INFICON MDC-370 Thin Film Deposition Controller - Manual (Page 92)

MDC-370 DEPOSITION CONTROLLER
soon as we’ve learned them, we know what we have to do to correct for errors
and we are back in good control. In other words the controller must compensate
for the characteristics or the “Plant”.
6.2 CONTROL LOOPS APPLIED TO VACUUM DEPOSITION
In the deposition control loop the vacuum system and evaporation supply make up
the plant. The output, deposition rate, is controlled by the source control voltage
which establishes the source power. If all plants were the same we could
predefine the characteristics of the controller for optimum control. Unfortunately,
plants vary widely, in their gain, linearity, response, noise and drift.
The question we are going to address here is how the controller adjusts the
sourcecontrol voltage, the “command signal”. The MDC utilizes a type 1 control
loop. A type 1 control loop does not require a continuous error to achieve a non
zero control voltage.
Many controllers utilize a type 0 control loop. In this type of loop the source
control voltage output is determined by multiplying the rate error by the
Proportional gain. For any given non zero output the error required to achieve the
necessary output is inversely proportional the to gain. High gain, low error, low
gain, high error. This would seem to call for high gain. Unfortunately, the higher
the gain the higher the chance of instability. We may go unstable before we get
the error down to where we want it.
In the MDC, the proportional gain parameter sets the rate at which the control
voltage changes in response to an error signal. Any error in the rate causes the
source control voltage to ramp to a new value. When the source control voltage
increases or decreases to the correct value, the value required to achieve the
desired rate, the error goes to zero and the output remains constant.
The Derivative Time constant is utilized to compensate for slow sources such as
boats and induction heated sources. Like a large truck, these sources take time to
get up to speed and to stop. The Derivative Time constant looks at the rate of
change of the error. If the error is decreasing rapidly we better take our foot off
the gas or we are going to overshoot our target. If the error is decreasing, but
decreasing very slowly, we need to goose it to get up to speed. The Derivative
Time constant instructs the controller on how much attention to pay to the rate of
change of the error. A value of zero tells the controller to ignore the rate of
change of the error. A large value tells the controller that this source is slow and
is going to be hard to get going and hard to stop. So if the rate starts to fall off,
give it power, or if we’re quickly approaching the target, begin to decrease the
power.
The Integral Time constant is used to keep the thickness profile on schedule. We
may have no rate error right now, so if we were not concerned about the thickness
profile, we would be happy and leave everything as it is. However if we are
trying to stay on a thickness profile, stay on schedule as it were, we may want to
speed up or slow down a little bit to make up for previously lost, or gained time.
For example, suppose our desired speed is 50 mph and that’s the speed we are
traveling. However we’ve been traveling for exactly an hour and we’ve only
Tuning the MDC-370 Control Loop 6-2