******************************************************************
* ARD9410.TXT -- Ver 1.0, February 21, 1995                      *
*                                                                *
* This file is most readable when viewed in a fixed-pitch        *
* type font, such as Courier or New Courier.                     *
*                                                                *
* Source: The American Radio Relay League (ARRL)                 *
*         225 Main St                                            *
*         Newington, CT 06111                                    *
*         USA                                                    *
*         tel 203-666-1541                                       *
*         e-mail: hq@arrl.org                                    *
*                                                                *
* Subject: Overview of the features and capabilities of ARRL's   *
* radio-frequency (RF) computer-aided-design (CAD) software,     *
* ARRL Radio Designer.                                           *
******************************************************************

For more information about ARRL Radio Designer, address an
e-mail message to ARRL's automated Internet information server:

info@arrl.org

including as your message's sole text the four lines

SEND ARD.TXT
HELP
INDEX
QUIT

You need not specify a subject unless your mailer requires it.
The HELP line nets you the info server Help file; the INDEX line
nets you the latest listing of all information files on the
server. You can also download the latest ARD.TXT from the ARRL
BBS at 203-666-0578.

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

This article appeared on pages 21-26 of October 1994 QST(R)
magazine, is copyright 1994 by the American Radio Relay
League Inc, and is provided herein solely for the personal use
of the recipient. (QST[R] is published monthly by the American
Radio Relay League (ARRL), a membership organization
of, by and for radio amateurs.) In the text to follow, \ /
delimiters indicate \superscript/; / \ delimiters indicate
/subscript\; and ** delimiters indicate *italic* type.

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

Introducing ARRL Radio Designer: New Software for RF Circuit 
Simulation and Analysis

If you've done any hobby computer-aided design at all, you've 
probably used CAD mainly for designing and enhancing your antenna 
system. Now you can put your PC to work modeling radio circuitry 
at the *station* end of your feed line!

By David Newkirk, WJ1Z
Senior Assistant Technical Editor

A glance into just about any current Amateur Radio magazine or 
club bulletin confirms that computerized antenna modeling now 
qualifies as a standard ham activity. With program names like 
NEC, MiniNEC, MN, ELNEC and ARRL MicroSmith well-established as 
household words--in ham households, at least--we don't even blink 
when someone gloats over another computer-optimized Yagi or gulps 
when modeling unmasks a new skywire as more of a worm warmer than 
an ether buster.

What, then, keeps so many of us from modeling radio *circuits* 
with our computers--designing, simulating and analyzing the 
innards of the "gray boxes" we connect to the antenna systems 
model with such enthusiasm? Availability, for starters. If 
versatile, affordable RF CAD software exists, how do we find it? 
Even the worthiest of the uncountable neat little (and not so 
little) utilities written by hams to solve or simulate or design 
particular radio-electronics problems or circuits rarely makes 
headlines.\Note1/

********
\Note1/Did you catch Dean Straw's "So What's New in The ARRL 
Antenna Book?" (the lead article in last month's QST) or checked 
out the companion software available for The ARRL UHF/Microwave 
Experimenter's Manual?
********

Program suitability to the task of realistic RF modeling is 
perhaps the biggest hurdle. General-purpose simulators like 
*PSpice*(TM) and *MicroCAP*(TM) are well-established in college-
level EE programs. They are offered as low-cost, general-purpose 
simulators, but the accuracy of their available active-device 
model libraries, especially in the important areas of noise 
(noise-correlation matrix calculations) and distributed parasitic 
reactances so important in RF modeling, significantly limits 
their usefulness above 100 MHz. What's more, these programs don't 
"speak RF"--they're not equipped to directly report circuit 
performance in RF-standard terms like S (scattering), Y 
(admittance), impedance (Z) and other network parameters.

Now there's a new choice. Working in association with Compact 
Software of Paterson, New Jersey, ARRL is proud to unveil ARRL 
Radio Designer 1.0--realistic, affordable (price class, $150) RF 
CAD Windows(TM) software for radio amateurs!\Note2/

********
\Note2/ARRL Radio Designer is available from Publication Sales at 
ARRL HQ for $150, plus $5 shipping/handling (UPS delivery). You 
can order by phone (203-666-1541) or use the order form in the 
ARRL Publications Catalog elsewhere in any issue of QST (the 
order number is 4882). ARRL Radio Designer software and example 
files are shipped on two high-density 3-1/2-inch IBM compatible 
diskettes. The instruction manual includes tutorial and reference 
information. See the article text for computer hardware 
requirements.
********

What is ARRL Radio Designer?

ARRL Radio Designer, a derivative of Super-Compact(R), Compact 
Software's industry-standard linear circuit simulator, analyzes 
the performance of linear, small-signal active and passive dc, AF 
and RF circuitry, including amplifiers, filters, matching 
networks and power splitters and combiners. ARRL Radio Designer 
tools include

* Analysis (prediction of circuit performance);
Optimization (automatic adjustment of circuit performance to meet 
goals you specify);

* Voltage Probe (predicts the signal level at any point in a 
simulated circuit);

* Statistical Analysis (simulates the effect of component value 
variations [such as those attributable to tolerance or 
temperature coefficient] on circuit performance using Monte Carlo 
techniques);

* Time Domain Analysis (simulates circuit performance in response 
to a steady-state time-domain signal using impulse, step, pulsed 
carrier or user-defined stimuli);

* Manual matching-network synthesis via Circles, an interactive 
Smith Chart utility; and

* Databanks (device-manufacturer-supplied S-parameter and noise 
data you can incorporate in your circuit simulations).

ARRL Radio Designer reports the results of its simulations in 
graphical (rectangular and polar) and tabular form, onscreen and 
via any Windows (TM)-compatible printer, in terms of

* S, Y, Z, group delay and voltage probe parameters for n-port 
networks;

* Chain (ABCD), hybrid (H), inverse hybrid (G), gain, voltage
gain, and stability parameters for two-port networks;

* magnitude of reflection coefficient, phase of reflection 
coefficient, VSWR and return loss parameters for one-port 
networks;

* gain, gain matching and noise parameters; and

* complex S, Y, Z, H, G, chain (A), gain matching, noise matching 
and voltage probe parameters.

Installing and running ARRL Radio Designer requires, at minimum: 
a 386, 486 or Pentium IBM PC or 100% compatible (math coprocessor 
not required, but strongly recommended); 8 Mbytes of RAM;\Note 3/

********
\Note 3/For those who must ask: ARRL Radio Designer has been 
successfully installed and used on a coprocessorless 386SX-16 
with 4 MB of RAM running in 386 enhanced mode. This required 
critical management of Windows resources, ARRL Radio Designer ran 
quite slowly because of the system's slow clock speed, 386 CPU 
(generally, 386s take two clock cycles to do what a 486 does in 
one) and lack of a coprocessor.
********

* a 3.5-inch, high-density floppy drive; a hard disk with at 
least 5 MB of free space;

* Microsoft Windows (TM) 3.1 or higher; and

* a mouse or equivalent pointing device.

So much for the fact dump. You've got to see ARRL Radio Designer 
in action to appreciate it, so let's put it work!

ARRL Radio Designer in Action

We'll use the circuit shown in Figure 1--the post-mixer amplifier 
from Hayward and Lawson's Progressive Communications receiver 
(November 1981 QST)--as our first test case. Reduced to its 
essentials, the process involves just four steps:

1. Mark each of the circuit's *nodes*--points of interconnections 
between its components or *elements* (Table 1 lists ARRL Radio 
Designer's entire set) with an exclusive number between 0 and 
999 (Figure 1);
2. Type the circuit's elements and node numbers into a *netlist* 
using ARRL Radio Designer's Circuit Editor (Figure 2);
3. Press ARRL Radio Designer's Analyze button; and
4. Graph, table, save and/or print the results to your heart's 
content (Figures 3, 4 and 5).

A 6-dB attenuator follows this post-mixer stage in the 
Hayward/Lawson receiver. With it in line, they reported the 
amplifier's gain as "about 16 dB." Subtracting 6 dB from the 
MS/21\ trace in Figure 4 or Figure 5 reveals that our model's 
HF/VHF gain agrees quite closely with the findings of Hayward and 
Lawson.

Circuit Tuning

How does a double-tuned-circuit filter's response shift as you 
tune it with a two-section capacitor? Sure, you can *hear* the 
peak sweep by if the filter's in your receiver, but what does 
that effect *look* like? If you happen to own a spectrum analyzer 
and tracking generator, you can demonstrate it easily enough. Or 
you can let ARRL Radio Designer's Tune feature show you (Figure 
6).

Statistical Analysis, Monte Carlo Style

Real components vary in value with temperature and tolerance. 
Using ARRL Radio Designer's statistical analysis feature, you can 
determine the effect of these variations on circuit performance, 
as I did for the op-amp audio filter described by Henry J. 
Perras, K1ZDI, in March 1994 Hints and Kinks. Figure 7 reports 
the happy news that 1000 out of 1000 K1ZDI filters built with 5%-
tolerance capacitors and 1%-tolerance resistors should work 
acceptably well with *no post-construction tweaking whatsoever*, 
and Figure 8 shows how those 1000 filters' 2-kHz losses vary 
around the nominal value of 2.52 dB as result of component 
tolerances.

Putting Optimization on the Case

It's one thing to use "cut and try" techniques to nail down one 
variable component value in a circuit when you've got all the 
others under control, but you can just about kiss a weekend's 
worth of experimenting goodbye if you need to vary more than one. 
If the optimization job you want to do requires esoteric test 
equipment--say, a noise figure meter--or presents a "solution 
surface" so complex that one-dimensional tweak-and-measure 
investigation is doomed from the start, you and ARRL Radio 
Designer's Optimization engine stand to become fast friends. 
Figure 9 shows a simple example of such a challenge--an antenna 
matcher with one variable inductor and one variable capacitor. 
Your job: Find the L and C values that let that tuner turn a 
highly reactive antenna load (16 - j2256 ohms at 1.83 MHz) into 
50 ohms, resistive. Your choice: Put ARRL Radio Designer on the 
case (Figure 10). Your reward: a match (Figure 11).

Let the Fun and Learning Begin!

That's all the space we can devote this month to QST's first look 
at ARRL Radio Designer. What happens next? For starters, you can 
expect to see more about ARRL Radio Designer in future QSTs--
whenever we can appropriately use it to confirm or improve a 
circuit design, or to illustrate a point, for instance. Our 
overriding hope, though, is that you will put ARRL Radio Designer 
to work, and share your experiences and findings with us and your 
fellow hams. The way we see it, it's only a matter of time before 
RF-accurate CAD becomes every bit as routine to radio amateurs as 
simulating an antenna with MiniNEC or cruising through a contest 
with CT or NA. We're pleased to offer ARRL Radio Designer as a 
doorway into that new phase of ham radio's growing computer 
tradition.

[sidebar]
Circuit Analysis, Circuit Synthesis--What's the Difference?

ARRL Radio Designer concentrates on circuit analysis as opposed 
to circuit synthesis, but what does that mean? With a circuit 
synthesis program, you define a problem ("Build me a fifth-order 
Chebyshev low-pass filter with 50-ohm terminations, 0.1 dB of 
ripple and a 3-dB corner frequency of 230 MHz") and the program 
responds with appropriate component values. Circuit analysis 
capability, if present, is usually limited to relatively simple 
simulations of the program's own solutions.

A circuit analysis program, on the other hand, predicts the 
performance of your solutions. You enter your circuit into the 
program in coded form (commonly, as in ARRL Radio Designer, this 
is a text file called a netlist, short for network list) and 
provide guidance for the program's simulation engine ("Calculate 
this circuit's behavior at 500 exponentially stepped frequencies 
from 100 kHz from to 100 MHz, and graph the real and imaginary 
components of its input impedance with its output terminals 
loaded by 1.5 kilohms in parallel with 3 pF"). The appearance, 
accuracy and quality of the results you get depend on how 
completely you specify (and how completely you the program lets 
you specify) your circuit's component characteristics and 
terminations, the accuracy of the program's mathematical 
component models, and the program's reporting capabilities.

ARRL Radio Designer goes far beyond plain-vanilla analysis, 
however. Its Circles utility can help you synthesize matching 
network values, Smith Chart style. You can command its Tune 
function to step selected components' values through sequences or 
ranges of values as you watch the results onscreen. You can 
predict the effect of component tolerances on circuit performance 
using Statistical Analysis. Even more excitingly, you can put 
ARRL Radio Designer's Optimizer to work tweaking your circuit for 
peak performance--hand it your best cut at a low-noise 2-meter 
preamp, say, and walk away with a design that's a dB or two 
quieter. True to its name, ARRL Radio Designer doesn't do "just 
synthesis"--it helps you design radios.--WJ1Z

[sidebar]
ARRL Radio Designer Versus Professional Simulators

Even circuit-simulation software costing tens of thousands of 
dollars or more--as the best such programs do--can't accurately 
model absolutely every circuit function in, say, a shortwave ham 
transceiver. Since even the best simulators can't do everything, 
what subset of "less than everything" can ARRL Radio Designer do? 
To put it another way, what do professional-grade simulators got 
that we ain't got? The answer has three parts: schematic capture, 
microwave/optical capability, and nonlinear simulation.

Schematic Capture

Schematic capture, standard with some high-end circuit simulators 
and optional in others, lets you draw a schematic onscreen and 
generate a netlist--the circuit's text equivalent in a form 
digestible by the simulator--from the drawing. Super-Compact, 
ARRL Radio Designer's big brother, doesn't include schematic 
capture (Compact's Serenade schematic editor is available as an 
option); its text-based netlist interface was designed for 
compatibility across platforms with widely variable graphics 
capabilities. ARRL Radio Designer therefore doesn't include 
schematic capture. For sprawling digital circuits, schematic 
capture, though tedious, is almost a necessity. For RF 
applications and simple circuits, however, manual entry is 
significantly faster and easier.

Microwave/Optical Capability

The lion's share of today's commercial and military radio R & D 
bucks flows into UHF/microwave and fiber-optics projects. At 
those frequencies, tuned circuits rarely consist of "lumped" 
inductances and capacitances--a tuned circuit consisting of 
0.0001 pF in parallel with 0.0001 mH resonates at about 1592 GHz, 
but I dare you to actually build one!--so stripline, microstrip 
and other transmission-line and waveguidelike structures do the 
job instead. Professional-grade circuit simulators can model 
these structures using actual physical dimensions and the real 
characteristics of their conductors, dielectrics and substrates. 
In contrast to this, ARRL Radio Designer models tuned circuits in 
terms of lumped L and C. Its practical applicability therefore 
declines wherever radio physics dictates that you must switch 
from LC circuits to stripline, microstrip, YIGs, dielectric 
resonators or waveguides--typically above 1 GHz.

Small- Versus Large-Signal Analysis

Oscillators stabilize their amplitudes; mixers mix; modulators 
modulate; amplifiers distort; rectifiers rectify; frequency 
multipliers multiply; and AGCed stages "AGC" because amplitude-
nonlinear device behavior. Because ARRL Radio Designer is a 
linear, small-signal simulator, it cannot simulate these large-
signal effects.

To get a handle on this, check out the bipolar junction 
transistor amplifier in Figure 1. That drawing includes no power 
supply connections--because the ARRL Radio Designer model doesn't 
need them! You can successfully analyze the circuit's 
performance, and plot its gain and frequency response, without 
connecting its bipolar junction transistor to V/CC\. How can this 
be?

This can be because the Figure 1 circuit is a linear amplifier, 
and ARRL Radio Designer is a small-signal, linear simulator. 
When, working at your radio bench, you power a real transistor at 
a particular voltage and bias it just so, you also set its 
supply- and bias-dependent parameters (starting with a bipolar 
junction transistors alpha or a FET's transconductance) to 
particular values. Modeling in ARRL Radio Designer, you specify 
your devices' parameters according to  the power-supply and bias 
levels you expect to exist in the modeled circuit, and ARRL Radio 
Designer then reports your circuit's performance with devices 
exhibiting those parameters.

The distinction between small-signal and large-signal analysis 
pretty much boils down to this: A device operates in its small-
signal region when its "operating" or "bias point" (which its 
power-supply and bias levels determine) doesn't shift in response 
to its input signal. Driving the device with a signal large 
enough to shift its bias point results in nonlinear, dynamically 
variable performance that the "hardwired" device parameters of a 
linear simulator's netlist don't reflect.

ARRL Radio Designer therefore can't model nonlinear effects like 
frequency translation, intermodulation, AGC, and the results of 
subjecting good devices, or well-modeled designs, to absurd or 
greatly divergent bias, drive or power-supply levels. If, for 
instance, you want to see what happens when you power a 13.8-V 
amplifier design at 6 V, you can model its gain under the new 
conditions. Working carefully, you can even simulate the 
frequency- and phase-response shifts that may occur as a 
transistor's internal capacitances change in response to 
different terminal voltages and currents. But you won't be able 
to tell if your underpowered amplifier is any more or less linear 
because ARRL Radio Designer's active devices are crunchproof and 
distortion-free by definition!

Nonlinear Analysis

Professional-grade nonlinear simulators, including Compact's 
Microwave SCOPE and Microwave Harmonica, can do most types of 
nonlinear AF and RF simulation extremely well--beginning at 
roughly the cost of a new car. PSpice, largely intended for dc, 
digital and audio design, and available (in a limited-capability 
evaluation version) as freeware from computer bulletin boards, 
can usefully simulate some nonlinear effects--if you don't need 
accurate results at VHF/UHF frequencies. So where does that leave 
linear, small-signal ARRL Radio Designer?

In good company, as it turns out. Compact Software gurus tell us 
that perhaps 90% of all professional microwave and RF circuit 
simulators sold are linear. Even though we stand on the threshold 
of the DSP age, our radios still consist largely of linear 
circuits. Because linear AF and RF design techniques will remain 
important in professional and Amateur Radio for the foreseeable 
future, an affordable RF-circuit simulator stands to be of 
significant use to hams interested in getting current and keeping 
current with modern RF techniques. That's why we're pleased and 
excited to bring ARRL Radio Designer to you.--WJ1Z

Table 1 [page 23]
ARRL Radio Designer Circuit Elements

Passive Lumped Elements
CAP     Capacitor models
DIOD    Diode models
DLY     Time delay
IND     Inductor models
MUI     Coupled inductor models
PLC     Parallel connection of inductor and capacitor
PRL     Parallel connection of resistor and inductor
PRC     Parallel connection of resistor and capacitor
PRX     Parallel combination of resistor, inductor and capacitor
RES     Resistor models
SHO     Short circuit
SLC     Series connection of inductor and capacitor
SRL     Series connection of resistor and inductor
SRC     Series connection of resistor and capacitor
SRX     Series connection of resistor, inductor and capacitor
TRF     Transformer models (ideal two- and three-winding types)

Black-Box Elements
IMP     Two-terminal impedance
ONE     Two-terminal element specified by admittance, impedance or 
          reflection coefficient
TWO     Three-terminal two-port specified by admittance, impedance 
          or S parameters

Distributed Elements
CAB     Coaxial cable models
TRL     Transmission-line models

Controlled Source and Active Elements
CCG     Current-controlled current source
CVG     Current-controlled voltage source
BIP     Bipolar transistor model
FET     Field-effect transistor model
OPA     Operational amplifier
VCG     Voltage-controlled current source
VVG     Voltage-controlled voltage source

Figure 1--We'll simulate the post-mixer amplifier from the 
classic Progressive Communications Receiver described in November 
1981 QST by Wes Hayward, W7ZOI, and John Lawson, K5IRK. The first 
step in readying this circuit for ARRL Radio Designer analysis 
merely involves marking its nodes--its points of component 
interconnection--with numbers, 0 being the default for circuit 
common (I circle mine to distinguish them from device pinouts, 
wire designators, and so on). How come no power supply 
connections? The "ARRL Radio Designer Versus Professional 
Simulators" sidebar tells why.

Figure 2--Next, we use ARRL Radio Designer's Circuit Editor--just 
an ASCII word processor--to specify each of the amplifier's 
components in network list--netlist--form:

FT:1.4E9 ; Hz (1.4 X 10\9/ Hz is 1.4 GHz)
Ic:30  ; mA
Rd:(26/Ic)
B:49
BLK
CAP 1 2 C=0.01UF
RES 2 0 R=470
RES 2 3 R=1000
BIP 2 4 5 A=(B/(B+1)) RE=RD CE=(1/(FT*2*PI*RD)) RB1=7.5
MUI 4 3 3 0 L1=46UH L2=46UH K=.999
CAP 3 7 C=0.01UF
RES 5 0 R=56
CAP 5 6 C=0.01UF
RES 6 0 R=5.6
BJTAMP:2POR 1 7
END
FREQ
ESTP 1MHZ 200MHZ 100
END

Highlights in this particular netlist:

* The four lines after the file header contain data for use by 
the circuit's transistor model. They are FT (f/T\, current gain-
bandwidth product), Ic (I/c\, collector current), Rd (r/d\, 
diffusion resistance) and B (beta, current amplification factor).

* BLK marks the beginning of a netlist block.

* The BIP line specifies a bipolar transistor intended to 
simulate a 2N5109 running at 30 mA of collector current, using 
equations to derive the transistor's alpha (A) from beta 
(B/(B+1)) and emitter capacitance (CE) from f/T\ and r/d\ 
(1/(FT*2*PI*RD)). The model's emitter resistance (RE) is set 
equal to R/d\.

* The MUI line specifies mutually coupled 46-uH inductors with a 
coupling coefficient of 0.999 (the number 9s in effect sets the 
number of decades of frequency range);

* The BJTAMP:2POR 1 7 line names the circuit BJTAMP and defines 
it as a two-port network with terminals at nodes 1 and 7; and

* The FREQuency block tells ARRL Radio Designer to simulate 
BJTAMP's performance at 100 exponentially stepped frequencies 
from 1 to 200 MHz.

Our netlist completed, we click on ARRL Radio Designer's Analyze 
button and go!

Figure 3--The analysis done, we pop ARRL Radio Designer's Linear 
Reports dialog to specify the simulated performance parameters we 
want to see (the S parameters MS/21\ [magnitude of forward gain], 
MS/11\ [magnitude of input reflection] and MS/22\ [magnitude of 
output reflection], and NF [noise figure], and how we want them 
displayed.

Figure 4--How ARRL Radio Designer delivers the goods onscreen. 
You can control the appearance of everything you see in an ARRL 
Radio Designer report--colors, fonts, titles, graph scaling, the 
works.

Figure 5--Need hard copy? Here's how the Figure 4 report rolls 
out of a laser printer in portrait orientation. ARRL Radio 
Designer's gives you wide control over its printer output, 
including line weights, fonts and color. (We don't have a color 
printer, so I set all of the report traces' colors to black for 
this version.)

Figure 6--Here's what happens when ARRL Radio Designer's Tune 
feature steps a double-tuned-circuit filter's two-section tuning 
capacitor through part of its range in 5-pF increments. The 
filter simulated (a bottom-coupled 80-meter design from Solid 
State Design for the Radio Amateur) uses a small inductor for 
coupling. The coupling inductor's reactance, and therefore the 
coupling between the filter's resonators, increases with 
frequency, and you can see the effect of this as a barely 
perceptible insertion-loss decrease as the filter sweeps from 3.5 
to 4.0 MHz. When you activate Tune, ARRL Radio Designer switches 
into Accumulate mode to let you see up to 20 simulations at once.

Figure 7--You can use ARRL Radio Designer's Statistics feature to 
explore the effects of component tolerances on circuit behavior. 
In this analysis, ARRL Radio Designer evaluated 1000 iterations 
of an audio filter design while varying its frequency-critical 
component values within specified tolerances, pass/fail-testing 
each trial on the basis of loss at 2 kHz. The target loss range 
was 0 to -5.52 dB.

Figure 8--Clicking on Histogram charts how the filter's 2-kHz 
loss varies around the nominal value of -2.52 dB across a 
population of 1000.

Figure 9--Ulrich L. Rohde, KA2WEU, described this antenna 
matching network in November 1992 QST's "Recent Advances in 
Shortwave Receiver Design." Finding the L and C values that let 
this network match a particular load--assuming that that the load 
is within the network's matching range--is a piece of cake for 
ARRL Radio Designer's Optimizer.

Figure 10--ARRL Radio Designer's Optimization engine lets you use 
random and gradient techniques to zero in on complex circuit 
solutions you couldn't hope to achieve by cut and try. This 
screen dump catches ARRL Radio Designer in the act of adjusting 
the Figure 9 network to match a 16 - j2256 ohm load to 50 ohms at 
1.83 MHz. (In practice, turning off the Optimizer's display 
feature greatly speeds a solution because the computer doesn't 
have to stop to calculate and redraw the graphs each time it 
tries new values.) The optimization goal: simultaneous 
achievement of tuner-input Z parameters of RZ/11\=50 (real part 
of impedance, 50 ohms) and IZ/11\=0 (imaginary--reactive--part of 
impedance, 0 ohms)--"50 ohms, resistive."

Figure 11--We have a match! (About 1800 iterations later, ARRL 
Radio Designer had fought the error function down to a number on 
the order of 10E-5, so I declared the job done.) Now compare 
these traces' 1.83-MHz values with the optimization goal I stated 
in the Figure 10 caption.

