Dancer Systems

[scottmurphy]

Unbelievable

Ok,
After getting the machine back late today, I though that I would go back to absolute basics, and try and simulate the previous system, that is, to control the axle directly off the dancer. So, all I did, was take the position of the dancer, and use that as a speed reference for the axle, no line speed reference, reel size, type of material etc.

And it worked. I ran the line up to 85m/min, and down again, the dancer moved slightly, but there was no hunting or bottoming out.
I only had a small reel, will go back tomorrow with a larger reel and see if the results are the same.

This is what really frustrated me, after all the theory, calculations and programming, and only 2 rungs did it. The problem that I have, is it is so crude, it is not funny. There is no line speed reference, no PID, just a reference directly related to the dancer position, when at preferred position, speed ref is 0, at max speed is max in direction of web, when at min, speed is max against direction of web.

So, all that is happening, is that it is like torque control, with the UW speed primarily being controlled by the pulling motor.

I haven't had a chance to put some decent thought into this method yet, I was too blinded by the fact that it worked so well. I still need to be able to match speeds with the new reel for splicing over, but this shouldn't be too bad to achieve.

I just do not like it as it defied the theory and calculations, plus it is so crude.

 

[jshiepe]

Sounds familiar....;-)

Congratulations Scott. I know there are some more complex applications of this process that warrant all the calculations, line speed reference, etc., but sometimes the KISS method does apply. Glad to hear you found a suitable solution.

Jeff

 

[Andrew Chan]

Genius configuration LM90

I came across this situation last week at out plant, we had a motor run by a Genius I/O and last week the Motor is stopped by it self couple time, then I went to look at the Genius block configuration and found no I/O address configured in to it; I reconfigured with the proper I/O address and never had a problem since then. My question is how the address got lost or wiped out? how is it possible, I think the base had a flash memory. Or is it possible to add a Genius block to a bus with no I/O address configured to it thru a Hand Held Monitor? When I looked at the original configuration documents (may created 10 years ago)its indicating no ref address needed. I am confused now, please any comments welcome.

Thanks.

 

[kamenges]

You're right, Scott. That doesn't make any sense. While I agree thata system like this can work and be stable, you would typically see some pretty significant dancer motion. I'm not referring to oscillation. I'm referring to the preferred run point.


It sounds like you are using proportional only control, that is unwind spindle speed command is based only on the error in dancer position. I am assuming your motor to load gearing requires that you use the majority of your motor speed range to reach line speed with an empty spindle.

Given this information, you would expect that with a full roll the dancer would start in the middle at zero line speed and move in the de-accumulation direction as the line starts. The dancer will pick a position where the error produces enough speed command to match the tangential speed of the unwind roll to line speed. Assume you stop at 25% of max line speed. The dancer settles into position. If you increase to 50% of line speed the dancer will move to a position that produces twice the error. This is required as you need twice the speed command to the spindle and all you have is dancer position error to drive speed.

Assume the line speed stays at 50% of max. As you run the roll diameter decreases. Assume you have run to half the original roll diameter, all at 50% line speed. When you reach 50% of origianl diameter the dancer will have moved to a settle point that produces twice the error again. This error position will have to double again when you get to 1/4 the original diameter.

If your system was originally designed to the very edge (max unwind spindle spoeed is requirewd to run a minimum diameter roll at maximum line speed) your dancer would need to be nearly bottomed out in the de-accumulation direction to get enough speed command for the unwinding roll to match line speed.

The inverse of the above description would apply to a torque mode system. The dancer would start with significant accumulation for large rolls and move toward center as the roll dimeter decreases.

Far be it from me to suggest you look a gift horse in the mouth, but if you are not seeing this there is something going on that you haven't accounted for. Either you have an integrator somewhere in the speed command path that is adding in speed for you or there is an additional speed command reference to the drive that you don't know about. As I said, since this system is working you may not want to dig any deeper. I just know this would drive me nuts since it makes so little sense.

Keith

 

[scottmurphy]

[but sometimes the KISS method does apply./QUOTE]

This is true, but this is so simple, that I just can't live with it.

Keith,
I don't know whether I can let it go, without adding some sort of superior control into the system.

It sounds like you are using proportional only control, that is unwind spindle speed command is based only on the error in dancer position

The only proportional control that was happening, is what the drive was doing. If the dancer was at mid position, then the speed ref is 0, and the speed is controlled by the web pulling it.

I am just about to run again, will see how it goes, and give some decent thought and try and convince myself that it will keep working this way.

Scott

 

[jshiepe]

Scott,

If it makes you feel any better, the VFFS machine that I mentioned in my original post a couple weeks ago has been working like this for years with no issues. The only thing we have ever done is add a little weight to the dancer to keep the film a little more taught, which greatly improved our tracking.

Jeff

 

[scottmurphy]

Jeff,

Thanks for your input, but control of the tension on this system is by air, and if the air adjustment is not fine enough, there is the facility to add weights as you have mentioned.

Just started with full reel, and with the system as crude as it was, there was too much unstability, since the dancer movement was so much, the speed corresponded.
I put this down to the reel being larger, hence heavier, and therefore, more torque required to overcome the load.

This was at zero speed by the way. But if it will not hold at zero, I don't fancy when accel /decel.


I will now put the PID in, and use it to control the dancer directly. Even though, I had it similar to this before, and the success limited.

 

[scottmurphy]

I am assuming your motor to load gearing requires that you use the majority of your motor speed range to reach line speed with an empty spindle.

This is correct, from the calcs that I have done, and with the core size, I get to app 74% of max line speed, bearing in mind of course, that the axle will never get to this size.

Far be it from me to suggest you look a gift horse in the mouth, but if you are not seeing this there is something going on that you haven't accounted for. Either you have an integrator somewhere in the speed command path that is adding in speed for you or there is an additional speed command reference to the drive that you don't know about

I don't know how exactly the existing system works, being 30+ years old electronics, is a bit grey. It definetly has the speed ref's into the controller, but these are needed for speed matching in splice control, they must be used for speed control in relation to dancer position, I will attempt to look at the circuit schematics again.

 

[jimtech67]

I only scanned the whole thread ( a little lazy ) so if this was mentioned already sorry

What brand drive are you using?? Many drives have macros or application software that can be added to make winders and unwinders work nicely with dancers and load cells.

I prefer SSD Drives (eurotherm) myself. All needed winding macros come with the drive out of the box

 

[scottmurphy]

What brand drive are you using?? Many drives have macros or application software that can be added to make winders and unwinders work nicely with dancers and load cells

The drive does have a lot of application blocks and features that can be configured for winding applications, I am preferring to use the PLC, as that way, I have all the informtion and data available in the same place, and can utilise / manipulate how I need to, and also present this information to the operator if need be.

Also, the PLC method, allows for non-drive experts ( myself included ) to attempt to see what is going on with real no data, rather than trying to diagnose a drive they know nothing about.

 

[jimtech67]

I hear what you are saying but.......

I went down that road of trying to program my own speed winder in a SLC500 and it never worked as well as the "canned" macros from the drive manufacturer did. Depending on the drive you can send and receive all needed info starting diameter, actual diameter, line speed, ect.... though analog inputs and outputs or if you want get fancy you can communicate directly to the drive with Device Net or something similar. And when you are done you will be a drive expert !!!!!!!

 

[scottmurphy]

I have looked at the drive only method, and was originally going to go that way. But the way that I was going to configure the drive, was based on the way that I would configure the PLC, the only difference that I could see, being that the update time of the drive would be different to the PLC.

I went down that road of trying to program my own speed winder in a SLC500 and it never worked as well as the "canned" macros from the drive manufacturer did

I can understand this, as the drive people put a lot of time into developing products to suit applications etc, and plc's are another way of acheiving the same thing. The difference being, when retrofitting, is that the time and backing is not in the same scale as the OEM's.

 

[kamenges]

Originally posted by scottmurphy:

Just started with full reel, and with the system as crude as it was, there was too much unstability, since the dancer movement was so much, the speed corresponded.
I put this down to the reel being larger, hence heavier, and therefore, more torque required to overcome the load.

While it is true you will need more torque to control the load your stability issue probably has more to do with tangential velocity than torque. You are controlling dancer position, which reacts to product roll surface speed. For a given spindle speed the product roll surface speed, and therefore dancer motion, is much greater for large rolls than for small rolls. You may need to run with lower correction gains with large rolls.

Did my last post make sense to you? When you ran the small roll did the dancer stay in one place at all speeds or was it simply stable? How you answer this will help us decide what is actually happening. Does the spindle drive have only one analog input; that being the speed command from the plc?

Keith

 

[jimtech67]

Program scan time was a problem for me as the unwind was a small part of a large program.


I found that a very good speed match between the unwind and line speed is the most important item. The surface speed of a empty core must match the line speed with all trim out of the picture. Also the surface speed at different diameters is a good thing to check with all trim disabled.

Calculate and accurate unwind diameter in the plc and this will make your life much easier

 

[scottmurphy]

I found that a very good speed match between the unwind and line speed is the most important item. The surface speed of a empty core must match the line speed with all trim out of the picture. Also the surface speed at different diameters is a good thing to check with all trim disabled.
Calculate and accurate unwind diameter in the plc and this will make your life much easier

I find the calculations the straight forward bit, putting into practice based on these calcs is the key. I don't think that I will be able to get the speed accuracy that I require, as the drive has no feedback, I am only using Armature Volts.

Did my last post make sense to you? When you ran the small roll did the dancer stay in one place at all speeds or was it simply stable?

Yes, it made sense. The dancer moved slightly when acc/dec, and then settled. I think that this position was slightly different to the starting position, however, I did not actually monitor this.

Does the spindle drive have only one analog input; that being the speed command from the plc?

Yes, I only have the one Analogue per spindle. And it is configured as a speed ref. The spindle drive has the facility to have up to 6 analogue IP's however. I could possibly add another OP from PLC, but am limited by the amount of IO available, as the spares have been reserved for the RW.