• Remove voltage sources and replace with "Nortonized" (current) sources.
• Add a small GSHUNT, perhaps 1e-12 to as much as 1e-10.
• Add a small CSHUNT, perhaps 1e-15.
• Use SPICE models for similar (IC or transistor) parts from other manufacturers.

开云体育

开云体育

There are indeed infinitely many reasons for that error. We can't help much unless you let us see more of what you are doing. Upload your .ASC file AND all the other files required to run the simulation, but not .RAW? and .LOG files or pictures,? in a ZIP archive to Files => Temp.

Go to the web page: /g/LTspice/topics. Click on Files in the list on the left. Then click on Temp. Then click on New Upload in the blue box at top left. Click on Upload File in the drop-down menu. Then send a message to tell us that you did that.

Hello everybody,
I'm working on a schematic and encounter a "singular matrix" error.
I know there are as many cause than people in the mankind, but I would like to have advices to avoid this sort of things.
I'm aware that there are a lot of maths in the background too :'(
Thanks a lot in advance for your help.

开云体育

View the extended netlist.?

开云体育

The failing stoi function expects an integer during parsing. Maybe 44.1k should be 44100 ?
Or the file contains non-integer values where a integers are 'required' (expected).

Arie

开云体育

Well, that is a very significant error message, isn't it? Thank you for sharing.(;-)

I haven’t tried that, Andy, will do.

Running XVII

?

“A C++ exception with the message:

‘invalid .stoi argument was thrown during simulation, which indicates a bug in LTspice. Please report this incident.”

(Can’t copy the text, can’t attach a screenshot!)

?

Dave

?

From: [email protected] <[email protected]> On Behalf Of Andy I via groups.io
Sent: Monday, March 24, 2025 8:53 AM
To: [email protected]
Subject: EXTERNAL: Re: [LTspice] .STEPping a .WAVE file output

?

It might depend on your LTspice version.? What's yours?

?

Did you try ".wave {stp} 16 ..."?? That is, remove the ".wav" extension from that command.? Then add it manually, in Windows.

?

What error message(s) did you get?

?

Andy

开云体育

开云体育

Exactly what error message do you get?

I want make multiple runs of a model which will output? a waveform to a .wav file.

Each run will step a parameter, all of which works as intended, except, I need to change the filename each time.

This illustrates the intent, but not? surprisingly, errors off on the last line:

?

.step param stp 1 6 1

.param Vcon table({stp}, 1,-2, 2, -1, 3, 0, 4, 1, 5, 2, 6,3)

.wave {Stp}.wav 16 44.1k V(Out)

?

Any suggested better syntax? I’ve tried several…

Thanks!

?

Dave

?

开云体育

开云体育

If I remember correctly, LTspice 24 v24.1.* screwed around with (messed up) AKO.? Since you are using v24.1.5, those errors might be a side-effect of that.

?

You could run an experiment (probably a DC sweep) to see if it picked up the NJ20 model, or if it fell back to the standard NJF model with defaults.

To clear, V24.1.x didn't screw up the AKO feature, as such - it still worked. The actual problem was with the model name not being recognised when AKO was being used with the resulting numeric "name" that was being .STEPed.

开云体育

The simulation runs but I'm not confident its using the proper JFET parameters because it doesn't show them when it produces the error. It may be using JFET defaults.

It works in 24.1.4 and 24.0.12. In 24.1.4, you get an error message in the logfile, but it still works. I don't have 24.1.5 installed yet. The NJF() is superfluous. You already have defined NJ20 as an NFT.

Yes, I know. But I had to that for BJT AKO to work so I tried useing it here.

Do you have braces in the instantiation? Try removing them. There are plenty of nuances here.

Already did. Doesn't help.

I had already checked that it does work even when an error is reported. 24.1.4 gave the same current as 24.0.12 in a .OP analysis, even when it reported the error. I doubled-checked by changing one of NJ20 model parameters (Beta) - the current changed accordingly. The result was the same whether or not the actual original model was used or the AKO name.

The error is being wrongly reported. Errors should be terminal, preventing analysis. If that doesn't happen, it should be reported as a warning, if "strict" syntax rules are not being followed. The classic "questionable curly braces" warning doesn't (normally) seem to affect the analysis.

I don't ever remember having to use the intrinsic model type (D, Q, MN, NJF, etc) when using the AKO syntax. It shouldn't be necessary. If you found that it was for BJT, it suggests that the base model you used wasn't in standard.bjt, and LTspice then fell back to the default BJT when the the AKO was used.

--
Regards,
Tony

I'm using ako: with a trailing (field) separator, as written in "General Form" in MicroSim's PSpice reference manual (as far as I remember, MicroSim introduced ako:).
But in the corresponding example, MicroSim omitted the space.
I hope LTspice will continue to accept ako: with and without trailing separators.

Excerpt from MicroSim PSpice A/D 7 Reference Manual:
Commands:
...
.MODEL (Model)
...
General Form:
.MODEL <model name> [AKO: <reference model name>]
+ ...
...
Example:
...
.MODEL QDR2 AKO:QDRIV NPN (BF=50 IKF=50m)

(from Berkeley SPICE3: Field separators are blanks, comma, equal (’=’) sign, left or right parenthesis; extra separators are ignored)

Bernhard

.model 2N2222mod ako: 2N2222 bf=5?; same except lower beta

.model 2N3904_mod ako: 2N3904 Cje=0