Positioning Control

pauly

Member
Join Date
May 2002
Location
South Wales,U.k
Posts
244
I've done a couple of projects whereby a load controlled by VFD has had to be positioned and the method used so far has been with simple comparators in the PLC and slow down speeds. I am looking for a better more "engineering" method where the speed of the drive is continuously controlled but as my maths skills are not what they used to be I am struggling. Any pointers would be gratefully reseived.
 
don't do positioning loop in PLC, it's too slow, use drive for that.
get VFD with encoder and let plc communicate with the VFD over fieldbus for example (so you know current position etc).
 
Originally posted by panic mode:

don't do positioning loop in PLC, it's too slow, use drive for that.

That's not necessarily true. If your allowable positioning times are relatively long you can do this in the plc. If the process currrently works using triggered speed switches from the plc a more 'continuous' control out of the plc will also work. Granted, you won't be able to do a semiconductor pick and place application this way. But many applications don't require 250 millisecond moves and 0.0001" accuracy. Whether it will work is very application specific.

Keith
 
The PLC may be too slow to use a traditional PID type position control. I would not do that either.

What I have done with great success is calculate a ramped up/down speed reference that is based on position error. Sort of like the P in PID but with customization to take into account the lag time of the controls, and allow for tuning the accel/decel distance, top speed, in position tolerance, minimum "stall" speed, whatever the system requires.

It can be a very simple and cost effective approach to positioning, and you'll be surprised how responsive it can be when fine tuned...not a servo controller, but a "poor-man's positioner"...
 
kamenges said:
That's not necessarily true. Many applications don't require 250 millisecond moves and 0.0001" accuracy. Whether it will work is very application specific.

True, but still not worth it. I've done couple of pick and place machines that had positioning in the PLC instead of motion controller and I know that it can be done. But even though it can be done, it's not right tool for the job. It is just not worth the trouble. Drives with built in positioning are readily available and they are cheap. What else is to be said?
 
panic mode said:
True, but still not worth it. I've done couple of pick and place machines that had positioning in the PLC instead of motion controller and I know that it can be done. But even though it can be done, it's not right tool for the job. It is just not worth the trouble. Drives with built in positioning are readily available and they are cheap. What else is to be said?

In my case, I already had position feedback in the PLC5 and A/B 160 series analog drives. There were 3 of these setups per machine times 19 machines, so it was very worthwhile to make a simple program change that cost zero dollars and greatly improved accuracy. It also allowed me to virtually eliminate the wear on the mechanical brakes that were previously responsible for position repeatability.
 
I'm right now comissionning system with positioning. PLC is ControlLogix, drives are Danfoss VLT 5000 with positioning card. Encoder, overtravel limit switches and home prox are wired directly to this card and all you need to do from PLC is send a target position and raise "Go" bit over fieldbus (DeviceNet in my case). Works very good. I'm not sure weather I would have been able to do it in PLC - I need to travel about 1 meter within 2 seconds max.
 
Originally posted by panic mode:

Drives with built in positioning are readily available and they are cheap. What else is to be said?

You are still not looking at this as an application specific item. If I am doing one axis and it takes me 10 more hours to program this in the plc than it does to use the drive internal positioner then the drive internal positioner is the way to go. The money value of time would win.

But look at OkiePC's point. He had 57 axes to deal with. He already had the PLC and associated I/O. He knows he can accomplish the task in the PLC based on his application. He figures out one axis and the result is repeated 56 times. Why would he add the cost of a drive internal positioner in his case? It makes very little financial sense.

Then take it the next step and look at it like a high volume OEM. We aren't talking 19 machines, we are talking 200 or 300 machines, 600 - 900 axes. If your application can support it why wouldn't you do this in the plc?

Keith
 
Another way I used about half a year ago is step motor with hight speed uoput on MicroLogix. Travel distance was about half meter and time was about 25 sec (cutting of rubber strip). Worked fine as well. The only problem is absence of feedback. You have to trust your pulse train output information. I have to use end of travel sensors to confirm end position...
 
So far we just have opinions. They don't count, I have some questions.

How do you prove that an application can be done in a PLC or a motion controller for that matter?

So far I like the advice that one should use the PID control in the VFD drive.

There is a lot of difference between velocity control and position control. Position control is harder. Why?

I am looking for a better more "engineering" method where the speed of the drive is continuously controlled

Is there a ROI?
Is the machinery and feedback worthy of a motion controller? Some machinery is so screwed up with sloppy gears and such that the first dollars should be spent on the machinery.

but as my maths skills are not what they used to be I am struggling. Any pointers would be gratefully reseived.
This is one of the main reasons to get a motion controller because a lot of math is done for you. The PID in the drive is not good enough unless you also have a motion profile generator. If your math skill are not good then you will be stopped by the motion profile generator.

How does one guarantee a scan time wit a PLC?
How good can the motion profile generator be?
How good are the diagnostics? This is very important.


 

Similar Topics

We have to do a positioning with an incremental encoder on the motorshaft that delivers 6 pulses per rotation. One rotation of the motor is 22.8...
Replies
5
Views
2,187
Anyone can share their experiences on performing precision indexing with PowerFlex using solely an absolute rotary encoder attached to the load...
Replies
2
Views
2,995
help please... i'm planning to control two stepper motors for a particular x-y motion and i want to use the omron cpm2a cpu. from the specs, it...
Replies
1
Views
1,772
Are there any S5 gurus in here? I have searched the forum for any info about the IP266 module, but found only one thread :( We have some banding...
Replies
4
Views
11,166
Hi everybody. I’m working on system controls blade position in cutter. Position of blades is actuating by hydraulic cylinders. Hydraulic valves...
Replies
10
Views
4,517
Back
Top Bottom