1769-L32E PID Loop

Strange problem today. PID stopped working. Output stuck at 100% from PID.

Newly installed unit ran 1 night flawlessly.

CONFIGURATION
PIDEQ=independent
CONTROL=SP-PV
dev of=PV
loop update=.1
CV High limit=100
CV Low=0
dead=0

SCALING
unscaledMax=30976
unscaledMin=6272
Engineering max=20000
Engineering min=0
CVmax=100
TIEBACK
Max=30976
Min=6272

TUNING
P=20
I=.05
D=0

I have upgraded firmware to 17.4

Using a periodic at 10ms

.EN=1

All other functions seem to work on the controller.

rocksalt said:

Strange problem today. PID stopped working. Output stuck at 100% from PID.

Newly installed unit ran 1 night flawlessly.

CONFIGURATION
PIDEQ=independent
CONTROL=SP-PV
dev of=PV
loop update=.1
CV High limit=100
CV Low=0
dead=0

SCALING
unscaledMax=30976
unscaledMin=6272
Engineering max=20000
Engineering min=0
CVmax=100
TIEBACK
Max=30976
Min=6272

TUNING
P=20
I=.05
D=0

I have upgraded firmware to 17.4

Using a periodic at 10ms

.EN=1

All other functions seem to work on the controller.

Click to expand...

20 seems like an awfully high gain. Also, it appears like your Periodic setting and PID Update Time don't match. They need to be the same value.

What Analog Input Card are your using for your PV?

What are you controlling with this PID?

Stu....

20 seems like an awfully high gain:
Crude as it was, I trend the PV and SP and bump P up until it doesn't overshoot. As long as I start with a low enough KI I can get it close.

Also, it appears like your Periodic setting and PID Update Time don't match. They need to be the same value.
I will change that.

What Analog Input Card are your using for your PV?
1769-IF4OF2/A 4-20mA Pyrometer PV and 4-20mA out to an inverter controlling an induction coil to preheat a welding process.

I have been dealing with some induction units at 1500Kw/s and have not had a problem like this. I thought firmware bug but did not fix it. Then, how can it work 10 hrs straight and then stop?

Just some of my thoughts.

What are the CV limits set to ?

Could it be a case that it wound up past the Value 100 and when it needs to come back it takes forever to get back down to 100 or less so that you see the difference

Could you try to enter a value less than 100 to the CV and see how it is responding

I concur with Gil that you may have a PID wind up problem. As well, your scaling for your Analog Card might be wrong according to the AB Manual. Have a look at Page 4-6, PDF Page 50.

http://literature.rockwellautomation.com/idc/groups/literature/documents/um/1769-um008_-en-p.pdf

This and the mismatch between your Periodic Task could cause your PID to perform incorrectly.

Stu....

Further thoughs

I looked at your figures again and see the Tie back figures relate to the unscaled max 30976, but see your scaled max at 20000 and wonder if it should relate to 20000 instead.

I see that as a potential point for wind up to 150 % even though the output will limit to 100%

Is the 20000 the temperature ?
What is your setpoint figure and is it related to unscaled or scaled figure.

I looked at your figures again and see the Tie back figures relate to the unscaled max 30976, but see your scaled max at 20000 and wonder if it should relate to 20000 instead
They do not give a good explanation of tieback. Your thoughts would be appreciated.
Under the help it states the tieback should be the unscaled max and min.

I derived the counts from a signal generator at 4 and 20mA. I could not find the .pdf that listed the counts. Thanks for sharing that with me. I am not sure I should change it due to it matching the pyro readout for temp. If I change it, albeit slightly, the temp will be off.

I rescaled the pyro in the PID for easier numbers of setpoint. The numbers I wrote originally were some changes I made last night. Sorry about that. The original settings were 30976, 6272 for unscaled and engineering as well as tieback.

This card does not have the option to engineering unit on the card itself. I use the CPT to multiply and add to get the correct setpoint and temperature display.

What is important to know is that setpoint was below the actual in count PV (as shown in the PID dialogue box when PID running). This should make the PID go to 0. At this point the .en was high. I tried to write to %out (I think it's .so or something) and it changed back to 100% on it's own.

What is your setpoint figure and is it related to unscaled or scaled figure.
Scaled figure

CV limit when tested were set at 50%max0. and output was 50. It shows 100% in the PID box but the CVOUT variable was 50 as it should be.

I am going to download my changes this AM and see what happens.

It's working now.

I changed the unscaled min to 0 and tieback min to 0.
Also, on the PID instruction there is a "tieback" variable that is separate from the internal one that I set to 0. It's an optional parameter.

I changed this after it did not work from changing the periodic to match the loop update time.

Can someone explain the tieback?

sthompson said:

I concur with Gil that you may have a PID wind up problem. As well, your scaling for your Analog Card might be wrong according to the AB Manual. Have a look at Page 4-6, PDF Page 50.

http://literature.rockwellautomation.com/idc/groups/literature/documents/um/1769-um008_-en-p.pdf

This and the mismatch between your Periodic Task could cause your PID to perform incorrectly.

Stu....

Click to expand...

I used a signal generator and forced 20mA and 4mA and counts represented 30976 and 6272. I also used the output function of the pyro to force 20 and 4 with the same results. The card datasheet states 31085@20 and 6217.1428@4. Which numbers should be used?

rocksalt said:

I used a signal generator and forced 20mA and 4mA and counts represented 30976 and 6272. I also used the output function of the pyro to force 20 and 4 with the same results. The card datasheet states 31085@20 and 6217.1428@4. Which numbers should be used?

Click to expand...

I apologize, I misread your scaling as 0-20,000. Based on the Manual listed range, it works out 1554.286/mA, and your numbers are correct for 4-20mA.

Stu....

XRocksaltX said:

It's working now.

I changed the unscaled min to 0 and tieback min to 0.
Also, on the PID instruction there is a "tieback" variable that is separate from the internal one that I set to 0. It's an optional parameter.

I changed this after it did not work from changing the periodic to match the loop update time.

Can someone explain the tieback?

Click to expand...

Tieback is used when there is a Manual Station the manipulates the controlled device directly, and bypasses the PID. The Tieback is used for Bumpless Tansfer. This is different than Software Manual function. See Page 505 of the Logix Instruction Reference, or the Logix Instruction Help.

http://literature.rockwellautomation.com/idc/groups/literature/documents/rm/1756-rm003_-en-p.pdf

If you are not using this feature, it should be set to zero in the PID Instruction setup. Then the numbers in the Tieback shouldn't affect the PID.

If you have a Techconnect Contract, search for PID Tieback on the Knowledgebase. Some of them are marked for everyone, so all you need is a sign on. For example;

http://rockwellautomation.custhelp.com/app/answers/detail/a_id/46517/kw/pid tieback/r_id/113025


Stu....