The one trivial change that can improve your process stability by 70-95%
There is one setting in a digital process control system that – if not correct – can move your process behavior from looking something like this:
To something more like this:
The setting you need to check is the transmitter’s damping constant.
The transmitter (or sensor) is the piece of kit that tells the control system what the Process Variable is at any one time.
The transmitter does this by sampling the measurement signal – temperature, pressure or whatever.
Most transmitters have a configurable filter. This filter is usually configured through a parameter called "Damping" or "Damping Constant". Its units are Seconds. This setting is vitally important – and relatively easy to get right
The problem you have to watch out for with the damping constant is “Aliasing”. If the measured Process signal contains noise (as almost all do) with periods less than the scan rate of the controller, we can get a phenomenon called aliasing.
What often happens is that the combination of transmitter sampling frequency and noise frequency combine to give the effect of a low frequency oscillation of the measurement signal. (a detailed explanation of why and how this phenomenon occurs is in the PID Tuning Blueprint)
The symptoms of this in a control loop are that the low frequency aliased signal will become superimposed on the controller's output, causing the Process Variable to oscillate at this frequency. This is almost always mistaken for instability due to PID tuning.
The solution as promised is dead simple: Set the damping constant in the transmitter to 1.3 times the PID algorithm scan-rate, and aliasing won’t happen. (why this works is covered in the PID Tuning Blueprint)
Check your transmitter’s settings – because if this setting is out then your tuning may be futile.
There are lots of other pre-tuning checks and hard earned, practical insights like this in the PID Tuning Blueprint. Each one can save days or weeks of frustration when trying to get the maximum performance from your loop.