Copy a set of tuning numbers from one controller into another and the loop can behave completely differently, even though the numbers are identical. The reason is that PID controllers are built in more than one algebraic form, and the same three parameters mean different things depending on the form. This guide explains the three structures, parallel or independent, standard or ISA dependent, and series or interacting, why identical numbers produce different behaviour across them, and what conversions are needed when moving tuning between controllers.
PID controller forms in one line: The parallel, standard, and series forms are three algebraic structures for the same PID controller. In the parallel or independent form each term has its own gain; in the standard or ISA dependent form a single controller gain multiplies the integral and derivative times as well; and in the series or interacting form the terms multiply together. The same tuning numbers behave differently in each form, so parameters must be converted, not simply copied, when moving between controllers.
All three forms combine the same three actions, proportional, integral, and derivative, but they arrange them differently, and that arrangement changes what the tuning parameters mean. The parallel form, often called the independent form, gives each of the three terms its own separate gain, so the proportional, integral, and derivative contributions are computed independently and simply added together. Its appeal is that the three gains are decoupled: changing the integral gain does not touch the proportional or derivative action. Its drawback is that the parameters are less intuitive in engineering terms, since they do not correspond directly to the classic notions of proportional band and reset time in the way practitioners expect.
The standard form, also called the ISA or dependent form, is the one most control textbooks and many practitioners treat as the reference. It uses a single overall controller gain that multiplies everything, together with an integral time and a derivative time. Because the one gain multiplies the integral and derivative terms as well as the proportional term, the parameters are dependent: change the controller gain and you change the effective strength of all three actions at once. The advantage is that its parameters map cleanly onto the familiar proportional-band, reset, and rate quantities engineers reason with, which is why it is the common language for tuning.
The series form, also called the interacting or classic form, arranges the terms so that they multiply rather than add, reflecting how many early pneumatic and analog controllers were physically built. In this structure the derivative section feeds the proportional and integral sections, so the three actions interact: the derivative time influences the effective proportional and integral behaviour. Many older and some current industrial controllers use this form because of that heritage, and it has the useful mathematical property that its parameters are guaranteed real and it cannot produce complex tuning results, but the interaction means its numbers are not directly comparable to those of the other two forms.
The heart of the problem is that a parameter labelled, say, gain or integral time does not mean the same thing across the forms. In the standard form the single controller gain scales all three actions, so it interacts with the integral and derivative times; in the parallel form the gains are independent, so a number placed in the integral gain field acts alone. Enter the same value into what looks like the corresponding field of a different form and you are not reproducing the same controller, because that field participates in the arithmetic differently. The loop will respond with different aggressiveness, different overshoot, and possibly different stability.
The series form makes this especially stark because of its interaction. Since its derivative section modifies the effective proportional and integral action, a series controller with a given set of numbers behaves differently from a standard controller with the same-looking numbers, and the difference grows as the derivative time becomes significant relative to the integral time. When derivative action is small the three forms happen to converge and the discrepancy is minor, which lulls people into thinking the forms are interchangeable, but as soon as derivative action matters the divergence is real and can be large.
The practical consequence is a genuine hazard whenever tuning is moved between systems: copying numbers from a controller of one form into a controller of another form, without conversion, can produce a loop that is far more or far less aggressive than intended, sometimes to the point of instability. This bites when replacing a controller with a different platform, when a technician transcribes settings from documentation that assumed a different form, or when tuning rules derived for one form are applied to another. The numbers look right, so the error is easy to miss until the loop misbehaves.
Because the forms are algebraically related, tuning can be converted between them rather than merely re-guessed, and the safe practice when moving settings is always to convert. Standard-form and parallel-form parameters relate through the controller gain, and standard-form and series-form parameters relate through factors that depend on the ratio of derivative time to integral time, reflecting the interaction. The exact conversion formulas depend on the specific parameter definitions each vendor uses, so the essential discipline is to know which form both the source and the destination controller implement, apply the correct conversion, and never assume the fields are directly interchangeable.
This makes it critical to identify the form a given controller actually uses before touching its tuning, and vendors are not consistent, with some defaulting to parallel, some to standard, and some to series, and occasionally offering a choice. The parameter names on the faceplate are not a reliable guide, since the same words are used for different definitions, so the form has to be confirmed from the controller's documentation. Tuning rules and lambda or IMC formulas likewise assume a particular form, and applying them to a controller of a different form without translating the result is a common source of poorly tuned loops.
In a cloud SCADA operation that spans many controllers and possibly several makes of equipment, keeping track of which form each loop uses is part of managing tuning as an asset. When tuning parameters are recorded and reviewed centrally, noting the form alongside the numbers prevents a set of values valid for one controller from being wrongly transplanted into another. Across distributed field sites this matters most during equipment replacement and standardisation, where the temptation to carry old numbers straight across is strongest, and where a form mismatch would otherwise show up only later as a loop that overshoots or drags. Treating the form as an explicit part of every loop's tuning record is the reliable safeguard.
In the parallel or independent form each term has its own separate gain, so the proportional, integral, and derivative actions are decoupled. In the standard or ISA dependent form a single controller gain multiplies all three terms, so it interacts with the integral and derivative times. The same numbers mean different things in the two forms, so tuning must be converted, not copied, between them.
In the series form the derivative section feeds the proportional and integral sections, so the derivative time influences the effective proportional and integral action, and the three terms interact rather than acting independently. This heritage comes from how early pneumatic and analog controllers were built. The interaction means series-form parameters are not directly comparable to those of the parallel or standard forms.
Only if both controllers use the same form and the same parameter definitions. If the forms differ, copying numbers directly can make the loop far more or less aggressive than intended, sometimes unstable, because the same-looking parameter participates in the math differently. The safe practice is to identify each controller's form and convert the tuning, rather than transcribing the raw numbers.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.