# Current versus voltage not exponential above forward voltage of diode?

I have done a simple current over forward voltage simulation for a PN diode with LTSpice. I wonder why the curve is only exponential around the "knee-point" but it seems to be linear when going above that point. I assume that this is some effect that is modelled in the SPICE model because If I use a simple model that only models the diode with the two parameters Is and n, it seems there is no such effect. When I look into the 1N4148 datasheet, I see the curves are "saturating" on the logarithmic y scale -> is this the effect I observe in my simulation? If yes, what is the physical reason behind that effect and which SPICE parameter takes care of it?

• Try RS (series resistance) in the model. Commented Dec 6, 2021 at 8:43
• It would help if you plot your sim result also as log over lin to compare the two graphs. Commented Dec 6, 2021 at 9:12

I wonder why the curve is only exponential around the "knee-point" but it seems to be linear when going above that point.

If I adjust RS (a model parameter for diodes), I get these curves using micro-cap and the 1N4148 diode model: -

The shallowest curve is with RS at 0.65 Ω. The next steeper curve is with RS halved, then halved again, then halved again.

Of some interest might be the log plot of current vs voltage over the full range from 0 volts to 1 volt applied: -

The blue curve is with RS at 0.65 Ω the orange curve is a straight line reference. Blue is not too far away from the standard graph (shown in the question for the 1N4148). With about 0.5 volts applied, the standard graph suggests a current of about 100 μA and, this pretty much tallies with the micro-cap prediction.

With 0.8 volts applied, the standard graph suggests a current of between 10 mA and 20 mA. Micro-cap comes up with about 30 mA. Not too shabby but, pretty dependent on the model parameters of course. It looks fairly linear (log displayed current) between an applied voltage of 0.2 volts and 0.8 volts (see orange marker line).

• Oh man =) I could have expected some "resistance parameter" if things go linear... Commented Dec 6, 2021 at 9:27
• I think you might of LOL @Junius Commented Dec 6, 2021 at 9:30