PID tuning (valve tuning)

What is PID tuning (valve tuning)?

Closed loop control algorithms are used to adjust valve responsiveness within Alicat flow or pressure controllers. Modifying the variables within these algorithms allows the user to control how fast a controller reaches a setpoint as well as control stability.

In this article, we will explain some of the theory behind valve tuning and give tips on tuning your own Alicat device.

When to tune your controller

Before shipping, each controller is tuned for optimal valve performance at the indicated process conditions. For example, if a controller is ordered to flow air at 25 SCCM, then it will be tuned with that in mind. That means if you use your controller for exactly what you configured it for, the valve tuning and therefore control stability should be pretty spot on.

However if process conditions change, for example if you decide to flow hydrogen instead of air or if you modify the inlet/outlet pressure, you may need to retune your device.

Valve tuning using PD/PDF and PD2I control algorithms

There are two control algorithms used to tune Alicat controllers:

  1. PD/PDF tuning for single valve controllers
  2. PD2I tuning for dual valve controllers (PCDs and MCDs)

Note: You can view your selected control loop by going to MENU > CONTROL > PID > LOOP TYPE on the front panel of your Alicat controller.

Valve tuning variables: proportional, derivative, integral

PD/PDF and PD2I control algorithms are composed of three terms each playing their part to minimize the error between the setpoint and the measured process variable. They are defined as follows:

  • Proportional (P): The proportional term decreases the error between the setpoint and reading.
  • Derivative (D): The derivative variable dampens the rate of change of the valve position.
  • Integral (I): The integral variable takes into account the sum of previous errors.

Note: You can view the current settings for your P, D, and I variables by going to MENU > CONTROL > PID on the front panel of your Alicat controller.

The PD/PDF Algorithm

The PD/PDF algorithm provides a simplified user interface.  It only requires the user to work with two variables: P and D.  The I control term is handled internally.  This substantially cuts the user’s level of difficulty.

When and how to use PD/PDF tuning

There are a couple telltale problems indicating a device in need of valve tuning. We discuss those problems below and provide the tuning solutions.

Problem 1: You notice oscillation around the setpoint or the control response is unstable.

  1. Decrease the P term in increments of 10%. The goal is to change P until the device quickly reaches setpoint with only slight overshoot.
  2. If there are still small oscillations, increase the D term in increments of 5-10%.

Problem 2: Your controller takes too long to get to the setpoint or never achieves the setpoint, but settles to a flow rate or pressure below the setpoint value.

  1. Increase the P term in 10-15% increments until you see the controller getting close to your setpoint.
  2. You can then decrease the D term to reach the setpoint more quickly. If you start seeing oscillations, D has been set too low.

The PD2I Algorithm

The PD2I algorithm was developed at Alicat Scientific for the express purpose of obtaining the highest possible performance in pressure control applications. It is also applicable to many flow control applications. As it utilizes proprietary predictive functions, this algorithm does not behave like a standard PID routine. Fortunately, PD2I tuning is simple if you follow a few simple guidelines:

  1. Set D = 10. This will work under almost all conditions.
  2. Initially, set the I term low. For example, I = 100. We will get back to this term later.
  3. Initially, set the P term to a moderate value. For example, P = 200. Now, drive the set-point to 25% of full scale.  If the process value oscillates around the set-point, increase the P  That’s right, increase. You may have to do this in fairly large jumps, until you get close to the right value.  (It is seldom necessary to exceed 10,000.) At some point, the oscillations will damp out.  If they do not, you may need to decrease I, even further.  (Do not go to zero.)  If you go too high on the P value, the oscillation is likely to transition to a more violent form, with large valves, this could even make a hammering sound! Once the oscillation damps out, start doing step changes to the set-point. For example, change between 25% and 75% of full scale, and back. The ideal response is to jump to something a bit short of the commanded set-point and slowly drift up to it.  If you overshoot, increase P. If on the way to the set-point, the process value first jumps towards the set-point and then backs away, decrease P.
  4. Once you have achieved the recommended response pattern, start increasing the I term, while doing the step changes described in item 3, above. As you increase I, you will note that the process value more quickly converges to the commanded set-point. You may continue to increase I until you start to see oscillation in the process value. Decrease I until the oscillation goes away.
  5. Now, exercise the controller through its entire range, making sure that the process value is stable at each point. You may have to adjust the P and I terms to ensure stability under all conditions.

These instructions should take care of you, in most situations. If you run into difficulties, please call or email the applications engineers here at Alicat.