Smith Chart Programs

[rob...@my-deja.com]

I am looking for a good free smith chart program for windows. I've
looked around the search engines and have only found old dos programs.
Does anyone have any suggestions?

Thanks,
Robert

Sent via Deja.com
http://www.deja.com/

[tom_b...@agilent.com]

Well, two suggestions, neither of which is exactly what you asked for:

First, I like WinSmith quite a lot. There are things that could be
better about it, but it lets you build a ladder network of R, L, C,
lines and transformers, and adjust them easily to quickly visualize
transformations on a Smith chart plot, simultaneously at a set of
frequencies. But...it costs around $80 as I recall.

Second, there are do-it-yourself solutions. You could program Excel to
plot reflection coefficient, and even arrange it to handle some simple
ladder network things. Plotting the typical Smith chart lines might be
a challenge! You could make quite a decent Smith chart program in
SciLab, which is freeware. There are tools to do a user interface
something like WinSmith. But it would be a fairly serious programming
effort. It would be a very nice contribuiton to the ham/RF community if
someone undertook this, though, and made it available.

I know there are a few freeware Smith chart programs out there, and I'm
sure someone will post pointers to them, but none I've seen makes me
even remotely interested in switching from WinSmith.

Cheers,
Tom

In article <959j1s$ms1$1...@nnrp1.deja.com>,

--
Tom Bruhns -- K7ITM

[Al Gerbens]

There's some software and links at this site.
http://rf.rfglobalnet.com/software_modeling/categories/25.htm

rob...@my-deja.com wrote in message <959j1s$ms1$1...@nnrp1.deja.com>...

[rob...@my-deja.com]

Thanks for the input. I looked at WinSmith. I would buy it, but I
can't find anywhere that will allow immediate download after purchase
since I need it asap. I've spent two days now writing my own in Python
and wxWindows to make it cross-platform. If anyone is interested then
I'll post a link when I get to a usable state.

I would still like to get a like to a good alternative in the mean
time.

Robert

In article <959na3$r2u$1...@nnrp1.deja.com>,

[rob...@my-deja.com]

Thanks for the link. I had been trying to download a few last week but
RFGlobalNet was having problems so I couldn't get any. I just found a
program I had quite a while ago and hadn't been able to find since.
The program is smith_v191.zip at
http://rf.rfglobalnet.com/software_modeling/software/25/471.htm if
anyone else is looking for a similar program.

Thanks again,
Robert

In article <3p_d6.1587$gC1.3...@news.uswest.net>,

[David Newkirk]

rob...@my-deja.com wrote:

> I am looking for a good free smith chart program for windows. I've
> looked around the search engines and have only found old dos programs.
> Does anyone have any suggestions?

Although it's not a Smith Chart program per se, Serenade SV, the
free-for-the-downloading student version of Ansoft Corporation's Serenade
Design Suite (the full version of which includes linear and nonlinear
circuit simulation, system simulation, layout and filter synthesis
tools), includes Smith Tool, an interactive, Smith-Chart-based
matching-network synthesis program that can be invoked from any polar
graph produced by Serenade's circuit simulator. You can download Serenade
SV's two (large) component files (SerenadeH.exe, which installs the
circuit simulator that includes Smith Tool, and SerenadeS.exe, which
installs Serenade SV;s system simulator) via the pages that begin at

http://www.ansoft.com/about/academics/sersv/index.cfm

*And* you can read more about Serenade SV in the January 2001 issue of
ARRL's QST magazine.

Best regards,

Dave Newkirk, W9VES
(a technical writer with Ansoft)
dnew...@ansoft.com
dpne...@home.com

[Reg Edwards]

It depends on what you want to use a Smith Chart for.

If the purpose is design or analysis or performance of transmission lines
with given terminations, or computation of what the terminating impedance
must be for a given input impedance, then small professional-grade programs
are available from the website below.

They are much more accurate, faster and easier to use than charts. They
cover the whole electrical spectrum whereas Smith charts are useful at only
at HF. Also transform from series R+jX to parallel equivalents, and
reverse. Smith charts are obsolete in this new Century.

Download in a few seconds each, programs RJELINE2, RJELINE3, COAXPAIR.

Not zipped up. Run immediately.
---
***********************************
Regards, Reg, G4FGQ
Free radio design & modelling software
http://www.btinternet.com/~g4fgq.regp

***********************************

<rob...@my-deja.com> wrote in message news:959j1s$ms1$1...@nnrp1.deja.com...

[Al Gerbens]

These programs are great, thanks.

Reg Edwards wrote in message <95bq4b$n2d$1...@neptunium.btinternet.com>...

[Craig A. Ferris]

Why do you say Smith charts are only useful at HF??? Please explain.

73,
Craig NR4E

[tom_b...@agilent.com]

In article <3A79C0FD...@aeronix.com>,

"Craig A. Ferris" <cfe...@aeronix.com> wrote:
> Why do you say Smith charts are only useful at HF??? Please explain.

Actually, I believe he went on to write that they are "obsolete". Mind
you, there are those of us who don't share that opinion. In fact, my
Smith chart seems to work just fine at audio frequencies...and hasn't
failed to work yet when I've called on it in 2001. To me, it remains a
valuable visualization tool which helps me anticipate what will happen
when I add or subtract elements in a ladder network, or change their
values, and with a computer-aided chart, there's absolutely no sacrifice
of accuracy. (Note that it's easy to read a paper Smith chart to the
accuracy needed for a great many engineering applications, and it's
easily powered by a brain and a pencil in the field.)

As with any tool, if you don't find it useful, don't use it. But don't
assume it's not useful to others. A bit like "proving" that a
bumble-bee can't fly??

--
Tom Bruhns -- K7ITM

[Ian White, G3SEK]

Reg Edwards wrote:
>If the purpose is design or analysis or performance of transmission
>lines with given terminations, or computation of what the terminating
>impedance must be for a given input impedance, then small professional-
>grade programs are available from the website below.
>
>They are much more accurate, faster and easier to use than charts. They
>cover the whole electrical spectrum whereas Smith charts are useful at
>only at HF. Also transform from series R+jX to parallel equivalents,
>and reverse. Smith charts are obsolete in this new Century.

I agree with Reg, if numbers are what you want. There's nothing more
frustrating than playing the geometry-construction game on a paper
chart, and watching the drawing errors mount up.

On the other hand, Smith charts are wonderful as a thinking tool. Don't
think of it as a calculator - that really is obsolete - but simply learn
to read it like a map.

There are about six things you need to know in order to read a Smith
chart like a map, but words are the wrong way to teach something that is
essentially visual. Find out more if you can - you won't regret it!

The ideal software is something that gives you the accurate numbers
*and* the intuitive graphical features of the Smith chart too. One
Windows program with those features is Z-MATCH marketed by Number One
Systems, UK: http://www.numberone.com/ds_zmatch.htm
The only problem is the "professional" price.

73 from Ian G3SEK Editor, 'The VHF/UHF DX Book'
'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.com/g3sek

[William E. Sabin]

Ian White, G3SEK wrote:

> On the other hand, Smith charts are wonderful as a thinking tool. Don't
> think of it as a calculator - that really is obsolete - but simply learn
> to read it like a map.

The Smith charts in programs such as ARRL Radio Designer and Harmonica are
very useful for designing amplifiers, oscillators and matching networks
using L, C and xmission lines in a Z0 system environment. The ability to
get gain, stability and noise figure circles and the input/output reflection
coefficients that provide the desired performance over a wide frequency
range is hard to duplicate using nothing but equations. The visual aspect
adds insight that is almost impossible to get otherwise.

There is a two part QEX article for the Radio Designer program that is
available from the ARRL QEX website.

Bill W0IYH

[rob...@my-deja.com]

I've actually been trying to download these for a couple of days. I
got serenades.exe, but I couldn't get serenadeh.exe. I kept getting a
124k file then the download would end even though the page showed that
it was a 31meg file. I just tried again today and the page now shows
the size as 124k, but when I run it I'm told that it isn't a valid
program. Any thoughts?

Robert

In article <3A78B1BE...@home.com>,

[Reg Edwards]

Francine wrote -

> > Why do you say Smith charts are only useful at HF??? Please explain.
>

> Yes,I wonder why that is too.
============================
Background -

Smith did not "invent" Smith charts so much as simplify much older charts
which had been in use even before the end of the 19th century to solve power
and telephone transmission line problems. And indeed a wide variety of
problems in other engineering fields.

In mathematical terms, the old charts provided a graphical means of
performing extremely complicated and tedious calculations involving complex
hyperbolic functions. Amonst other uses they solved equations such as
Tanh(A+jB) = X+jY which occur very frequently in the analysis of performance
of all transmission lines.

Before the age of computers they were absolutely essential labour-saving
tools. They were faster than slide rules, trig-tables, log tables, etc, but
not so accurate of course. They consisted of a whole stack of charts and
were still very cumbersome to use.

When in the 1930's many transmission engineers were shifting attention to
carrier telephony, HF world-wide and VHF radio, etc, it was realised the old
graphical aids could be simplified and greatly reduced in number by assuming
the characteristic impedance Zo of HF lines to be a pure resistance Ro, the
reactive component jXo of real lines in many cases being small enough to be
ignored. This, for example, reduced hyperbolic functions to Sin(A) = X and
InvSinh(A+jB) to Log(x), with corresponding simplification of charts.

Smith, around 1940 I think, then an employee of Bell, was the first person
to have the newly simplified charts printed and his name attached to them.
They were quickly snatched up and put to use in the design of ground and
airborne radar systems. After the war came TV.

One consequence of setting the angle of Zo to zero is to imply that
conductor loss per unit length is identical to dielectric loss per unit
length - which it never is ! On the other hand this does not begin to
matter until a line is long enough for the actual loss to be noticeable.
Radio amateurs are concerned only with short lengths of low-loss lines.
There are other consequences.

The exact formula for Zo is -

Ro+jXo = Sqrt((R+jwL)/(G+jwC)) ohms.

So Smith charts can be misleading and inaccurate unless w = 2*Pi*F is large
enough to make

Zo = Sqrt(L/C). ie., a lossless transmission line.

But the angle of Zo of real lines approaches -45 degrees at power, audio
frequencies and VLF

Similar discrepancies occur when considering velocity. Smith's lines always
have a velocity factor of unity but, of course, it is easy to apply a
correction if it is remembered a correction may be needed. But even the
correction is guesswork.

A Smith Chart is incapable of showing how an SWR as high as 2.414 can arise.
According to Smith an SWR greater than unity is impossible. Beware of using
his charts as educational tools.

The sentiment attached to Smith Charts is understandable. But in the age of
the PC the young busy engineer has tools of exact precision, labour-saving
in the extreme, and covering all frequencies from 1Hz to 1GHz. And just as
educational.

I have a book of 6-figure mathematical tables. Also a beautiful 12" slide
rule in a pigskin leather carrying case, with a strap for attachment to the
owners belt, in excellent working order. Ocasionally I take it out of its
drawer and handle it for a while with fond memories.

But I never use it.

And in 57 years of association with transmission lines, from 0.1 Hz to 3
GHz, I have never used a Smith Chart either.
----

[William E. Sabin]

Reg Edwards wrote:

> A Smith Chart is incapable of showing how an SWR as high as 2.414 can arise.
> According to Smith an SWR greater than unity is impossible.

Baloney.

> Beware of using
> his charts as educational tools.

More baloney.

Bill W0IYH

[Gerhard W. Hoffmann]

rob...@my-deja.com wrote:
>
> I've actually been trying to download these for a couple of days. I
> got serenades.exe, but I couldn't get serenadeh.exe. I kept getting a
> 124k file then the download would end even though the page showed that
> it was a 31meg file. I just tried again today and the page now shows
> the size as 124k, but when I run it I'm told that it isn't a valid
> program. Any thoughts?

I had the same problem here.

73, Gerhard dk4xp

[Craig A. Ferris]

Thank you for your enlightenment. And to think the amount of money that was
wasted on my education. I was under the impression that I could calculate
attenuation, swr, stability and gain, and most importantly, match impedances
using this antiquated chart. What was I thinking?

George P. Burdell is rolling over in his grave.

73,
Craig NR4E

[tom_b...@agilent.com]

In article <95ehjf$5j$1...@uranium.btinternet.com>,
"Reg Edwards" <g4fgq...@btinternet.com> wrote:
...

> A Smith Chart is incapable of showing how an SWR as high as 2.414 can
arise.
> According to Smith an SWR greater than unity is impossible.

Perhaps you were thinking of reflection coefficient magnitude? (There
are those of us who might disagree even if that were the case...)

--
Tom Bruhns -- K7ITM

[John Moriarity]

A masterpiece of obfuscation.

John, K6QQ, a Smith Chart user.

> Background -
>
> Smith did not "invent" Smith charts so much as simplify much older
charts
> which had been in use even before the end of the 19th century to solve
power
> and telephone transmission line problems. And indeed a wide variety of
> problems in other engineering fields.
>
> In mathematical terms, the old charts provided a graphical means of
> performing extremely complicated and tedious calculations involving
complex
> hyperbolic functions. Amonst other uses they solved equations such as
> Tanh(A+jB) = X+jY which occur very frequently in the analysis of
performance

> of all transmission lines...

> And in 57 years of association with transmission lines, from 0.1 Hz to
3
> GHz, I have never used a Smith Chart either.

Much to your loss.

[Al, N2NKB]

I learned a new word today...
ob·fus·cate tr.v. ob·fus·cat·ed, ob·fus·cat·ing, ob·fus·cates. 1. To make so confused or opaque as to be difficult
to perceive or understand. 2. To render indistinct or dim; darken. --ob”fus·ca“tion n.
Al

John Moriarity wrote in message ...

[Reg Edwards]

Tom wrote:
> "Reg Edwards"wrote:

> ...
> > A Smith Chart is incapable of showing how an SWR as high as 2.414 can
> arise.
> > According to Smith an SWR greater than unity is impossible.
>
> Perhaps you were thinking of reflection coefficient magnitude? (There
> are those of us who might disagree even if that were the case...)

=============================

Thanks Tom for correcting my slip of the tongue (pen, keyboard).

Yes, it is the Reflection Coefficient which the Smith Chart asserts cannot
rise above unity whereas on real lines the RC can approach 1 + sqrt(2).

Would you be one of those chart worshippers who disagrees ?
;o) ;o) ;o)
----
Reg, G4FGQ

[Reg Edwards]

Francine, I've nothing against Smith's transmission line charts except their
limitations. They've proved their worth over the last 60 years. But, as with
any other tool, users should understand or at least be very aware of the
limitations.

Here's another one: It can arise that least loss on a line does not occur
when the line is in a matched condition. But any analysis method which makes
the same assumptions as Smith's simplifications will give no indication of
this.

Your habit of calculating first and then checking by an alternative,
independent, graphical method is what I have done myself on countless
occasions. If both answers fall in the same ball park then all's right with
the World. If not - then panic stations.

Incidentally when RC = 1.5, SWR = -5 (minus 5). ;o)

Regards, Reg, G4FGQ
http://www.btinternet.com/~g4fgq.reg

francine <fran...@multiweb.nl> wrote in message
news:1eo92zq.1wmmwopvx76tcN@[212.58.191.56]...
> Hi Reg,
>
> Thanks for your reply.

>
> > Francine wrote -
> > > > Why do you say Smith charts are only useful at HF??? Please explain.
> > >
> > > Yes,I wonder why that is too.
> > ============================
> >

> > A Smith Chart is incapable of showing how an SWR as high as 2.414 can
arise.
> > According to Smith an SWR greater than unity is impossible. Beware of
using
> > his charts as educational tools.
>

> As Tom said, you probably meant reflection coëfficient magnitude.
> When reflection coëfficient mag = 1, vswr is infinite.

>
> > The sentiment attached to Smith Charts is understandable. But in the age
of
> > the PC the young busy engineer has tools of exact precision,
labour-saving
> > in the extreme, and covering all frequencies from 1Hz to 1GHz. And just
as
> > educational.
>

> I agree with you that you do not have to use the chart to design these
> days. I must admit I calculate and plot the results on a smith chart,
> accept to determine the length of a line. I use the chart to do that for
> me, but that's because I like to do something with the program to give
> me a feeling that I have a visual interaction of things. i.e. play a bit
> with the controls and see what happens with the results. It gives one a
> good idea in "What happens if I do this".
> But, hé...that's personal.
>
> --
> francine

[Donaly]

Hi Reg,
I'm not a chart worshipper, but I do appreciate the fact
that when I draw a curve on a Smith chart, the chart shows values
at an infinite number of points between the ends of the curve.
No digital computer can do that. Of course, computers can draw
curves, too, which are good enough, but if you measure the time it
takes to turn on a computer, load the software, enter the parameters,
and click the mouse, you'll find you haven't saved much time over
just reaching for a piece of paper and drawing a curve on it with
a compass.

Tom Donaly KA6RUH

[budgie]

On Sat, 03 Feb 2001 08:42:26 -0800, Donaly <dtdo...@pacbell.net>
wrote:

>Hi Reg,
> I'm not a chart worshipper, but I do appreciate the fact
>that when I draw a curve on a Smith chart, the chart shows values
>at an infinite number of points between the ends of the curve.
>No digital computer can do that. Of course, computers can draw
>curves, too, which are good enough, but if you measure the time it
>takes to turn on a computer

(snip)

You mean to say that you turn your computer OFF?

[Reg Edwards]

Ian said -

> The ideal software is something that gives you the accurate numbers
> *and* the intuitive graphical features of the Smith chart too. One
> Windows program with those features is Z-MATCH marketed by Number One
> Systems, UK: http://www.numberone.com/ds_zmatch.htm
> The only problem is the "professional" price.

=====================================

Hi Ian,

It's hardly the ideal program when you get a precision version of the Smith
chart plus 8-digit (?) numbers, complete with all the errors in the first
digit which Smith built into it. At low enough frequencies the errors are
orders of magnitude.

They who use such programs are liable to suffer from delusions of accuracy.

Do they come with a Government health warning on the package ?
----
Reg, G4FGQ

[Ian White, G3SEK]

Reg Edwards wrote:
>Ian said -
>> The ideal software is something that gives you the accurate numbers
>> *and* the intuitive graphical features of the Smith chart too. One
>> Windows program with those features is Z-MATCH marketed by Number One
>> Systems, UK: http://www.numberone.com/ds_zmatch.htm
>> The only problem is the "professional" price.
>=====================================
>
>Hi Ian,
>
>It's hardly the ideal program when you get a precision version of the Smith
>chart plus 8-digit (?) numbers, complete with all the errors in the first
>digit which Smith built into it. At low enough frequencies the errors are
>orders of magnitude.
>
>They who use such programs are liable to suffer from delusions of accuracy.
>

As far as I'm aware, the only limitation that particular program shares
with the paper chart is the assumption that the characteristic impedance
of the line is purely resistive. Otherwise I believe it uses the full
transmission line equations (including line loss) to compute its
results. The chart format is merely for visualisation and optional
input.

You may be almost unique in having spent your career in areas where the
Smith chart *doesn't* work - VLF, very long lines and so on. Thus you've
learned all about its shortcomings, but you say you have never actually
used it to see what it *can* do. For the rest of us, it works just fine
(or close enough) as a visual aid to *understanding* transmission lines
and matching problems.

I agree with you that the days of the Smith chart as a calculation tool
are over. I've tried lots of impedance calculators, and even wrote one
of my own (http://www.ifwtech.com/g3sek/netcalc/netcalc.htm), but all
they give me is numbers. They don't give me the *understanding* that a
Smith chart can.

[Reg Edwards]

Ian White, G3SEK wrote

> As far as I'm aware, the only limitation that particular program shares
> with the paper chart is the assumption that the characteristic impedance
> of the line is purely resistive. Otherwise I believe it uses the full
> transmission line equations (including line loss) to compute its
> results. The chart format is merely for visualisation and optional
> input.

==================================
Obviously the program seller hopes to sell it at an inflated price by
associating it with the deserved popularity of the Smith Chart. A sales
gimmic !

But how ridiculous it is to deliberately spoil the ability of a computer
program by obliging it to conform to the technical shortcomings of an
accompanying status symbol.

[Craig A. Ferris]

A reflection coefficient greater than 1??? When RC is equal to 1, ALL power
is reflected.
With a RC of 2.414, 5.8 times the power is reflected. Your transmission line
generates power.
You should patent it.

73,
Craig NR4E

[tom_b...@agilent.com]

Craig,

Consider if you will a piece of RG-58 or other small line at 1MHz. No
need to go to audio as Reg might suggest (to get really large reactive
components in Z0). Then put a reasonably high Q inductor with about
+j50 reactance on that line as a load. Let's say it's 0.5+j50. The
line characteristic impedance will be somewhat capacitive, let's say
50-j3...I think it's actually a bit more than tnat. Now...what's the
reflection coefficient magnitude? When I calculate it, it's greater
than unity.

I'd suggest that you NOT get hung up in assigning a physical
significance to that. Someone once suggested to me that the formula for
calculating reflection coefficient (given Z0 and Zload) is incorrect,
but on careful, ahem, reflection, I decided it was just fine as it's
normally given. Good idea to go back to a clear definition of
reflection coefficient and not read anything into it that isn't given by
the definition or justified by math relationships with other concepts.

Cheers,
Tom

In article <3A7F2632...@aeronix.com>,

"Craig A. Ferris" <cfe...@aeronix.com> wrote:

--

[Reg Edwards]

Hello Craig,

As you have found, your trial calculation gives an impossible result.
Perhaps the formulae you have used to calculate forward and reflected power
make the same assumptions as Smith and others adopted in order to generate
curves which could be usefully printed on a flat rather than on a useless
3-dimensional sheet of paper.

Hint 1: When calculating power from volts and current it is essential to
take into account their relative phases.

Hint 2: On REAL lines the current and voltage are not in phase with each
other because the characteristic impedance Zo is never a pure resistance.

As Ian, G3SEK, pointed out, one of Smith's convenient assumptions was that
Zo is always a pure resistance. It never is. Oliver Heaviside, of
ionospheric-layer fame, realised this way back in 1875 or thereabouts when
he solved fundamental problems relating to distortion of telegraph signals
along transmission lines and a little later distortion in telephone lines.
Data rates even in those days were of great importance. Those solutions,
ridiculed for years by Oxbridge PhD blockheads, are today equally applicable
to transmission of digital signals and TV waveforms.

It was American engineering mathematicians such as Carson and Armstrong who
took Heaviside's work into the 20th century and into radio. The Yanks, if
nothing else, are essentially practical !
--
Regards, Reg, G4FGQ
***********************************

Craig A. Ferris <cfe...@aeronix.com> wrote in message
news:3A7F2632...@aeronix.com...

[tom_b...@agilent.com]

In article <95gkvr$38p$1...@uranium.btinternet.com>,

"Reg Edwards" <g4fgq...@btinternet.com> wrote:
...

> Yes, it is the Reflection Coefficient which the Smith Chart asserts
cannot
> rise above unity whereas on real lines the RC can approach 1 +
sqrt(2).

Hi Reg,

OK, can we get down to brass tacks here? Are you suggesting that when I
scale all impedances by Z0 that I can't read correct values from the
chart, that they are inherently incorrect? Did you happen to quit
drawing the grid lines right at a unit circle from the origin??

Specific example: scale to Z0 = 50-j10. Then, rho = 0+j1 results from
a load of Z = 10+j50. The normalized impedance point that plots to is
0+j1, which seems right to me. For rho = sqrt(1/2)+j*sqrt(1/2), the
load is +j2.414*Z0, which again is just what I expect from the
chart...in fact quite in general, the chart simply plots for grid lines
a field of Z(normalized) = (1+rho)/(1-rho), and you are not constrained
to do that for |rho| <= 1 even though the printed charts typically stop
the grid lines there, out of convenience. So I come down to the
ultimate question: is it not true that Z = Z0 * (1+rho)/(1-rho)?

Note that the instructions for plotting normalized Z for rho = 0+j2.4142
say it will be -0.707+j0.707; so if my transmission line Z0 is 705-j705,
then a load of 0+j997 ohms results in a reflection coefficient of
0+j2.4142. I _still_ don't have any problem with that... the bumble
bee is still flying just fine, thank you.

[Roy Lewallen]

The Smith chart is easily extended to include values of reflection
coefficient greater than one. This is often used for amplifier stability
analysis and oscillator design. For examples, see Fig 6.2 in Vendelin,
Pavio, and Rohde, _Microwave Circuit Design Using Linear and Nonlinear
Techniques_ (Wiley, 1990), Fig. 2-27 in Rhea, _Oscillator Design and
Computer Simulation_ (Noble, 1995), or Figs. 2.2.4 and 2.2.5 in
Gonzalez, _Microwave Transistor Amplifiers_ (Prentice Hall, 1984).

The limitation of real characteristic impedances would be a severe
handicap in the MF and below world where Reg apparently spent most of
his career. But with decent quality cable at HF and above, the error
introduced by the limitation is typically much smaller than your ability
to discern it. And 50 ohms resistive is so widely used for network
design that the limitation is no handicap at all in the vast majority of
circuit applications. The Smith chart continues to be a wonderfully
intuitive tool for many design problems. If accuracy beyond a couple of
significant digits is necessary, the engineer should think carefully
about the reproducibility of the design using real world components in a
real environment. And in that case, computer analysis of not only the
design itself but of the effect of component and environment variations
is called for.

Roy Lewallen, W7EL

tom_b...@agilent.com wrote:
>
> Without going back to specific articles, I'm remembering that you've
> commented on two specific shortcomings of the Smith chart: that you
> must scale only to real values of reference impedance and that you
> cannot use the chart with reflection coefficient magnitudes greater
> than unity. I don't have any trouble using it with reference
> impedances which are reactive, and I don't have any trouble using it
> to see reflection coefficients greater than unity. The numbers I gave
> were my perhaps feeble attempt to demonstrate that.
>
> In words, the Smith chart is simply a normalized impedance (and/or
> admittance) grid overlaid on a linear complex reflection coefficient
> plane, and some enhancements to use with that grid/plane. As far as I
> know (and indeed as far as the equations and specific numerical examples
> show), the grid may extend well beyond the commonly plotted circle, if
> it suits your needs for a particular problem.
> . . .

[Reg Edwards]

Roy Lewallen wrote

> The limitation of real characteristic impedances would be a severe
> handicap in the MF and below world where Reg apparently spent most of
> his career

=====================================
Entirely irrelevant and wildly incorrect !

But it does raise questions about the motive in mentioning it.

[Donaly]

For low frequency lines, it's possible to modify the line in
such a way as to make Z0 a real number by making R/L=G/C. This
makes Z0 equal to the square root of L/C. Heaviside and Preece
argued this all out in the 19th century in relation to distortion
in telephone lines. This is elementary and I'm surprised none of
the controversialists on this newsgroup has bothered to mention it.

Tom Donaly KA6RUH

[tom_b...@agilent.com]

In article <95tunh$8sk$1...@neptunium.btinternet.com>,

Hmmm...I think you've quoted that out of context. I think Roy's comment
was firmly in the context of Smith charts: _IF_ they are limited to
handling (being normalized to) real Z0 only, that would be a handicap in
the in a world where lines commonly have a complex Z0. I quite frankly
have a lot of respect for someone who spent a lot of time working on
lines where the simplification of assuming real Z0 was inappropriate, in
an era before ready availability of calculators with hyperbolic
functions, and personal computers with scientific spreadsheet programs
and the like.

(Of course, as I've said, I don't believe that Smith charts _are_
limited to real-valued normalizations, nor to |rho|<=1. And I also
don't believe that one always has the option to use lines that have
real-valued Z0 at all frequencies of interest. A Smith chart displayed
on a computer screen, where the computer takes care of rapid, accurate
calculations, lets me easily use the visualizations I find so helpful
while removing the tedium of part value calculations,normalization and
plotting, and making it easy to read all sorts of different values to
whatever precision is appropriate.)

Cheers,
Tom

[Roy Lewallen]

I have to plead ignorance in not having read Heaviside's and Preece's
19th century works, nor ever having had to design telephone transmission
systems. Is this modification commonly done, and why? And briefly, how
is the line constructed or modified to make this happen? You're pretty
well stuck with R for a given conductor size, so it appears you'd have
to intentionally increase shunt loss G and/or make the nominal
high-frequency impedance (sqrt(L/C)) extraordinarily high. You can't get
a very high L/C ratio in practice with reasonable conductor spacing, so
my guess is that you'd end up having to intentionally increase G. Or is
it just done by periodically inserting lumped inductors? But this is
pure speculation on my part -- it's not clear to me why you'd
necessarily want or need a purely real Z0 in the first place. Why is it
done?

It's possible that it hasn't been mentioned because of the general lack
of experience of the posters (some of us at least) in dealing with low
frequency lines. I had thought Reg had, but I seem to have been mistaken
on that account.

Incidentally, my motive for mentioning it was an attempt to understand
why Reg feels that not being able to easily normalize a Smith chart to a
complex impedance is such a handicap, while the multitudes of daily
users of the charts aren't bothered. Since I was wrong about his MF and
below experience, I'm again at a loss to explain it.

Roy Lewallen, W7EL

[Donaly]

Reg was going on about the Smith Chart assuming a real Z0, when
in fact, as Tom Bruhns pointed out, the assumption isn't necessary.
Evidently, in the early days of telephony, the distance one could
transmit a signal was limited by the dispersiveness of the line.
Heaviside realized that, given the condition that R/L = G/C,
the phase velocity of a given line would be the reciprocal of
the square root of LC. This is a constant which would, theoretically,
produce a distortionless line. Heaviside advocated the use of
lumped inductors spaced along the line to provide the needed inductance.
British electrical engineers of the time didn't think much of this idea,
so it was left to an American, George A. Campbell, working for The
American Bell Company to design the first practical coil.
I don't know whether or not anyone cares about low frequency
transmission
lines any more. I posted what I did to gently remind Reg that the
assumption that all transmission lines have a large reactive component
at
low frequencies is not always valid.

Tom Donaly KA6RUH

[Harry Brown]

Robert,

If someone else has not already suggested it, an excellent FREE Smith Chaer
Program is Motorola's MIMP program. It should be available from the Motorola
semi web page. If not, I can e-mail a copy.

73, Harry, W3IIT
harry...@earthlink.net

<rob...@my-deja.com> wrote in message news:959j1s$ms1$1...@nnrp1.deja.com...

> I am looking for a good free smith chart program for windows. I've
> looked around the search engines and have only found old dos programs.
> Does anyone have any suggestions?
>

> Thanks,
> Robert

[Roy Lewallen]

You're correct that a circuit has to generate power to have a reflection
coefficient greater than one. You'll encounter this negative resistance
condition in some active circuits such as oscillators or amplifier
inputs, where the Smith chart is also used. Also, in some phased arrays,
one or more elements send power toward the source, so there's a negative
resistance looking into the transmission lines to those elements from
the source. The power is, of course, coupled from the other elements.
For sure, though, you'll never see a reflection coefficient greater than
one in a totally passive circuit.

Roy Lewallen, W7EL

[Roy Lewallen]

Tom Bruhns dropped me a note pointing out that the magnitude of the
reflection coefficient can indeed be greater than one in passive
circuits. Apparently Reg also mentioned this in a posting, which I
didn't see or I overlooked. My apology for missing this.

I'm going to use rho to mean the magnitude of the reflection
coefficient, not the complex reflection coefficient itself, here.

It turns out that the condition for rho to be greater than one at the
load is simply R * R0 + X * X0 < 0 where R0 + jX0 is the transmission
line characteristic impedance Z0, and R + jX is the load impedance Z.
When doing analysis at HF and above, Z0 is almost always assumed to be
purely real, because it is very nearly purely real for ordinary
transmission lines. (An exception would be ones which are much lossier
than would normally be considered for common uses.) When Z0 is purely
real, X0 is by definition zero, so the only way that the condition for
rho to be greater than one can be satisfied is for R to be negative.
(Negative R0 could also do it, but this could surely never happen with a
physical transmission line.) Consequently, it's common to assume, as I
did, that the only way rho can be greater than one is for R to be
negative, which can't happen in a purely passive circuit.

But when Z0 is complex, even slightly so, rho can indeed be greater than
one, and it's easy to find combinations of Z0 and Z that'll make rho
greater than one while maintaining positive R and R0.

Again, my apology to Reg for missing his reference to this, and my
thanks to Tom for getting my attention and pointing it out to me.

Roy Lewallen, W7EL