How Can I Correct for Noise Using External Filters?

External Filters Provide an Effective Means for Addressing Noise, But Like Other Noise Compensating Options They Have Limitations

Choices, choices. In the realm of process control practitioners are regularly forced to choose between competing options. Consider a PID control loop: Should it be tuned for faster disturbance rejection or tighter Set Point tracking? Should the Derivative Term be used or does the PI configuration provide a sufficiently fast Settling Time? And the choices go on and on. In that sense there are multiple choices for filtering noise too – options that provide very different benefits. Fortunately when it comes to filtering for Signal Noise the choice is typically clear.

Another recent post focused on Internal Filters – algorithms available within many modern PID controllers that mathematically modify the associated Process Variable (PV) signal to correct for noise. As defined in that post: Noise is random variation in a signal. It is generally categorized as either high frequency Signal Noise or low frequency Process Noise. Whereas Internal Filters are built into the controller and address Process Noise, External Filters are generally hardware-based solutions that can be used to correct for Signal Noise. An unfortunate aspect is that as a practitioner you’re forced to choose a different solution for one or both types of noise. If the goal is to compensate for Signal Noise, however, then the answer is simple.

The following offers an overview of the two leading types of External Filters:

• Process Variable Filter
  

PV Filters are typically analogue devices designed to minimize high frequency electrical interference associated with low frequency process variability. They are limited to addressing Signal Noise which has marginal effect on a controller’s Proportional action. Signal Noise does not impact Integral action and filtering for Signal Noise only adds delay.

In newer smart sensors like those from Endress+Hauser the PV Filter is often built into the device itself. They are typically positioned between the sensor transmitter – or multiplexor – and the controller. By correcting for Signal Noise, External Filters like the PV Filter reduce variability which in turn reduces wear and tear on the Final Control Element (FCE).

• Controller Output Filter
  

CO Filters are also built into equipment but they’re on the receiving side of the controller’s signal. Whereas PV Filters smooth the signal fed directly to the PID controller, External Filters such as the CO Filter correct for high frequency chatter in the CO signal as it enters the FCE. As such they provide an effective means for correcting for signal Noise in general, and they are particularly useful in correcting for noise associated with Derivative action.

Like the PV Filter a CO Filter reduces wear and tear on the FCE. GE Industrial’s variable frequency drives (VFDs) and the Fisher series of positioners from Emerson are both recognized as highly effective solutions for filtering Signal Noise and improving control.

The Set Point (SP) Filter is another type of External Filter that was purposely excluded from this post. These External Filters are a valuable tool for maintaining effective process control. That said, SP Filters are not used to correct for either Signal Noise or Process Noise. SP Filters will be covered in a future post.