MKS T3PIA, Instruction Manual, Page: 34 / 70

Chapter Three: Product Overview Control Mode
26
Model Based Pressure Control
The T3P controls the chamber pressure by utilizing a Model Based Control technique. The valve position is
adjusted at each control cycle to regulate the valve conductance and, thus, the chamber pressure itself. An
accurate knowledge of the system parameters is required for the optimal operation of the pressure controller.
These parameters include the chamber volume and the valve conductance.
LEARN Function
The purpose of the LEARN function is to obtain the chamber pump speed versus valve position using the real
operating conditions of sensor type, chamber volume, inlet flow, and pumping system. The LEARN function
should be executed prior to normal valve operation. The LEARN function is generally required only once
during initial setup/installation. The sequence of steps required to LEARN the valve is as follows:
1. Enter the chamber volume in liters utilizing the valve software interface. An approximate value for
volume may be used provided the volume estimator is turned on (see the following step).
2. Determine appropriate setting for the volume estimator:
a. If you are confident of the actual chamber volume entered, turn off the volume estimator.
b. If you are unsure of the actual chamber volume, turn on the volume estimator. The volume estimator
will determine the volume during the system learn and use this volume for pressure control.
3. Set correct pressure transducer ranges. Refer to the MKS Type T3BIA/T3PIA Valves with RS-232 Interface
Supplement , as necessary.
4. Obtain the recommended learn flow setting from the valve software.
5. Turn on flow to the desired rate for the duration of the LEARN test.
6. Enter the actual flow rate that is being used.
7. Start the LEARN function.
The flow rate should remain constant until the LEARN function is completed. The LEARN function moves
the valve over a set of non-uniformly distributed positions and records the pressure data from the valve high
channel transducer. The typical duration of the LEARN function execution is less than 45 seconds. The
pressure data versus time resembles Figure 5. If the actual flow is significantly lower than the recommended
flow (dotted line), the accuracy of the obtained chamber pump speed may be insufficient. If the actual flow is
much higher than the recommended flow (dashed line), the pressure transducer saturation may occur. This
will result in incorrect pump speed curve and will negatively affect the quality of pressure control.
Adjust the flow, if necessary, and rerun the system learn starting with Step 5 above.