Greg, As already suggested it is better to use the OUTSEL block. We use a lot of the 'override' or 'constraint' controls. An important parameter that affects the behaviour of this scheme drastically is the LIMOPT parameter of the PIDA (we only use PIDA, but I believe the PID block has the same functionality regarding LIMOPT). If you leave the LIMOPT at 1 you might see some strange behaviour and tuning might not be made as robust as you would like. When leaving LIMOPT at 1 the integral action of the 'not-selected' controller will freeze as soon as it becomes 'not selected'. This means that the output of the TC could be very close to the output of the LC while the temperature is still significantly above setpoint, and as such, the P action of the temperature controller (which does not freeze) could already effect the setpoint of the FC when temperature drops but still is not below setpoint. We typically set the LIMOPT op both controllers (the LC and the FC in your case) to 3. This means that the temperature controller will move it's output away from the output of the Level controller so it will not already start to interfere due to the P action of the temperature controller, even though the temperature is still above setpoint. The smaller the offset on the TC becomes, the closer the output of the LC will 'track' the output of the LC and (if LC would have no offset), it will take over exactly at the point where the offset of the TC changes sign (temperature drops below setpoint). If you can spare the time I would suggest you build a small simulation test configuration, set-up some trend display (use foxanalyst if you have it) and do some 'step-testing'. Be aware that control is a 'dynamic' thing and therefore what you see does not always have a straightforward explanation. I guess in your case tuning of both loops will be quite different (TC versus LC) ? If tuning of both controllers is significantly different (e.g. you could have TC as PID and LC as PI with very long integral time) it will not make things easier but it should still work. You might find that, even though the temperature controller is above setpoint it will still be the 'selected' controller if the LC also has a significant offset. I have seen configurations where output limits of the LC and/or TC where 'back-calculated' (with a +offset for the High Limit and a -offset for the Low Limit) to make sure the LC and TC do not 'wind-up'. Al of this is really not necessary as the PID(A) algorithms take care of al of this. It should work like a charm! Success, Patrick Martens Total Raff. Ned. N.V. -----Oorspronkelijk bericht----- Van: foxboro-bounce@xxxxxxxxxxxxx [mailto:foxboro-bounce@xxxxxxxxxxxxx] Namens Gregory A Hurwitt Verzonden: dinsdag 23 oktober 2007 22:07 Aan: foxboro@xxxxxxxxxxxxx Onderwerp: [foxboro] cascade/constraint control I have a control scheme that will normally run as level-flow cascade. If the temperature of the downstream equipment falls too low, then the flow must be restricted. I plan to use a constraint control scheme: +-----------+ +-----------+ +-----------+ | | | | | | | LC | --------> | SIGSEL | --------> | FC | | | | (low) | | | +-----------+ +-----------+ +-----------+ ^ +-----------+ | | | | | TC | -----------------+ | | +-----------+ What are the recommended FBK and BCALCI connections for the LC and TC? How do I ensure that the TC does not wind up while it is not selected? ____________________________________________________________________________ __________ Greg Hurwitt Engineering Associate E-Mail: gregory.hurwitt@xxxxxxxx Postal Address: BASF Corporation 602 Copper Road Freeport, TX 77541 USA BASF - The Chemical Company _______________________________________________________________________ This mailing list is neither sponsored nor endorsed by Invensys Process Systems (formerly The Foxboro Company). Use the info you obtain here at your own risks. Read http://www.thecassandraproject.org/disclaimer.html foxboro mailing list: //www.freelists.org/list/foxboro to subscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=join to unsubscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave _______________________________________________________________________ This mailing list is neither sponsored nor endorsed by Invensys Process Systems (formerly The Foxboro Company). Use the info you obtain here at your own risks. Read http://www.thecassandraproject.org/disclaimer.html foxboro mailing list: //www.freelists.org/list/foxboro to subscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=join to unsubscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave