Open source alternatives for Matlab?

[jan.didden]

I want to do some triode modelling for LTspice, and there are a few apps around that can do that, based on Matlab.
I don't have Matlab, and was wondering about open source alternatives.
I've heard of Octave, anyone has experience with it? Any other alternatives?

Jan

 

[wrankin]

I used Octave years ago in academia and remember that it was a pretty decent alternative to Matlab. The only potential issue is the lack of the various ToolKits available with Matlab.

Good luck,

-bill

 

[BlacK_Chicken]

Go with Octave, it is really decent nowadays.

If you need any help with the LTSpice2Matlab() function shoot me a PM. The orignal one was incompatible with Octave, but I managed to find/modify a version that is working reasonably well. I use it regularily.

 

[mgb1965]

Scilab is also an alternative to Matlab.

Home Page | www.scilab.org

But personally I have never used it.

 

[jan.didden]

Go with Octave, it is really decent nowadays.

If you need any help with the LTSpice2Matlab() function shoot me a PM. The orignal one was incompatible with Octave, but I managed to find/modify a version that is working reasonably well. I use it regularly.

LTspice2Matlab, what is that??

Jan

 

[BlacK_Chicken]

LTSpice2Matlab() is a function that is able to read LTSpice .raw files and show the SPICE results into a array in your workspace.

If you use LTSpice as a commandline tool it is possible to do some, even more, sophisticated analysises than LTSpice features in standalone mode.

Sometimes I just use it for convinient post-processing.

 

[NickKUK]

Octave works, although it's single threaded for the most part which means for big parallel jobs it can be a little slower. Have a look at the octave forge as it pretty much has most of the Matlab libraries.

Currently thinking of writing a octave neural network to solve the best set of resistors for an attenuation network.

 

[jackinnj]

I want to do some triode modelling for LTspice

I just use the optimizer in Excel -- works for tubes.

 

[jan.didden]

I dabbled a bit it that - it is very powerful indeed!
I thingk I have all I need right here in my PC.

Jan

 

[jackinnj]

There was a free-ware program "Nutonian Eurequa" which worked well, but now it is very expensive.

In any event, one neat thing from Eurequa was the "solve for" equation, log squashed error: LN(1+ABS(I_actual - I_predicted))

Optimize to the minima of the sum of errors.