Exp so slow [LTSpice]

[Chris Carlen]

Hi:

This simple thing:

* Y:\spice\LT-SWcad\impulse.asc
V1 N001 0 EXP(0 10 1n 1n 10n 1m)
..tran 10m
..backanno
..end


Simulates at a snail's pace.

Why is that?

I simply want to make an exponentially decaying source, with effectively
a step on the rising edge, which is why the rise tau is so short.

But it sims real slow.

Any suggestions?


Thanks.


Good day!

--
____________________________________
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA
crcarle@sandia.gov

 

[Jim Thompson]

On Thu, 22 Jan 2004 10:22:21 -0800, Chris Carlen
<crcarle@BOGUS.sandia.gov> wrote:

Hi:

This simple thing:

* Y:\spice\LT-SWcad\impulse.asc
V1 N001 0 EXP(0 10 1n 1n 10n 1m)
.tran 10m
.backanno
.end


Simulates at a snail's pace.

Why is that?

I simply want to make an exponentially decaying source, with effectively
a step on the rising edge, which is why the rise tau is so short.

But it sims real slow.

Any suggestions?


Thanks.


Good day!

Those sharp edges probably force small simulation steps over that
whole 10ms.

How large is the data file ?

...Jim Thompson
--
| James E.Thompson, P.E. | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona Voice480)460-2350 | |
| E-mail Address at Website Fax480)460-2142 | Brass Rat |
| http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.

 

[Mike Engelhardt]

Chris,

This simple thing:

* Y:\spice\LT-SWcad\impulse.asc
V1 N001 0 EXP(0 10 1n 1n 10n 1m)
.tran 10m
.backanno
.end

Simulates at a snail's pace.

Why is that?

It's making a mistake about the timestep size it needs to take to
not outrun the time constant specified for rise time constant,
the 4th parameter.

I simply want to make an exponentially decaying source,
with effectively a step on the rising edge, which is
why the rise tau is so short.

Any suggestions?

Use

V1 N001 0 EXP(0 10 1n 1n 10n 1m)

for the time being.

--Mike

 

[Mike Engelhardt]

Chris,

* Y:\spice\LT-SWcad\impulse.asc
V1 N001 0 EXP(0 10 1n 1n 10n 1m)
.tran 10m
.backanno
.end

Simulates at a snail's pace.

Why is that?

It's making a mistake about the timestep size it needs
to take to not outrun the time constant specified for
rise time constant, the 4th parameter.

OK, this problem is fixed on the version currently on
the web. LTspice looks at the details of the sources
to pick timesteps that allow it to run both faster and
more accurately than standard SPICE's. The error in
the algorithm used for exp() sources has been fixed so
LTspice will be able to run your deck more accurately
and faster than anything other SPICE out there.

--Mike

 

[Chris Carlen]

Mike Engelhardt wrote:

Chris,


* Y:\spice\LT-SWcad\impulse.asc
V1 N001 0 EXP(0 10 1n 1n 10n 1m)
.tran 10m
.backanno
.end

Simulates at a snail's pace.

Why is that?

It's making a mistake about the timestep size it needs
to take to not outrun the time constant specified for
rise time constant, the 4th parameter.


OK, this problem is fixed on the version currently on
the web. LTspice looks at the details of the sources
to pick timesteps that allow it to run both faster and
more accurately than standard SPICE's. The error in
the algorithm used for exp() sources has been fixed so
LTspice will be able to run your deck more accurately
and faster than anything other SPICE out there.

--Mike

Thanks Mike! Geez, this new version seems to compute results
*instantly* compared to the one I was using (from about 09/03/03), even
for stuff that wasn't giving me any trouble.

Good day!


--
____________________________________
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA
crcarle@sandia.gov