diode model and soft reverse recovery

2004-04-06
2004-04-06
  • Jon Engelbert
    Jon Engelbert
    2004-04-06

    "The diode models widely used in analog circuit simulators such as SPICE do not contain the  equations for forward & reverse recovery", Ma & Lauritzen. 
    I've become aware of shortcomings in the berkeley spice diode model because of some simple power supply circuits that fail with my simulator (which uses the berkeley spice diode). 
    Please forgive me if I don't get this exactly right... The "reverse recovery" of the diode involves a negative dip for the current as the voltage goes from positive to a large negative rapidly, followed by an 'exponential decay' shaped recovery of the diode current back up to its expected value (around zero amps)... somehow, this either occurs and/or is significant in the case when the diode is connected in series to an inductor. 

    I've read some articles on this topic, and some (Ma & Lauritzen) have proposed new equations, others (William Robbins) have proposed subcircuit models.
    Now, I've noticed that there have been significant changes to the diode in ngspice.  Does ngspice's diode model address reverse recovery?
    Does anyone know of diode source code that does address this? 
    Should it be a priority for the ngspice team... if so, I'll help with its development... just let me know.

    P.S.  Please reply to my email address (jon@beigebag.com) as well as to the forum.