README.optran

Adding transient simulation top generate the operating point of a circuit

The operating point of a circuit has to be established before a transient, ac, 
or noise simulation may be started.

The procedure, as found in cktop.c is as follows:
1) Iterate the matrix with the starting values
If this fails:
2) Start gmin stepping
If this fails:
3) start source stepping

If this again fails, bail out.

Step 1 may be skipped by
option noopiter

Step 2 may be skipped by
option gminsteps=0

Step 3 may be skipped by 
option srcsteps=0

Recently a fourth method has been added:
4) Start a transient simulation for a specific amount of time.
   Keep all data local. Keep the matrix values as an
   operating point for the simulation following.
If this fails, finally stop.

Operating point by transient is now standard. The fefault optran
data are 1 1 1 100n 10u. That is: first iteration, gmin stepping
and source stepping are set to ON. The optran step time is 100n,
the total optran time is 10u, supply ramping (not yet established)
is OFF.

You may change its performance by the command 'optran' in spinit,
.spiceinit or in a .control section.
The first three entries will override noopiter, gminsteps, srcsteps.
A typical command may be
optran 0 0 0 50u 10m 0
(no initial iteration, no gmin stepping, no source stepping,
stepsize for optran 50 us, optran run until 10 ms, no supply ramping

The time for transient is determined a priori. If a transient
simulation follows, choose 100 time the TSTEP value. If an ac
or noise simulation follows, take the inverted starting
frequency divided by 10 as the optran simulation time.

A method to stop the op transient simulation, when all nodes
have become stable, might be desirable, but has not yet been implemented.

The transient op calculation uses the 'UIC' flag for its startup.
Basically two applications come to mind: You may use this with
constant voltage and current sources and simulate until the 
output is stable. A method to automatically stop the op transient
simulation when all nodes have become stable, might be desirable,
but has not yet been implemented.
One also may use the optran for a first transient simulation, maybe
with different step sizes, before the next (tran, ac, pz, noise)
simulation is commencing. So you may reach a certain bias point
quickly and with coarse time steps, and then continue with fine
steps to see details. For example a power amp in electro-thermal
simulation may run for some seconds (minutes?) until the final
temperature is reached, and then the detailed transient or
an ac simulation starts.