VisSim vs. Matlab/Simulink

May 19, 2004 6 Replies

Greetings:



I'll soon be experimenting with using TI DSPs for control systems. I'd also like to have some good modeling and simulation tools. Right now I use SPICE and Mathematica. SPICE is good for continuous time stuff, but not so good for discrete. MMA is good for error-free algebraic simplifications of paper and pencil analysis, as well as plotting and simulations, but it has a peculiar programming language. For that reason, I haven't yet attempted to do any discretely sampled system modeling with MMA. It's also not well integrated with the outside world, compared to MatLab.



Thus, I am considering becoming acquainted with either Matlab or VisSim. Both are quite expensive, so it will not be possible to purchase both. I will likely try evaluations of them as much as possible before making a decision. I know Matlab is well entrenched in the world of DSP and control simulation. But I didn't hear about VisSim until recently.



Is VisSim a good, stable product?



Why might one choose Matlab over VisSim, or vice/versa?



Has anyone used the target code development tools from either vendor? It seems both offer tools to generate code to target TI C2000 DSPs and C67x DSPs.



Not that I'd want to depend entirely on such code development tools, of course. I rather enjoy hacking out numerical algorithms in C and ASM. But it seems that such hacking is actually unwise at times, when a proper tool will get a project completed to meet specs in a much shorter time than by starting from scratch. This will often be the case in my projects, since we can always afford to throw excessive CPU horsepower at a problem, rather than a lengthy development time. This might relegate my hand coding work to experimentation only, but that is Ok.



Thanks for input.


Good day!


Short form:

I don't know about VisSim, but Simulink should answer the mail for you and you may want to look at SciLab.

Long form:

I remember playing with VisSim 10 years ago, but I never had the opportunity to go beyond a demo. At the time it was as solid as Simulink, so if they've tracked it should be fine.

Simulink is very good for mixed-mode simulation (and costs bucks). Three years ago their code generation wasn't much to write home about, and I haven't had occasion to use it since then. I'm personally rather paranoid about using their code generation, but it has more to do with the efficiency of the resulting code than it does with any possibility that it won't perform exactly as expected. My preferred technique has been to write my code and simulate it in Simulink, which works quite well.

I've heard that Scilab

formatting link
is a gnu-ish version of Matlab, complete with simulation, but I haven't tried it yet. I suspect that it wouldn't have the code generation capabilities of Simulink, however.

Check into Scicos. It's part of open-sourced Scilab, a free MatLab-like alternative.

Scicos URL:

formatting link
I copied this from Scilab's site: "The newsgroup comp.soft-sys.math.scilab is a good place to exchange experience, ideas and Scilab programs."

That newsgroup currently has a thread asking about using MatLab programs with Scilab. Check it out...

Gary Schnabl

Hi Chris,

I don't use any of those modeling tools, but I just thought I'd mention that LabView (National Instruments) is gaining in popularity and may well be worth adding to your list of candidate signal processing software packages.

See Ya', [-Rick-]

Hi Rick. Thanks for the reply.

I don't much like LabView. I like programming in a text language like C. Actually, I don't know if I'd like the more high-level stuff Matlab has to offer. But I might certainly benefit from Matlab as a modeling tool, and perhaps Simulink, but even that might be overkill.

We'll see what happens once I finally get Z-transformed from my continuous time thinking and understanding of things into some actual discrete stuff. Maybe I can just figure out how to do in in Mathematica.

Another reason for the interest in Matlab is that it seems to be the most commonly used package for signal processing and controls modeling. Most of the controls texts I encounter talk about it. It might simply be a good lingua franca.

Have a good day!

Hi Rick,

May I ask, what do you use for modeling tools? Or don't you do any modeling at all?

Not only for the software side of the things... It's real power comes from the fact that you can interface directly to hardware, to play with signals form the real world, in a graphical and easy-to-program environment.

Kindest regards,

JaaC

Hi,

I have an old copy of the Student Version of MATLAB.

[-Rick-]

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required