Is there any hope of using the mfr models in linear mode for ZVN and ZVP4525G? They simulate way too well to be true..
Now that C3503 and A1381 is dead there isn't much to choose from in the higher voltage range. But yeah, if you could buy on other than the hoard/salvage market I wouldn't even be looking. 🙂 It's never going to be a production issue so you might ask who cares, but I just don't like it.
30 years ago there were MOSFET SPICE models that had a transconductance greater than ID/(kT/q) at low drain currents, because the gate-source voltage to drain current relation went abruptly from near zero to a quadratic curve, without any weak or moderate inversion regions. A transconductance greater than ID/(kT/q) is definitely too good to be true.
I think these transistors may be getting almost that old, but I question how hard it would be to get the model closer by now. I guess they never had any hope of it being an amplifier so didn't bother. I can't even follow the circuit in the model yet to "fix" it the way they broke it apart. The MOSMOD1 line seems to make sense, so for now I'm assuming the problem isn't there.
It's easy to check in simulation. Just do a DC sweep: apply an ideal 10 V source between drain and source, an ideal voltage source that gets swept from 0 V to 10 V in 10 mV steps between gate and source, plot the drain current versus the gate-source voltage with a logarithmic current scale and a linear voltage scale, and see if the current increases faster than a decade per 60 mV anywhere. If so, it's wrong.