Hi,

fine tuning is trying to create an oscillation of the process value around the setpoint by modifying the PID block output accordingly.

This means by modifying the output value between the outptut limits, it must be possible for the PID block to make the process value cross the setpoint in both directions, upwards and downwards.

To give the tuning algorithm a chance to influence the process value easily, you should select a tuning setpoint that requires a steady state output value with some distance to the ouput limits e.g. something between 20% and 80%.

I am not sure if this is the case in your scenario.

In the first screenshot, that you shared, the output value is constant 0% for around 10 minutes and the process value is above the setpoint all the time and seems not to change.

How should the tuning algorithm make the process value cross the setpoint in the downwards direction, when the output value is already at its lower limit and process value is still above setpoint?

Maybe you could try a different (probably higher) tuning setpoint?

If you do a recording of this, please choose a longer time interval than the default 0.3 sec so that we can see a longer period than 10 minutes.

The "noise" on the process value should be no problem for the controller and its tuning. There is a specific tuning phase where the standard deviation of the process value is observed to consider it for the following tuning phases.

Regards

Mach11

Hi,

fine tuning is trying to create an oscillation of the process value around the setpoint by modifying the PID block output accordingly.

This means by modifying the output value between the outptut limits, it must be possible for the PID block to make the process value cross the setpoint in both directions, upwards and downwards.

To give the tuning algorithm a chance to influence the process value easily, you should select a tuning setpoint that requires a steady state output value with some distance to the ouput limits e.g. something between 20% and 80%.

I am not sure if this is the case in your scenario.

In the first screenshot, that you shared, the output value is constant 0% for around 10 minutes and the process value is above the setpoint all the time and seems not to change.

How should the tuning algorithm make the process value cross the setpoint in the downwards direction, when the output value is already at its lower limit and process value is still above setpoint?

Maybe you could try a different (probably higher) tuning setpoint?

If you do a recording of this, please choose a longer time interval than the default 0.3 sec so that we can see a longer period than 10 minutes.

The "noise" on the process value should be no problem for the controller and its tuning. There is a specific tuning phase where the standard deviation of the process value is observed to consider it for the following tuning phases.

Regards

Look at my sxy noiseless processed signal . Noise in a closed loop creates lot of problems.
So much to explain in detail. To brief it . If you give mad (Zigzag) signal to PID it is forced to act mad, suffering to survive against your raw signal consuming lot of energy.
PID needs clean signal . Not Noise. Noise dominates the main signal leading system to oscilate to be unstable. It affects more to small systems which is visible . But large process like one I'm working on it is not that much visible but consumes more energy in background than it is supposed to be. . 

Hi,

I repeatedly getting this error during finetuning. Pls suggest wat to do about it. 

GaJEnDrA D S

Hi,

I repeatedly getting this error during finetuning. Pls suggest wat to do about it. 

 Im not able to do fine tuning anymore. Why pretuning is totally not possible throwing error that setpoint is too close to input. but my setpoint is beyond  50% of actual.. 

Hello.

Why nobody is responding?

Pls ｒｅｆｅｒ　ｔhe image above . Why theres are spikes of PID　output for each step drop in Process value. It's because of parameter D in PID. How to avoid it? What's the cause? 

Hi,

regarding the requirements for starting pretuning:

This is described in online help.
You can also find the description here:
Pretuning V2
Parameters State and Mode V2

Regarding your Output spikes:

It seems that you are calling the PID_Compact in shorter cycles than you receive new process values.

In the cycle where the process value change happens, the derivative action of the PID algorithm will respond with an output change. For the next (around 8) cycles the process value remains constant and the derivative action reduces (derivative of a constant signal is 0 so the derivative action moves with delay in this direction).

To avoid this, I would recommend to call the PID_Compact in the same cycle time like you receive new values.

Regards