This is a correct simulation result?

In so far as the results don't look too different from Figure 1 on page

They aren't realistic - a more comprehesive simulation would show the gate-source junction breaking down when the gate-source voltage exceeded 40V and the channel would vapourise fairly quickly with 20A running through it - but computer modelling is all about using over-simplified models to see roughly what is going on.

No, it's a good example of GIGO. Your model is WRONG. Don't blame spice when you use defective worse-than-useless models.

I wouldn't say that. Most real parts are even more wimpy than shown in figure 1, but basically the maximum current vs drain voltage is shrinking as it goes above 2A, as shown by the curve bending to the right. Even at high voltages the FET won't be able to conduct more than 4 - 5 amps. Compare that to the defective spice model, which shows the current gaily increasing to 22A and beyond, as if a simple Rds = 2.27 ohms was all that mattered. But as the datasheet shows in figure 2, Rds increases dramatically at high currents. They don't plot the curves very far, if they did the model's serious discrepancy would show more dramatically. That's one of the model's defects at high current. At low gate voltages the model isn't even in the same ballpark. It completely fails to model the threshold scene. All the detail we show for MOSFET behavior in AoE pages 121-123 is missed by the model. Anyway, as a result Boki gets GIGO.

I repeatedly say, vet your spice models with bench measurements.

"Boki" schrieb im Newsbeitrag news: snipped-for-privacy@v46g2000cwv.googlegroups.com...

Hello Boki,

To be honest, the simulation setup is nonsense! The MOSFET will not withstand a Vgs of 50V.

Try the following. .dc V2 0 10 0.01 V1 0 3 0.6

Garbage in -> garbage out.

Please read the datasheet before you setup a simulation.

Best regards, Helmut

But that is the model provided by that company, what can we do more...

Best regards, Boki.

The answer is simple. Don't rely on simulations. There's no substitute for

Graham

"Pooh Bear" a écrit dans le message de news: snipped-for-privacy@hotmail.com...

No. Don't blindly rely on crappy models. This case is trivial, but when doing serious analog design you have to check your models, know their limitations... and it's often that we have to make our own ones, carefully check them... This can be a non negligible part of the design time.

Can't agree more.