Sorry, another PID question!

[MartinW]

Hi,

I've been searching the Internet for days now and there's so much confusing information about PID terms. I'm not new to the general concept of PID but lacking the hands-on application...

I'm using Siemens FB41 to control a motor using a dancer roll. The motor drive is in speed control.

The SP is 0 (i.e. dancer in the centre of its travel). The PV is the dancer input, which is +/- 5V ish.

Why is the integral term in Siemens application called "reset time"? I'm assuming from my testing that a smaller reset time means that the integral term ramps up more quickly but that it also "unwinds" a lot quicker when the PV overshoots the setpoint.

I'm also using D for the first time on this project. I understand that it's the ramp of the PV etc. I'm using a value of 2 Seconds which seems to work ok, but I would welcome some guidance here - what does 2 seconds actually mean? If I made it say, 10 seconds, what effect would that have? Would it damp more aggressively or less?

Thanks,
Martin

 

[nettogrisen]

As always when people ask about PID...

Read these posts by Ron Beauford


What is P in PID?:
http://www.plctalk.net/qanda/showthread.php?t=13459&highlight=pid

What is I in PID?:
http://www.plctalk.net/qanda/showthread.php?t=11242

What is D in PID?:
http://www.plctalk.net/qanda/showthread.php?s=&threadid=11969

 

[Roy Matson]

Unfortunately the PLC manufacturers dont always follow the classic PID terminology. Reset is normaly either in
Repeats / minute or the inverse minutes / Repeat
A Repeat is normally related to the Proportional band or it's inverse Gain
For example if you have a setpoint of 50%, PV of 55% and Gain of 1 a repeat would be 5%. If the reset time was 10 seconds, every 10 seconds you would get an output change of 5%.
Hope this helps a little
Roy

 

[Ron Beaufort]

thank you, Jesper, for the kind compliment ...

to MartinW ... the material that Jesper mentioned is also available on my website in a handy PDF format ... just look in the Sample Lessons area ...

note that I only deal with Allen-Bradley - but I hope that the information will be helpful to you ...

and welcome to the forum ...

 

[MartinW]

Hi all,

The info posted so far is most helpful. It gives a lot more insight into how the actual values work. Does anyone know if Siemens (particularly FB41 in a CPU318-2) works in a similar way to the AB methods Ron has described?

Thanks again.

Martin

 

[Pandiani]

You can also look at this for reset time

 

[MartinW]

Another related question....​

What's TM_LAG and how does it work? Is it the filtering action on the derivative term described in Ron's text as a "feature" of AB's implementation?​

Thanks,​

Martin​

 

[MartinW]

Just answered my own question with 5 mins and PLCSIM. And for those looking for enlightenment of their own, TM_LAG does seem to control how quickly the derivative term decays off after the PV has settled.

In fact, that would appear to make it the opposite of the AB implementation. That filtered UP to a given value. This seems to reach it's value instantly but then take a while to decay off.

Anyone for a debate on why or how this affects processes?

 

[MartinW]

For some reason I can't edit my last post.

Clarification - FB41 filters both ways. And so did AB. My bad....

 

[JRW]

Martin,

Im curious regarding your application. Are you going to take the speed actual from a previous nip roll (or something) and trim this section or are you trying to control this section (winder or roll) with 100% control?
Do you have a picture or block diagram you would like to share?

 

[MartinW]

System diagram

Here's the diagram.

Basically, the material is unwind from a mother roll of material, over a dancer roll, via a series of idler rollers to a winding drum (main drive) which is motor driven and is the machine's speed master.

The main speed reference for the unwind PID is taken from the speed feedback from the main drive

An LVDT connected to a dancer roll provides tension feedback, and is the PV for FB41. The connection to FB41 is via the PV_PER input and PV_FAC is adjusted to that PV is 100% when the LVDT gives 5V (the maximum travel). (obviously, PVPER_ON is ON)

The PID produces an output +/- 100%. This is multiplied by a % of maximum line speed to give the maximum amount of trim the PID can produce, in real units. In this case, m/min. This gives a "trimmed linespeed". The reason that this is done before converting to a rotational speed demand is that this way, you get more effective output at smaller diameters, than you do at larger diameters. I've seen other systems which multiply the P by diameter, diameter squared, or diameter cubed to produce the same effect.

The trimmed linespeed is then converted to an RPM demand.

The machine operators have a switch which changes the unwinding direction of the material on the unwind.

The drive (Siemens 6RA70) is in speed control.

(By the way, the + in the diagram is actually a -)

 

[JRW]

Makes perfect sense Martin-
I have a block written in SCL that might be a good starting point for you.
I personally would do this application in the 6RA70 if it was
me but regardless, if you want the unwinder/winder block I mentioned- just PM me with your e-mail address.
Again, the blocks written in SCL but its a starting point and very close to everything you are doing.