Raghu:
I was able to achieve about +/- 0.5°F with the LUT scheme in a previous project that I implemented. My thermal response time was however quite slow because the heater in this case warmed a metal object to which the 1N418 diode was attached. The LUT table in my example was 32 entries and the range of the LUT values through the table corresponded to about 0->5°F of error.

You need to keep in mind the fact that the measurement resolution of the SENSE_VALUE must be at a finer resolution than the smallest step size in the LUT (which you generally have at the lowest indexes). Another issue is that there are "thermal inertia" characteristics of the system that must be considered. It can be that a heater that is too large can overwhelm any attempt at regulating the sensed temperature.

In your system there may have been other issues as well. The circulating fluid may have been a factor to the good or to the detriment. Placement of the temperature sensor could also have an effect, especially if there was a weird thermal gradient between the fluid and the heat exchanger plate where the sensor was located. The control system could also have been complicated by the pumping rate of the fluid if it was not constant (whether that was intentional or not).

Please note that I am an electrical engineer and not an expert in thermodynamics, but I have found the look up table system to work well for temperature control. Note that the LUT system that I had described works well where the thermal mass that one is trying to control is kept at a higher temperature than the normal static temperature from the nearby environment. If one is trying to heat something up to 70°F and the temperature of the surrounding environment is say 63->72°F (or higher) it is going to be rather difficult to control the temperature.

My most recent application of a LUT temperature control was to keep an LCD module glass from going below 0°C when the nearby temperature of the glass was detected at or below 0°C. LCD crystals can become rather slow to change state at cold temperatures and many vendors will not spec their modules below -10 or -20°C. When working with a military application where a display has to function to -50°C we found it necessary to install a heater wire in a serpentine loop on the back side of the display glass. Actually the heater pad was made up with two wire loops, one for the heater and the other wire was used as the temperature sensor.

An earlier application involved the stabilization of the mirror positioning motors on a laser beam galvo mechanism. The mirror position feedback was built into the galvo motors as a variable capacitance device where the motor rotational position was proportional to capacitance. (Note that galvo motors are designed to run in onl;y a partial turn of rotation - similar to the needle movement in an analog volt/current meter). The sensors were temperature dependent as well as the motor position was itself temperature dependent due to winding resistance changes from self heating in the motor. These factors were compensated by wrapping a heating pad around the galvo motor with its integrated position sensor. The motor temperature was controlled at a level that was at least 25°C above the maximum temperature that the motor would have operated at in a self heated mode of operation.

Hope this info helps
Michael Karas