I run Spectre several times but stop at the same space,why?

[Harryzhu]

Hi,

I run simulation with spectre, but it cannot be finished everytime and
stops at the same space, how strange is it! I don't know why it do so, who
may help me or give some advice? Thanks for your help!

Have a good day!

Best Regards,

Harryzhu

 

[S.Badel]

just add this to your .cdsenv

spectre.envOpts dontCrash boolean t

 

[S.Badel]

I'm kidding of course, but...
What kind of answer do you expect if you do not give any clue, error
message or excerpt from the log file???

Please give some more info.

stéphane

 

[G Vandevalk]

Figure out what happens at the bi-stable point.

What is happening is that mathematically something
changes in the network at the time/freq in question.

Usually it can be tracked to some input condition changing.

The other main source is usually caused by a mathematical
quirk in the forcing matrix (i.e. the feedback to some node
reached the same drive as the input)

At this bi-stable condition,the mathematics goes nuts. Usually the
circuit will pass this location in real life because one of the inputs
into the bi-stable condition is driven by a source that is not directly
tied to the output of this node. (But mathematically, the simulator
cannot make this choice!)

Try relaxing your simulation criterion and examining the circuit at this
point. ( I have seen this caused by several "unconnected" noise sources
that were left in the circuit for convenience ... Each time a source has a
discontinuity, the integration algo. changes and causes a glitch ... ARGH)

I recommend that you read the book "The Designer's Guide to Spice & Spectre"
by Kenneth S. Kundert ... ISBN 0-7923-957109



"Harryzhu" <harrytone@sohu.com> wrote in message
news:d2iesf$17ba$1@mail.cn99.com...

Hi,

I run simulation with spectre, but it cannot be finished everytime and
stops at the same space, how strange is it! I don't know why it do so, who
may help me or give some advice? Thanks for your help!

Have a good day!

Best Regards,

Harryzhu

 

[Andrew Beckett]

On Fri, 01 Apr 2005 12:10:47 +0200, "S.Badel"
<stephane.badel@SPAMMERDIENOWepfl.ch> wrote:

just add this to your .cdsenv

spectre.envOpts dontCrash boolean t

Very amusing ;->

 

[fogh]

S.Badel wrote:

I'm kidding of course, but...
What kind of answer do you expect if you do not give any clue, error
message or excerpt from the log file???

Please give some more info.

stéphane
Nice one.

Hope we get more april fool posts next year.

spectre.envOpts pleaseOhPleaseGiveAShyBowBeforeYouCrash boolean t