|
How TSTEP Affects a Transient Simulation
How does the value of TSTEP affect my transient simulation?
3 J; V0 n, ]1 a6 a R- G! q1 m& ?TSTEP is specified in the transient analysis netlist command. r/ N# u! \/ N* @7 n
.TRAN TSTEP1 TSTOP1 TSTEP2 TSTOP2 ... TSTEPN TSTOPN
. V! |* {# I8 [5 m) kThe most common usage is a single TSTEP/TSTOP pair. For example:
/ A& I( j5 U5 z& E! u) ~.TRAN 0.1ps 100ns
# O! ^) n: F1 l( d' q' hTSTEP has a variety of effects on the operation of HSPICE.: q: G9 N' \' t) [
1. When the RUNLVL algorithm is in use:
& y. l0 c# n% u7 f% n0 ]% l - RUNLVL "in use" is defined as HSPICE version >= 2004.09 and OPTION RUNLVL > 0.7 a+ y7 `6 H' a& u/ E+ H; L) |
For HSPICE version >= 2006.09, RUNLVL=3 by default.+ y4 t, p/ @6 H; H" V" A, V
- The minimum timestep is determined by the RUNLVL algorithm. The RMIN option is ignored.
: r( i5 ^& S2 F4 m) {; z& o - The maximum timestep has a "soft" limit of RMAX*TSTEP. The RUNLVL algorithm
: U6 a" t$ o. F2 V g: @6 q allows the timestep to exceed RMAX*TSTEP as long as tighter Newton-Raphson
3 S$ {% j" {, d: k% f tolerances are met (voltage, current, charge).The default value of RMAX is 5.# a) ], n5 B, y, P8 ?
- The DELMAX option can be used to override the RUNLVL algorithm's choice of; s+ g, `! L3 @3 n
maximum timestep. When DELMAX is used, TSTEP has no impact on simulation accuracy (except
5 V+ A! r9 @7 x8 x4 D: m( l6 k when used in conjunction with .OPTION INTERP, see #3 below).& ~/ Q" U: } @! w0 t4 t& l' l- t
2. When RUNLVL=0 (discouraged):
% y! o0 y, ^3 e0 d - The minimum timestep is defined as TSTEP*RMIN. The default value of RMIN is 1e-9.
) Z8 Q- c2 Z3 G+ a1 c# d - The maximum timestep is defined as TSTEP*RMAX. The default value of RMAX is 5.: H, ^3 l8 W* C5 o! G; W6 K
- These limits are "hard" limits./ y, Z2 P2 t1 I. M9 e
- If the simulation engine attempts to choose a timestep < TSTEP*RMIN, a
0 |2 U/ d; u* y! M' y4 K Timestep Too Small Error will result.
7 z6 _9 T& K# E3 E - The simulation engine is never allowed to take a larger timestep than4 ` m: h& ?5 q. \6 L
TSTEP*RMAX., f5 l k- i ]! Y7 }( D" Q
Summary$ K! S- d' s/ x9 y& G; d: ^# b T
With the introduction of the RUNLVL algorithm, the impact of TSTEP has been
2 l. Q* ]0 e2 U2 Mgreatly reduced to the point that it can be ignored except for niche applications# H7 Z4 D8 H, B8 r- A) Y# f& \
like INTERP. One should still be careful not to set TSTEP too small for long,$ q4 r9 q; L! p7 ]& D4 a) I3 N
low frequency simulations in which the soft timestep upper limit of RMAX*TSTEP, ^7 {# D. \6 s9 A4 ~4 w; v+ {
may still force more timestep solutions than is necessary which will cause& C7 K+ d8 N# } U
excessive simulation time. |
|