How to get a fixed timesetp in hspice?

[ysz]

hello all!
I need a fixed or approximate fixed timestep for .tran sim, who can tell me how to do?

 

[frankon_o2]

for example:
.tran 1u 20m
then 1uSec is the time step of transient simulation
the length of simulation is 20mSec
Is that what u want to know?

 

[nathan]

use delmax (maxdel ???)= ...

nathan

 

[ysz]

for example:
.tran 1u 20m
then 1uSec is the time step of transient simulation
the length of simulation is 20mSec
Is that what u want to know?

hspice will change the timestep, not always 1u

 

[ysz]

use delmax (maxdel ???)= ...

nathan

I use this opt:
.option post=2 accurate DELMAX=0.0u RMAX=1 DVDT=2 LVLTIM=2 FS=0.1 FT=0.1
it has a little benifit .

 

[always@smart]

hi ysz,

how do u know that Hspice will change the timestep? sorry i;m new in using hspice.

can you please also explained the command u have shown us?

.option post=2 accurate DELMAX=0.0u RMAX=1 DVDT=2 LVLTIM=2 FS=0.1 FT=0.1

thanks in advance

regards,

 

[ysz]

Star-Hspice allows the selection of three dynamic timestep control algorithms:
n Iteration count
n Truncation
n DVDT
Each of these algorithms uses a dynamically changing timestep.

you can look it up in the manual.

 

[ysz]

this is my ugly wave, who can help me?

 

[leo_o2]

Set DELMAX to a value smaller than the minimum pulse width. It might help. Avoid signals with steep edge.

 

[ysz]

Set DELMAX to a value smaller than the minimum pulse width. It might help. Avoid signals with steep edge.

I already set DelMax=0u

 

[leo_o2]

What is TStep that you used for .tran analysis?

 

[ysz]

.option post=2 accurate DELMAX=0.0u RMAX=1 DVDT=2 LVLTIM=2 FS=0.1 FT=0.1
.tran 10u 120m

 

[leo_o2]

Obviously, 10us for Tstep is too big. You must decrease it first in my opinion.

 

[ysz]

I have changed it to 1us, the wave is same, I'm so doubt.

 

[leo_o2]

Try 10ns. However, it will be very slow.

 

[joskin]

The main problem is your simultion time is too long(bigger than 100ms).
Because in (*.tr#)transient sim result file,the valid digital number that describe simulation time is fixed,the wave can't show detailed information when the time exceeds 100ms.

 

[ysz]

The main problem is your simultion time is too long(bigger than 100ms).
Because in (*.tr#)transient sim result file,the valid digital number that describe simulation time is fixed,the wave can't show detailed information when the time exceeds 100ms.

really?
but if I can set a fixed timestep, I will set the timestep longer, so the number of steps are reduced, maybe I can have longer sim time. do you think so?

 

[ysz]

Nobody can help me?

 

[Hughes]

.option post=2 accurate DELMAX=0.0u RMAX=1 DVDT=2 LVLTIM=2 FS=0.1 FT=0.1
.tran 10u 120m

Don't set DELMAX=0, set to a small value other than 0. For example:
.option delmax=0.1n

 

[fashion]

If you set more accurate, the speed of simulation will be very long, sometime set 1/10 of total time.