¿ªÔÆÌåÓý

power dependent resistor - thermodynamic model


 

Hello,

I'm trying to simulate (transient run) a power dependant resistor, the resistor is used to dump an high energy capacitive bank (like an RLC circuit). During the fast discharge, the resistor temperature rises due to Joule effect and it's value change.
I think to use behavioural source to model the resistor.
What I'm not able to do is to get the power variation (between 2 time steps ?) in order to calculate the temperature rise: delta RxI? = mass x heat capacity x delta Temperature. Then i will use the temperature to compute the resistance (Temperature don't need to be linked with the LTSpice Temp variable).

Any tips ? Even if it's not the appropriate method I'm sure that ltspice is able to do it !
This is a common for power electronics ...

Thanks for your help or comments !!

Micha


 

Hi Micha.
In this group, there is information dedicated thermistors. There is an option which takes into account the inertia of thermal processes. Look thermistors "NTC" and "PTC".
You need to replace the functional dependence of the temperature of the resistor.

Bordodynov.


 

Thanks Mr Bordodynov for the tips, i saw enough information in the lib section to accomplish the first step.


An other question: Anyone know if there is a way to link ltspice with an other program to work in a co-simulator mode ? I would like to implement physical model that can be complex to define just with an electrical solver (for example, mechanical actuators or light intensity emitted from a flash lamp, ...)

For example, LTSpice will be use "only" as an electrical solver and the other software (not a brand soft like matlab, but something developed by myself with C-code) will compute all other physical parameters and will change some values in the spice netlist between time steps.

I know Synopsis Saber is able to do it and also matlab/simulink with cadence but it?s not in my budget :-)


Thanks.?



 

Hello,


As far as I know you cannot link user code to LTspice.? However, LTspice can be used solve ordinary differential equations, difference equations or z-domain problems and most other algebraic equations.


Rick




 

I will try to run LTSpice with command line and use .savebias/.loadbias to restore all nodes settings from previous run. Labview (or Scilab ) will compute physical parameters and update the netlist.


I think I will face some convergence issues ... Let?s see.


Regards.


Micha


 

Micha,

I don't think .savebias/.loadbias will help you much, but they shouldn't cause an convergence problems.? You need to explain more clearly what you want to do.

Rick