I'm popping MOSFETS....linear derating factor involved?

Awhile back I made a half-hearted attempt and scaled a Level=7 (monolithic) device up into the ampere range to show it could be done, but Win didn't like it... complained about the ~0.8V threshold ;-)

...Jim Thompson

Point taken, I didn't look, but there's the time-honored case of servo-loop integrator windup to consider.

And an incorrectly-high transconductance. But those issues, I imagine, could be easily fixed if one understood level 7 models. Perhaps you could give us a little tutorial?

Remember that battery terminal glitching is better modelled by a switch, rather than by a changing voltage at the terminals.The voltage source is low-z, regardless of its voltage, the contact intermittence would be high z, outside of an arc.

I don't think the mosfet model is very relevent in getting first-order performance of current surges which are dominated by either the op amp slew rate or the gate circuit's ability to keep the gate pulled below a threshold and the stray impedances of the voltage source, depending upon the initial settings that set up bias on the gate.

RL.

Never saw your earlier statements about MOSFET models so I'm glad you were willing to say it again. :-)

I'll be adding RL's circuit modofications to my prototype in a day or two (crazy busy now) and will compare it to the sim. I will also do the transconductance curve plotting then. It'll be good learnin' for me.

I've also found your programmable slew rate controller (using a programmable transconductance opamp) and everyone pretty well everyone seems to have the LM13700 in stock. I also found Steven M. Sandler and Charles E. Hymowitz's article in the May 2005 issue of Power Electronics Technology with their updated SPICE model for a MOSFET.

LOTS and lots to learn. :-) Thanks Win, John

-- remove SPAMMENOT for e-mail responses --

I've also started taking a look at op-amp integrators and slew-rate limiters. :-)

John

-- remove SPAMMENOT for e-mail responses --

I would except I'm not sure the extraction can be done by hand. There are companies, for example, Silvaco, that just push the data into a computer and their software "UTMOST" spits out the model.

It might be possible to do it with Excel. I'll give it some thought.

...Jim Thompson