The Cyber-Spy.Com Usenet Archive Feeds Directly
From The Open And Publicly Available Newsgroup
This Group And Thousands Of Others Are Available
On Most IS NNTP News Servers On Port 119.
Cyber-Spy.Com Is NOT Responsible For Any Topic,
Opinions Or Content Posted To This Or Any Other
Newsgroup. This Web Archive Of The Newsgroup And
Posts Are For Informational Purposes Only.
Reply-To: "Kevin Aylward"
From: "Kevin Aylward"
References: <firstname.lastname@example.org> <email@example.com> <firstname.lastname@example.org> <email@example.com> <12j4LzC+fEp9EwgG@jmwa.demon.co.uk> <firstname.lastname@example.org> <1U9p9.32946$cS4.email@example.com> <firstname.lastname@example.org>
Subject: Re: Problems with Electronics Workbench - timestep errors
X-Newsreader: Microsoft Outlook Express 6.00.2800.1106
Date: Mon, 14 Oct 2002 08:10:58 +0100
NNTP-Posting-Date: Mon, 14 Oct 2002 08:11:03 BST
"carltons" wrote in message
> In article , "Kevin Aylward"
> > You seem to be discussing two issues here. The bit about simulators
> > don't design, and putting up a window for guidance seems to be
> > to real, practical design help, which is independent of any
> > method. The bit about convergence is related to the simulator
> So what? You want the simulator to hold the hand of the person
> (I am not advocating that) or do you expect the person using it to
> what is behind the simulation tool?
An overview, certainly.
>The latter would require another
> several courses of schooling for the user and would not solve the
> at hand.
A basic understanding of what is going on is about half an hour study.
Look, if you connect a scope to a circuit an engineer or technician
*must* understand its basics. He must know hoe to set its controls. He
must know that a scope has capacitance and resistance and will load down
the circuits causing errors. He *must* know that spikes seen on the
trace might not be real spikes, it might be a ground problem e.g. short
the scope probe ground and probe tip together and probe a point on a
To be competent in electronics, you *must* know the limitations of your
tools. You cannot just use them blindly.
>You could go to school forever and not know enough to write a
> Spice program that would converge no matter what. If you want the
> aid the average engineer, then do that much, but when he doesn't know
> about a given portion of the algorithm behind the Spice, don't give
> bunch of Academic grief to the point that he won't use Spice at all.
I don't think that anyone is suggesting that. This was certainly not the
reason why the circuit that was forwarded to me had problems. It was not
really a simulation issue at all, imo. It was equivelent to wanting to
view a 1Khz signal (1ms), but having the time base set to 10us/cm
I think that you are under the illusion that simulation tools can be
made perfect. I agree that some simulators can converge better then
others, however it is *impossible* to have a simulator converge on all
circuits. And I mean impossible. A simulator is a Turing machine, and as
such it is proven that you cannot guarantee a Turing machine will halt.
The reality is, is that at some point, a simulation *will* fail. Life is
simple not as neat as you would like it. You do have to get your hands
dirty. So there is no alternative, at some point, you will have to try
and work around the limitations of the tool. This is the way it is. Its
Certainly, I don't suggest that someone become an expert in algorithms,
however you must understand basic issues involved. For example, *any*
computer has a finite number of bits. Typically spice is only good for
12 digits on typical computers. There is no way around this. If you try
and switch infinitely fast, you are also going to get problems.
> I first started work as an engineer we built a circuit and made it
> We then built lots of them and made them work. We then shipped the
> product. This was silly and time consuming. Now that we have the
> capability to do otherwise, we should. Make the design tool and not
> design instrument. Please give the new guys some breathing room.
So, I am still not really clear on what your are trying to say. "Make
the design tool and not a design instrument." is not really clear to
me. As I said, it not possible to make a fool proof simulator. Make a
tool such that any fool can use it, and it will only be used by fools.
What we have here is essentially, a new widget. How do you operate a new
widget without learning how the new widget operates? If it did operate
the same, it would probably be an old widget.
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.
Go Back To The Cyber-Spy.Com
Usenet Web Archive Index Of
The sci.electronics.design Newsgroup