so you could build a lookup table and index into the table as the roll gets smaller.

[Bill Linne 0]

Because the curve is linear, you should be able to mathematically calculate the output as a function of the input sensing roll diameter. A bit of experimentation would give you the output required to get certain feed rates at various diameters, then calc the curve. A lookup table would also work. Bill

[Gerry 13]

I don't understand how you can describe a curve as "linear". For a centre-driven unwind, RPM vs diameter is non-linear for a given web speed. Sorry I can't recite the formula, but it involves square-roots and the thickness (caliper) of the web. The 5/03 may be capable of preparing a look-up table to approximate the curve, but I would expect the results to give significant variation in web tension. I would recommend using a dancer roll to trim the unwind speed and maintain tension.

[Spedley 0]

I don't see why you thinks it's non-linear? Surely the unwind rate is the circumference of the drum x the rotation speed and the the circumference is the diameter x PI. I assume the value from the sensor is linear giving a distance from the sensor to the drum. All that is required is to find the correct speeds with a full and an empty drum and then put those values in a SCP instruction (scale with parameters). P.S. controlling input to inverter because obviously it will be a non linear reponse ie higher rate of change of speed as roll gets smaller Does that mean you control acceleration rather than speed? Edited 23 Jul 2005 by Spedley

[fooltech 0]

Hi The output required to the inverter to control speed is non linear I have been playing with a fixed line speed for the moment trying to establish some parameters and will get back when I have sorted it out. Dancing roller is not possible because of space constraints and the nature of the product. Have never used look up tables before has anybody any idea where I could find a sample. Thanks

[mlarkinp 0]

I don't know what your budget is, but you can get inverters with onboard software for wind/unwind applications. You could possible run the sensor directly to the drive. Given some basic info, the drive could run the roll on it's own.

[jimtech67 2]

Are you following line speed from another machine or device? Is the unwind feeding another process or machine? Do you need to keep a tension set point on the web? What brand drive are you using? The following will assume you are feeding the unwind into another process 1: Speed match the unwind (with the equipment being fed or the driven master roll) at core with max speed. Output of analog signal 10volts = max speed. 2: Set the diameter sensor like this: 0.5 VDC for core 10 VDC for max dia. Using the SCP Scale with Parameters instruction to find an actual diameter from the sensor feedback. (for example lets say 3” core to 32”) 3: analog output for rewind speed = (speed set point ) * (core/ actual dia). This will give a speed match throughout the diameter range BUT you will not be able to keep tension if you are feeding another piece of equipment or another driven roll for a million reasons. You will need to close the loop with some sort of tension feedback device. This can be done in the PLC or in the drive if it is capable. If you go the open loop route you will need to under speed the unwind and control tension by drive torque or a mechanical clutch. I hope I did not make too many assumptions about what you are doing and this info is helpful. Post ant other questions.

[fooltech 0]

Hi I am using a Telemecanique ATV 31 inverter newer ATV 71 would probably do the job but have got to use what I've got for the moment. Decided to split curve into 4 linear sections on the basis that even a circle is made up of straight lines. Using the SPC instruction and limiting the line speed I am having reasonable success just trying find the roll over points from 1 line to the next at the moment. When I think I understand what is going on with one line speed range I will look at the others, biggest problem is that web has to be more or less tensionless.

[jimtech67 2]

Do you have any type of tension feedback into the PLC or drive ?? If you do not it will difficult to keep the proper speed. “Using the SPC instruction and limiting the line speed I am having reasonable success just trying find the roll over points from 1 line to the next at the moment” I do not understand what you mean by this. What type of device are you using to sense diameter ultra sonic, layon roll with a pot??? I am not familiar with Telemecanique ATV 31 inverters but call the manufacturer they might have a macro or boards that will do closed loop speed unwinding. I if you are going to replace the drive I suggest SSD (Eurotherm) 690+ it comes with all Winding and unwinding macros to do what you need. Jim Edited 2 Aug 2005 by jimtech67

[panic mode 246]

visit http://curveexpert.webhop.net/ and download Curve Expert. with this little thing everyone can find decent approximation for any curve. all you need to do is enter few points and then implement the formula in your PLC (maybe by using CPT instruction). I agree with Spedley though. why do you think relationship is nonlinear? circumference of a circle (a good aproximation for this application) is l=2*pi*r=6.28*r this is linear function. speed of the unwound material is v=l/t this is still linear function. is there some property of the material that we don't know?

[fooltech 0]

Thanks for everyone's help have just about succeeded using the following (Line Speed / Radius of Roll) X (Freqmax/(2 x pi x Nmax)) = Required Frequency Freqmax = Maximum Frequency of Inverter (Hz) Nmax = Measured Output of Gearbox at Max Frequency (rpm) Line Speed = mm/min Radius of Roll =mm Scaled the two inputs to engineering units and worked from there, slight difficulty in that for as yet unexplained reasons at certain speeds the drive is cycling around the set speed but think this is inverter set up and will need to do more work on this. Have cured it by altering the Nmax number using an SPC function based on roll diameter which seems to indicate that load on the drive is having an impact as well. This generates a number in Hz from 0 to max which I then put into an SPC to drive output card Eurotherm good drives but Telemecanique usually 2/3 cost well to us anyway

[OkiePC 14]

I work with wind/unwind applications all the time. I've found the best approach for tensionless operation is to feed into a "loop" that is monitored with analog feed back. (Banner ultrasonic, or STI Valuscan are my favorites) I use the line speed as the master, and use PID on the analog feedback to "find" the ratio. The loop height is the PV and the CV is scaled to a ratio, let's call it R. R is clamped between known limits and multiplied by the line speed. If the line speed is zero, and the loop is empty, the result will be zero, but needs to be offset to give the line some material so it can start. So there's usually a rung to deal with that. Then this value is sent to the drive as a speed ref. I try to set the drive with a minimal accel and decel time, and adjust the PID for smooth ramping. It's also important to have good data for the line speed. I often have to use the speed command going to a VFD as my line speed feedback for another function. In those cases, I will ensure that the ramp rates in the drive match my PLC code to make the value as accurate as poosible. The result, once tuned, is that the "variable diameter unwinder" will quickly attain the correct loop height, and handle stops and starts and speed changes without too much upset in the loop height. My method is quite simple if your comfortable with PID. There's isn't a lot of math involved, you just have to calculate the minimum and maximum ratios needed by your application. Also, you don't need the ultrasonic measuring the diameter...you can move it to sense the loop height. Edited 5 Aug 2005 by OkiePC

[kcor 1]

which drives are you referring