3 Replies Latest reply on May 21, 2010 12:03 PM by Bill McEachern

    Power requirements in electronic thermal model.

    Ron Reiners

      Hi all,

       

      I am new to Simulation Flow. I am working on thermal model studies of our completed product assembly. Our product is a small electronic device enclosed in metal enclosure. We operate the device at an elevated temperature. This temperature is created by a heating circuit that is part of our product. When the unit is turned on the heater circuit works by applying a high wattage to a power transistor to create quick heat. As the temperature increases a temp sensor cuts back the power when a specific temperature is reached. As the temp sensor cools off it turn on more power to heat it back up. Because of overshoots and undershoot of the temp sensor the power level oscillates back and forth until the entire unit stabilizes at a specified temperature.

       

      I understand the heat source using time or goal dependencies but what I think I need is formula definition. I need to define a formula that indicates the maximum amount of power that can be applied, then use this max value to start up the heater, once a specific temperature of a point is reached it will need to cut back the power, when the unit cools down power is added. This power level needs to be increased and then decreased until a stable state temperature is reached. (Example: Start at 5 watts, cut back to 1W, increase to 3.5W, decrease to 1.5W, increase to 2.5W, decrease to 1.9W, then after numerous steps it stabilizes at 2.1 watts.)  Ultimately I would like Simulation Flow to tell me what the final stable state power level it.

       

      Is this possible? Does anyone do this or are there any manuals to tell me how to define these dependency formulas? Is there a better way to achieve this goal?

       

      Currently I am using a trial and error method to find the stable state power requirements and then back tracking to find the power up and down times to reach the steady state level as fast as possible. This method does not work well when testing multiple design configurations.

       

      Thanks

       

      Ron