options savecurrents_bsim3, savecurrents_bsim4, and savecurrents_mos1
will save all device currents as listed in chapts 31.6.1, .8 and .9
of the ngspice manual. Pure savecurrents will yield a subset of MOS1
currents, as seen before.
* Enable current mesurement for I source
* Look at only the first 2 nodes of S and W switches
and F and G controlled current sources, measure current
with = voltage source added to second node. So we get
I(Fx), I(Gx), I(Sx), and I(Wx).
Update the linear switch: add the limits to resistance ron, roff
Update the log switch: correct the resistance calculation for
von < voff
Add some examples for the pswitch.
If no .save or save command given by the user, add '.save all' internally.
To do so, check for .save or save.
Update to the .probe p command (power measurement)
Add probe_int_ to internal nodes and instance names, where appropriate.
These nodes will be disregarded when setting up the plot structure.
table lookup models (XSPICE) continuously outside the defined
rectangle or cube, as it may help convergence and avoid ambiguity caused
by rounding at the boundaries. See bug #591 "Wrong 2D table model output".
The value is that of the nearest point in the region while the reported
derivative is ramped to zero outside the boundaries.
Also fixes incorrect output at the upper limit in each dimension.
or control mode. ngspice enters interactive mode and waits for user
commands. Simulation result vectors are available up to the
(time) point of stopping. Simulation may continue by command
'resume', however now in interactive mode, a control flow stays
interrupted.
when TC or scale is used on the instance line of C and L.
Reason is the missing reset when CKTtemp is called once during each
simulation command. TC or scale is then applied again and again,
a reset is missing.
This patch adds a reset, i.e. the capacitance or inductance is
reset to its instance value in a call to *temp. It fixes bug#594.
Command 'alter' has been tested.
Resistance has not been tackled by this bug, because the line
here->RESconduct = here->RESm / (here->RESresist * factor * here->RESscale);
differs from
here->CAPcapac = here->CAPcapac * factor * here->CAPscale;
in that there is no accumulation of correction factors with here->RESconduct
to after numparam expansion, thus allow parameters in functions agauss etc.
Replace gettok by new gettok_np to ignore characters ( ) ,
which may be left over after parameter expansion.
This fixes bug 593.
which finds a token, while scanning over (), like
they were white spaces. It also proceeds with s while
still finding (), (which gettok_noparens does not do).
Holger Vogt
Giles Atkinson
Brian Taylor