pid

help im useing a slc 5/04. im running a pid on a winder. as of now i run the pid all the time so my cv will build to 100%(before i start the winder). but i need it to start at 100% instead of the value it was last on. someone told me to look up bumpless trasfer and use that. but being im not the smartest i really cant figure it out. is it saying to change the p and i to 0 and then switch to manuel to move in the 100% cv. thanks freddie

That's the way I do it on different applications. Create a rung that turns on the Manual bit of the loop. Then whenever you want the loop to freeze (or start from a known value) put the loop in manual. Then move whatever you want as the manual setpoint. When you release the manual bit the loop switches to auto, taking off from the manual setpoint you last gave it.

To make it bumpless when going from Auto to manual you can move the output of the loop to the manual setpoint when the loop is not in manual.

http://www.plctalk.net/qanda/showthread.php?t=49866

Thanks for identifying your controller. It will be easier to follow this if you keep all your posts in the same thread rather than starting a new one each time you have something to add.

The concept of bumpless transfer between auto and manual involves adjusting the stored integral value as you change the CV in manual so that when you switch back to auto, the first solution of the PID equation results in the CV you've established. I don''t know enough about the SLC 5/04's PID to be able to tell you if that's feasible or not with your hardware, but luckily for you, Ron Beaufort is an AB guru. As long as he's available to help you out, you're in the best pair of hands you could ever hope for.

thanks for that info. so for i have tried switching to manuel and then moving 100% into my cv but when ever it switches back to auto the last value from auto is put back i cant make the pid start at 100%. i've tried many ways and found nothing to work.

do i need to give it my 100% before i switch to manuel.

When the PID is running in auto, you need to constantly move the PID CV value into a buffer, I'll call that buffer "Manual Value".

When you place the PID into manual, you need to use that "Manual Value" buffer as the control value, because you constantly move the PID CV value to the buffer when you transition the PID to manual, the output value will retain the last PID CV value.

When the PID is in manual you now need to constanly move that "manual value" to the PID CV, so if you manipulate the "manual value" using an HMI(or directly in the PLC), the PID CV value must be updated to match.

When you put the PID back into auto, the CV has now been preset to the previous manual value, thus creating a bumpless transfer.

BIG NOTE: When you do this, you must be sure you are scaling your PID CV and "Manual Value" properly!!

I have an old SLC instruction set reference manual (1761-6.15). On page 9-21 it states,

"In the manual mode, the PID algorithm does not compute the value of the control variable. Rather, it uses the value as an input to to adjust the integral sum (words 17 and 18) so that a bumpless transfer takes place upon re-entering the Auto mode."

That sounds like what you want to have happen. Are you making the rung containing the PID function block false when you go into manual? The same manual says on page 9-3,

"Normally, you place the PID instruction on a rung without conditional logic. The output remains at its last value when the rung is false. The integral term is also cleared when the rung is false."

Of the two threads you started on the same subject, this is the one in which you've posted more than once. I'll continue to use this one.

i have got the pid disabled when the winder stops. im changing every value i know to change like kc gain is 0 manuel mode, sp ,pv , and im hold the 100% in the cv for 3 seconds after i start using a timer. even with the pid false the old value is there when it goes true.

in an SLC-5/04 system, if the PID rung is ever executed as FALSE, then the Integral effect is forced to ZERO ...

so ... when you say "disabled the PID" do you really mean that you are allowing the processor to execute the PID as FALSE? ...

there are a lot of us here who are willing to help you - but you're not giving us enough to work with ...

once again (third time now) ...

can you post your program file? ... (note that you'll have to ZIP it first) ...

Click to expand...

if the answer is "no, I can't post it" then we'll go from there ... (or at least the other members will continue on - I'm offline now until Monday) ... if you CAN post it, then you can probably save yourself a LOT of time and trouble ... no offense intended, but your descriptions of what you're seeing aren't exactly "nailing things down" for us ...

the best thing to do is post the RSS program file and let the members take a look at it ... with the SLC-5/04 system there can be a LOT of "scaling issues" and "timing issues" and other things along those lines ...

good luck with your project ...

PS Edit ... in re-reading this, it MIGHT appear that I'm not trying to be helpful here ... I can assure you that I am indeed trying to help you solve your problem ... BUT ... probably 99% of the time when a PID "won't act right" the problem is in how it's been set up (as in BIG PICTURE type stuff) ... attacking the problem just one little piece at a time is going to be frustrating and time consuming FOR YOU ... you really need to post that program file - unless it's "trade secret" or "proprietary" information that your company won't let you post ...

gone now - wife is calling me to come home ...

thanks for all the help, finally got it i was using the scaled cv when i use my actual cv n7;25 it works goes straight to 100%. thank you all. i dont know how to post a program but i will try to figure it out i know it will be a big help in the future. i have a hard time explaining things , i thank you for your patience.