33
2
Chapter 2 - Front Panel Operation
Theory of Autotune PID
4. Temperature Sensors and Autotune Optimization
Sensor type also affects the magnitude of the proportional gain value
calculated when using Autotune optimization. For sensors with high
sensitivity (a large change in sensor signal per °C), the resulting proportional
gain value will be smaller than if a lower sensitivity sensor were installed
into the same thermal load. Correspondingly, the integrator time constant
increases when using highly sensitive temperature sensors and the
differentiator time constant decreases.
Thermistors are temperature sensors that exhibit high sensitivity and
therefore require less propor tional gain for optimum performance.
Unfortunately, thermistors only perform well over a limited temperature
range and their sensor sensitivity is non-linear. This means that thermal
loads using thermistors are optimally tuned around one temperature but
performance degrades as the setpoint temperature is moved away from the
initial tuning temperature. The LFI-3751 compensates for a thermistor’s
non-linear sensitivity in temperature mode by recalculating the optimum
P
,
I
,
and
D
terms whenever the setpoint temperature is changed.
Linear sensors such as the Analog Devices AD590 and National
Semiconductor LM335, and Platinum RTDs exhibit poor sensitivity and
therefore require considerably higher proportional gains, but do not require
recalculation of the PID terms.
5. Thermal load variation and Autotune Optimization
Ideally, we would measure the load’s thermal lag time,
L
, the maximum
temperature versus time slope,
R
max
, and the load’s thermal time constant,
T
for a given output step current,
TE I
at any temperature and achieve
the same results. Unfortunately, most thermal loads are only linear over
a small range of temperatures.
Several factors cause a thermal load to respond differently from one
temperature to another. One such factor is a thermoelectric’s effi ciency,
which varies depending on the current being pumped through the
thermoelectric and the temperature difference between its cold and hot
plate. At some point, a thermoelectric’s heatsink temperature raises to a
point where it absorbs heat more slowly from the thermoelectric. Both these
factors affect the maximum temperature versus time slope,
R
max
.
Other factors such as ambient temperature change and power being applied
to an active device affect the thermal load’s lag time,
L
, and time constant,
T
.
Fortunately, PID controllers are relatively robust and still control quite well
even if they are not optimized. Depending on the size of your load, you may
fi nd it good practice to perform a PID Autotune each time the LFI-3751 is
powered on to compensate for changes in the thermal load.
Summary of Contents for LFI-3751
Page 10: ...10 This page intentionally left blank ...
Page 13: ...1 Chapter 1 Quick Start 13 1 ...
Page 20: ...20 This page intentionally left blank ...
Page 21: ...2 Chapter 2 Front Panel Operation 21 2 ...
Page 34: ...34 This page intentionally left blank ...
Page 54: ...54 This page intentionally left blank ...
Page 68: ...68 This page intentionally left blank ...
Page 69: ...3 Chapter 3 Rear Panel Operation 69 3 ...
Page 75: ...3 4 Chapter 4 Remote Interface Reference 75 4 ...
Page 88: ...88 This page intentionally left blank ...
Page 127: ...5 Chapter 5 Specifications 127 5 ...
Page 131: ...Appendix CAT 220 Cable Accessory Diagram 131 ...