[Gnucap-devel] JFET and initial conditions

[Xavier Romeuf]

Hi,

While developing the Python wrapper for Gnucap, I have encountered an
issue with the JFET ngspice model.

Here is a netlist containing a JFET component based on the ngspice model:

-------------------circuit.ckt--------------------------
*** NJF Test ***

.attach jfet.so

.model J2N3819  NJF(Beta=1.304m Rd=1 Rs=1 Lambda=2.25m Vto=-3 Is=33.57f
+ Cgd=1.6p Pb=1 Fc=.5 Cgs=2.414p Kf=9.882E-18 Af=1)

R3 N528817 N528869 4k
I1 N528855 0 DC 6.1m
R4 N530614 N528973 50
R2 N528849 N528869 4k
V2 N530614 0  SIN 0 1 1e9 0 0
J2 N528817 N529031 N528855 J2N3819
R5 0 N529031  50
J1 N528849 N528973 N528855 J2N3819
V1 N528869 0 18

.END
--------------------------------------------------------

Doing a transient analysis on the node N528849 gives the following
results, which are similar to those given by Spice:

Gnucap> get circuit.ckt
Gnucap> print tran v(N528849)
Gnucap> tran 0 1n 0.1n
#Time       v(N528849)
 0.         5.8
 100.p      6.1586
 200.p      6.3693
 300.p      6.2136
 400.p      5.7429
 500.p      5.1405
 600.p      4.6337
 700.p      4.4201
 800.p      4.5832
 900.p      5.0749
 1.n        5.7175

However, if I add the "uic" option, the initial voltage becomes 0,
although there is not any specified initial condition:

Gnucap> get circuit.ckt
Gnucap> print tran v(N528849)
Gnucap> tran 0 1n 0.1n uic
#Time       v(N528849)
 0.         0.
 100.p      0.64275
 200.p      1.0197
 300.p      1.0951
 400.p      0.88528
 500.p      0.49818
 600.p      0.11335
 700.p     -0.090857
 800.p     -0.01006
 900.p      0.34247
 1.n        0.84256

Shouldn't it have given the same results?
Can the initial voltage of N528849 be defined to another value than 0?

Regards,

Xavier