General explanation about PID algorithm operation
1. PID regulator algorithm works always as a mathematical model function, with some differences and limitations depending on the hardware capabilities. Multi24 and FX PID regulators are not the same on the level of parametrizing: FX substation's PID has more parameters to set and thus is more sophisticated.Multi24's PID is more simpler and has fewer parameters. However, results of both operations may be considered as very similar in basic applications.
2. PID "stages" are dedicated for real building automation applications where could be several physical devices to control one goal parameter (e.g. temperature). PID regulator's "output" will be always just "rise the control level" or "lower the control level". If you have (for example) an AHU where "water cooler", "heat exchange", "water heater", "electrical heater" installed, this PID "output" shall be divided between all of those devices to control one goal parameter - supply air temperature. Thus, you need to set 4 stages with balance at 2. When PID starts, it will follow the temperature by setting all above devices to 0..100% each.
Each stage outputs may be supply to 0..10VDC controls. That explains deadzone and comfort zone operation: there is no reason to stop adjusting "cooling" or "heating" devices if setpoint is far from reaching - dead and comfort zones are applicable only when resulted temperature is near the setpoint and PID "output" is almost zero: "no need to rise neither to lower". Thus, you avoid excessive control device "rattling" when goal parameter is near the setpoint - valve will turn only if significant change needed.
P, I and D parameters may be set to any integer value, there is no limit like 0..100. In some big and slow systems (AHU for sport arena), "P" may be set to 5000 or more, "I" may be set to 600 sec or more because of long delay between changing of impact and feedback. In general, the faster the system, the smaller are P and I.
“P” is a proportional impact, in other words, how much heat shall be supplied to the system.
“I” is a speed impact, in other words, how fast the heat shall be supplied to the system.
“D” is an acceleration impact, in other words, how much you need to increase the speed of bringing the heat to the system.
Additional information about PID operation, its reaction on a input change (P), change speed (I), change acceleration (D) you can find there:
http://www.cds.caltech.edu/~murray/books/AM08/pdf/am06-pid_16Sep06.pdf
3. Basically, we usually set the PID regulator always "on" during the operation. If terms of Multi24, which has cyclic program structure, being once called in a program, PID is collecting information about input value and calculates the outputs continuously. If you wish to stop its usage, you should set outputs manually by some conditional operator.
However, when returning in operation, PID regulator will give out calculated outputs based on current circumstances: input temperature and how far it is from setpoint. Thus, you may face the situation where PID regulator output will be 100% from very beginning. There is no known way in Multi24 to correct this situation by parametrizing the PID.
FX substation PIDs (Control points) have way larger parameter number.
In FX substation, PID regulator may be set to "Balance point" when starting to operate. This parameter may be set to e.g. 50% and starting output value will be set to "middle" (depending on stages it could be e.g. Stage1:0%, Stage2:0%, Stage 3:50%, Stage 4:100%), if for example you are sure that system needs to be pre-heated before normal operation.
In FX substation PID there is "Indication point" where you can set the DI/DO to control PID operation. PID will keep the last output values when DI/DO equals 0.