Re: P-spice Error --> Time step too small..
- From: miso@xxxxxxxxx
- Date: Sun, 30 Mar 2008 20:01:43 -0700 (PDT)
On Mar 30, 6:03 pm, jason KIM <js...@xxxxxxxxxxxxxxxxx> wrote:
hi, all
I'm lcd engineer who design the electric circuit.
recently, i use spice to pridect my circuit and i'm in trouble now.
there is an error -"Time step too small"
to solve the problem, I add some options.
The options are:
-. ITL4=100
-. RELTOL=0.01
- change the time step : 1us --> 50us
but error is still with me.
show me the way, Plz...
regards,
jason
Off the top of my head, check for something floating. Perhaps you are
muxing something and leaving a node undefined. Also a bad model can
cause this problem. Is there a section of the circuit with high gain?
Making the time step bigger won't help. The problem is the circuit
state is changing rapidly so the internal time step is getting smaller
in order to keep the iteration to iteration change within limits.
Sometimes I've got around this problem by using all 4 parameters on
the .tran line. The spice gurus can do a better explanation, but
basically if you just give spice a time step and time limit, the
simulation picks times that it thinks are appropriate given the
changing state of the circuit. But for display (print out) purposes,
the value at the time step you desire is interpolated from the time
steps that spice used. You see this all the time of you are trying to
simulate a circuit for THD. That is, for such simulations you need to
set the time step to be say 1/100th of the period of the input sine
wave. The .tran card goes step stop start max. So you may need to set
a small MAX limit.
.
- References:
- [Help!!] P-spice Error --> Time step too small..
- From: jason KIM
- [Help!!] P-spice Error --> Time step too small..
- Prev by Date: Re: John Larkin the Posturing CUNTHEAD
- Next by Date: Re: State- of-the-art robotics?
- Previous by thread: [Help!!] P-spice Error --> Time step too small..
- Next by thread: pregnant woman
- Index(es):