I know about time step control and its deviousness...! I think minimum
time step size puts an u necessary computational burden when little or
no activity is present.. it is better to have variable time step
resolution which is user guided...
Thanks
cheers
AG
toggle quoted message
Show quoted text
On 9/14/2011 5:02 PM, Tony Casey wrote:
--- In LTspice@... <mailto:LTspice%40yahoogroups.com>,
Ganesan <dg1@...> wrote:
Thanks Nice idea...
This will work for transient analysis too, if LIST is supported in
.tran....
Is list supported in .TRAN ?? This would allow you to have lot of data
points during fast edges and few during flat portions of the clock.
Cheers
AG
================================================================================================================
On 9/14/2011 2:48 PM, Tony Casey wrote:
<snip>
--- In LTspice@... <mailto:LTspice%40yahoogroups.com>
<mailto:LTspice%40yahoogroups.com>,
Ganesan <dg1@> wrote:
I can run different ac simulations with different frequency
resolutions
and range..
Is there a way to merge the plot files.?
Cheers
A. Ganesan
</snip>
Hello Ganesan,
I just thought of another, more direct way for you to do exactly what
you want without messing with raw files.
You need to use the .ac list... syntax; but why not put your
(unequally-spaced or indeed random) frequencies in an external file?
1. Calculate, or determine your list of frequencies in Excel and
export as a space delimited text file. Open the file in a text editor
and prepend ".ac list ". (Don't forget the space)
2. Save file as Frequency_list.txt
3. Add a SPICE directive .inc Frequency_list.txt
4. Add .op analysis directive (this will be ignored when the text
file
is read).
Job done. I have uploaded a working example to Files>Temp.
I even surprised myself with this one. :-)
I guess this would work with a .dc list analysis too, but I don't see
any point in this.
Regards,
Tony
No virus found in this incoming message.
Checked by AVG - www.avg.com
Version: 9.0.914 / Virus Database: 271.1.1/3896 - Release Date:
09/14/11 01:34:00
Hello Ganesan,
This kind of interference is not needed or even desirable in .tran
analyses because SPICE itself determines the step size depending on
circuit activity.
Sometimes it needs some help, though, in deciding what's really
important. For this we can set the minimum step size; or the cunning
designer might put tripdt/tripdv limits into a bogus B source to force
the step size reduction just in the areas in most interest. Another
means of doing this is to have a high(er) frequency source that drives
the variable step size algorithm. This could be gated to run only when
it's necessary to increase the resolution.
Regards,
Tony
