Hi,

I'm using the LMX24_LM2902 spice model, taken from the Texas Instruments website, and it is giving me a lot of issues when simulating my circuits. In particular, it gets stuck when running a transient analysis, sometimes even stopping it altogether. Sometimes it would not even allow me to run the transient. 
I've tried the same circuit with different Op Amps and it runs smoothly. I've also tried other models for the LM124 I've found in here and they work very nicely. But I'm a bit surprised that my issue comes from a model produced by the manufacturer. Has anyone experienced the same issue before?

Thanks!
Marcos

Are you operating it with a single supply or a dual supply? Some TI models do not work in one or the other mode.

Hi,

I'm using the LMX24_LM2902 spice model, taken from the Texas Instruments website, and it is giving me a lot of issues when simulating my circuits. In particular, it gets stuck when running a transient analysis, sometimes even stopping it altogether. Sometimes it would not even allow me to run the transient. 
I've tried the same circuit with different Op Amps and it runs smoothly. I've also tried other models for the LM124 I've found in here and they work very nicely. But I'm a bit surprised that my issue comes from a model produced by the manufacturer. Has anyone experienced the same issue before?

Thanks!
Marcos

Hi John,

Thanks for your input. I've tried both configurations and both give me issues.

Thanks,
Marcos

On 15/10/2020 09:35, m.compadre.t@... wrote:

I'm using the LMX24_LM2902 spice model, taken from the Texas Instruments website, and it is giving me a lot of issues when simulating my circuits. In particular, it gets stuck when running a transient analysis, sometimes even stopping it altogether. Sometimes it would not even allow me to run the transient.
I've tried the same circuit with different Op Amps and it runs smoothly. I've also tried other models for the LM124 I've found in here and they work very nicely. But I'm a bit surprised that my issue comes from a model produced by the manufacturer. Has anyone experienced the same issue before?

Unfortunately, it's very common for there to be problems with manufacturers' models, especially those that were written a long time ago. Sometimes, these issues can be mitigated by changing some of LTspice's settings, sometimes not.

Many models are poorly written. It doesn't matter that they come from manufacturers' websites; often the SPICE modelling is outsourced by semiconductor companies. This was particularly true many years ago. These days, more manufacturers realise that providing good models is important to designers now that almost everyone with a brain simulates their designs before building them.

--
Regards,
Tony

OK, go to the list web site and download all_files_z_yahoo.htm.

Open it in your browser and search (CTRL-F) for LM324. You will find messages about problems and models that work.

======================================================================================

Hi John,

Thanks for your input. I've tried both configurations and both give me issues.

Thanks,
Marcos

John wrote, "OK, go to the list web site and download all_files_z_yahoo.htm."

That file is here:
https://groups.io/g/LTspice/files/z_yahoo/all_files_z_yahoo.htm

    "Open it in your browser and search (CTRL-F) for LM324."

There are MANY examples of it.  The LM324 was a popular op-amp.  Probably best to start with anything under the "Lib" directory in our archives:

https://groups.io/g/LTspice/files/z_yahoo/Lib/LM324/
or
https://groups.io/g/LTspice/files/z_yahoo/Lib//LM324_NS_test.zip
or
https://groups.io/g/LTspice/files/z_yahoo/Lib//LM324_test.zip
or
https://groups.io/g/LTspice/files/z_yahoo/Lib//LM324_TI_test.zip  (description says it "has some convergence problems"; it is a frequent issue with models from TI)

For the LM2902, again search within that "all_files" file and see what comes up.  This one looks promising because it's in "Lib" and was uploaded by Helmut:

https://groups.io/g/LTspice/files/z_yahoo/Lib//LM2902_test.zip

In many cases, the files that are in our group's archives have been "repaired" to fix known problems with manufacturer's SPICE models such as small incompatibilities between PSpice and LTspice, or to add aides that help make them converge better.  Many manufacturer's SPICE models are poorly written and have discontinuities in a function or its derivative, both of which lead to convergence problems.

Andy

Marcos wrote, "But I'm a bit surprised that my issue comes from a model produced by the manufacturer. Has anyone experienced the same issue before?"

Yes!  All the time.

Some manufacturers don't know how to write good SPICE models for their own parts.

Some manufacturers make their models so obfuscated (difficult to read and decipher) that they compromise how it works.

Some manufacturers use complicated functions and nonlinear elements in their models (often which sharp inflections or "corners"), that they lead to convergence problems.  Many models contain few if any actual semiconductors in the model, instead using mathematical elements.  Many of those mathematical functions have if(...) functions which introduce inflections, which are bad.

Some manufacturers are so tied in to the capabilities of one SPICE program, that they ignore the fact that their models do not "play" well with other SPICE programs.  Some exploit the unique features of one particular SPICE program, causing their models to not work (or to work poorly) in other SPICE programs.

Andy

Thanks everyone for the great inputs. It was very surprising, but glad to see that it wasn't me doing something wrong. Thanks again for your experienced inputs!

Marcos

Can anyone recommend good material, books, articles on how to start writing your own SPICE models for someone that has NEVER written one?

Many thanks,
Marcos

So, where did you look in your quest for a LM124 model?

Obviously not on the manufacturer's website.

If you had searched this group's Files archives, you would have found a number of files of interest.