Engineering Mathematics

[Tuurbo46]

Hi

What math package do all you clever Fourier, Z and laplace people use. I
have Matlab 6.5, but im not very good at it. Do you enter sums into command
window, or is there a special toolbox? Also i have Mathmatica. Which is
the most user friendly?

 

[Don Pearce]

Hi

What math package do all you clever Fourier, Z and laplace people use. I
have Matlab 6.5, but im not very good at it. Do you enter sums into command
window, or is there a special toolbox? Also i have Mathmatica. Which is
the most user friendly?

I use Mathcad for almost everything. It hasn't the raw power of Maple
and Mathematica (both of which I also use), but it is incredibly easy
to use, and layout and format are completely freeform. I am currently
a beta tester for version 12 - which is looking good.

d

Pearce Consulting
http://www.pearce.uk.com

 

[Jonathan Kirwan]

I am currently
a beta tester for version 12 - which is looking good.

That is a relief considering the crappy behavior under Win98SE it has had in
earlier versions I have (memory loss is so substantial, I'm rebooting each hour
of use.) I will look forward to the possibility that it has become at least
usable.

Jon

 

[Leon Heller]

Hi

What math package do all you clever Fourier, Z and laplace people use. I
have Matlab 6.5, but im not very good at it. Do you enter sums into command
window, or is there a special toolbox? Also i have Mathmatica. Which is
the most user friendly?

I use Scilab. A big advantage is the cost - nothing.

Leon

 

[John Popelish]

That is a relief considering the crappy behavior under Win98SE it has had in
earlier versions I have (memory loss is so substantial, I'm rebooting each hour
of use.) I will look forward to the possibility that it has become at least
usable.

Jon

Have you been to their web site to download the patches?

 

[Active8]

That is a relief considering the crappy behavior under Win98SE it has had in
earlier versions I have (memory loss is so substantial, I'm rebooting each hour
of use.) I will look forward to the possibility that it has become at least
usable.

Jon

This might come as a surprise to you, but winders 98 can only handle
(manage) 64 MB of ram, IIRC, so if you have more ram in your system,
it's wasted. You may try the suggested patches. Do a defrag on your
drive and manually set your virtual memory (swap space) setting.
Make sure there's enough free drive space to handle the swap space,
though.

 

[Active8]

I use Scilab. A big advantage is the cost - nothing.

Leon

And when the day comes that it will do 3D graphs for pole/zero that
look like the ones I can get with Matlab, it'll *really* be the
balls.

 

[Jonathan Kirwan]

This might come as a surprise to you, but winders 98 can only handle
(manage) 64 MB of ram, IIRC, so if you have more ram in your system,
it's wasted. You may try the suggested patches.

No, it "eats" up the memory. I can sit and watch with a memory tool as the
memory simply "goes away." Close the program? Memory is still unreturned.

I loaded the fixes. No good. And tried it on at least four separate win98
based boxes, one of them with a freshly loaded O/S (plus Microsoft "critical
updates," of course.)

Jon

 

[Active8]

No, it "eats" up the memory. I can sit and watch with a memory tool as the
memory simply "goes away." Close the program?

Ugh. Memory leak. NT and above *should* solve that prob. These
programs dynamically allocate memory on the heap every time you
create an object like a variable or a matrix - a varable is just a
single valued matrix or vector. The pointers have to be freed. Hard
to beleive the programmers left that many leaks.

Memory is still unreturned.

If that were truly the case, nothing else could run.

I loaded the fixes. No good. And tried it on at least four separate win98
based boxes, one of them with a freshly loaded O/S (plus Microsoft "critical
updates," of course.)

Jon

So how much does it eat? Before you say "all of it", like I said, 98
can only manage 65 MB ram so don't tell me it's more than that.

 

[Jonathan Kirwan]

If that were truly the case, nothing else could run.

Yup. That's what happens. That's why I have to reboot. (Or, sometimes, just
turn the sucker off and restart.)

Jon

 

[Jonathan Kirwan]

So how much does it eat? Before you say "all of it", like I said, 98
can only manage 65 MB ram so don't tell me it's more than that.

It's the "system resource" pool and the GDI resource pool that gets sucked down.
If I recall correctly, these are design-limited special areas whose size isn't
anything close to 64Mb. Different thing.

Jon

 

[Jonathan Kirwan]

It's the "system resource" pool and the GDI resource pool that gets sucked down.
If I recall correctly, these are design-limited special areas whose size isn't
anything close to 64Mb. Different thing.

Here's an article from Microsoft with some information:
http://support.microsoft.com/defaul...port/kb/articles/Q190/2/17.ASP&NoWebContent=1

Absolute limits for these areas are:

16-bit User heap 64K
32-bit User window heap 2M
32-bit User menu heap 2M
16-bit GDI heap 64K
32-bit GDI heap 2M

That's true, regardless of how much system memory there is. The last version of
Mathcad (I own several, including the 2001 thing) that didn't have these
problems under '98 was several versions back. (And years ago.)

Unfortunately, because of my work practices, I cannot readily use the newer
operating systems from Microsoft.

Jon

 

[PaulCsouls]

Hi

What math package do all you clever Fourier, Z and laplace people use. I
have Matlab 6.5, but im not very good at it. Do you enter sums into command
window, or is there a special toolbox? Also i have Mathmatica. Which is
the most user friendly?

There is an error filled book by Proakis for doing DSP with MATLAB,
that explains how to transforms. It explains everything well, but
someone should have proofread it.


Paul

 

[John Woodgate]

I read in sci.electronics.design that Jonathan Kirwan

That is a relief considering the crappy behavior under Win98SE it has
had in earlier versions I have (memory loss is so substantial, I'm
rebooting each hour of use.) I will look forward to the possibility
that it has become at least usable.

I think you'll have speed trouble if you want to use it with Win98SE.

 

[Andrew Holme]

What math package do all you clever Fourier, Z and laplace people use. I
have Matlab 6.5, but im not very good at it. Do you enter sums into command
window, or is there a special toolbox?

I use SCILAB which is similar to Matlab but free. Everything must be
entered in the command window. It was hard work at first. It took
several evenings and a post to comp.soft-sys.math.scilab to make it do
what I wanted. It's a good package, the demos are helpful, but the
documentation could be better.

 

[Reg Edwards]

Mathematical programs are of any use only when you already understand
exactly how to do it using pencil and paper. They have no educational value
regarding the subject matter in hand.

The most disastrous, catastrophic mistakes have been made using maths
programs in the dark, without knowing exactly what the computer is doing
with the input data. Only mathematicians should be licensed to use them.

They are purely grey-matter, time and labour-saving devices at which they
make excellent tools. All due only to the fantastic speed of modern
computers. Otherwise they would be no better than log tables and slide
rules.

 

[Jonathan Kirwan]

I think you'll have speed trouble if you want to use it with Win98SE.

The several versions I've tried run at a far more than adequate speeds on at
least several of my machines. Why are you imagining this?

Jon

 

[Kevin Aylward]

Mathematical programs are of any use only when you already understand
exactly how to do it using pencil and paper. They have no
educational value regarding the subject matter in hand.

The most disastrous, catastrophic mistakes have been made using maths
programs in the dark, without knowing exactly what the computer is
doing with the input data. Only mathematicians should be licensed to
use them.

They are purely grey-matter, time and labour-saving devices at which
they make excellent tools. All due only to the fantastic speed of
modern computers. Otherwise they would be no better than log tables
and slide rules.
---

I have to disagree with the overall concept here. I agree, it is true
that one needs to understand in reasonable detail what the maths is.
However, today it is simply not possible to understand *exactly* the
implications of the equations from pen and paper. In general, equations
are too intractable to deal with in the sense that you are alluding to
here. For example, consider nuclear bomb simulations, or black hole
collision simulations. The equations are way too complex to understand
in detail. Its *only* by doing the simulations that one can actually
develop a feel for how the equations operate. Your view is the "nice
idea if in an ideal world", but it isn't. The world simply can't be
analysed using only pen and paper. There isn't even a finite closed form
solution to algebraic equations of degree greater than 4.

Kevin Aylward
[email protected]
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

 

[John Woodgate]

I read in sci.electronics.design that Jonathan Kirwan

The several versions I've tried run at a far more than adequate speeds on at
least several of my machines. Why are you imagining this?

I'm not imagining it. I have Mathcad 11.2 on a Win 98SE machine.
Obviously, speed depends on what you are doing.

 

[John Woodgate]

I read in sci.electronics.design that Kevin Aylward

For example, consider nuclear bomb simulations, or black hole
collision simulations. The equations are way too complex to understand
in detail.

But we don't normally do those calculations in Engineering, at least, I
don't. The clients don't ask for it. (;-)

The legitimate excuse for relying on simulation without much of an
insight into the nuts and bolts of the mathematics is that the problem
is non-linear. An analytical solution of a 'simple' full-wave rectifier
with a small inductor in series on the a.c. side is really TOUGH, but
finite element analysis will do it easily, at the expense of more
computation that you would want to do in a week. A current model
computer will do it in less than a second.