Section 1.
0 : Introduction
By Martin J. King, 07/05/02
Copyright © 2002 by Martin J. King. All Rights Reserved.
Section 1.0 : Introduction
Over the years, many articles have appeared in Speaker Builder (now
audioXPRESS) presenting test data and theories related to transmission line
loudspeaker behavior. In the issues following each article, letters to the editor have
debated the merits of the author’s results. Some of these articles have been heatedly
discussed over the Internet without any consensus of opinion being reached.
Transmission line designs elicit strong differing opinions that seem to be largely based
on personal experience or vague design guidelines, rules of thumb, that are quoted as
fact without any specific reference source being provided. There does not appear to be a
generally accepted transmission line mathematical model, similar to the closed and
ported box models, where one can give the numerical values of several key parameters
and uniquely define the configuration being discussed.
Over ten years ago, I heard my first transmission line system at the home of a
local audio club member. The quality of the bass reproduction was impressive and my
interest in this exotic enclosure was stimulated. Since then I have read most of the
technical literature on the design of transmission lines, followed the frequent discussions
on several e-mail lists and bulletin boards, and visited a number of websites devoted to
this type of enclosure design. After many years of reading and thinking about
transmission lines, I decided three years ago, that it was time to build another set of
speakers and that I was going to seriously consider a transmission line system.
Since there was no acceptable design method available, my first step in
exploring the transmission line’s acoustic potential was to write the necessary design
software and correlate it against test results. I had already designed and built several
sealed and ported speaker systems based on the lumped parameter circuit models used
in the classic papers by Thiele(1) and Small(2-4). For each of these designs, I wrote my
own software using the MathCad(5) computer program. For me, formulating and solving
the computer simulation of a speaker system is as interesting and challenging as the
construction of the speaker system itself. After the speaker construction is completed,
measuring and achieving good correlation with the mathematical predictions is extremely
rewarding. Hopefully, a well-designed speaker system that measures as predicted also
sounds right. So I started the transmission line project by putting a MathCad computer
model together using equations from the articles I had collected and studied.
Before I go any further, I want to introduce my definition of a transmission line
loudspeaker. I define a transmission line loudspeaker as a driver mated to a resonant
tube where the natural frequencies and mode shapes of the air in the tube are used to
tailor the total system response. This definition does not include any restrictions on the
location of the driver in the tube or the boundary conditions at either end of the tube.
Also, this definition does not place any requirement on the amount or type of fiber
stuffing material that may be placed inside the tube to attenuate the standing waves
associated with the tube’s natural frequencies. This is a very broad definition. What
follows is my design method for a quarter wave length transmission line with a closed
end and an open end, or terminus, which emits sound that contributes to the system
response over the bass frequency range.
Page 1 of 5
Section 1.0 : Introduction
By Martin J. King, 07/05/02
Copyright © 2002 by Martin J. King. All Rights Reserved.
First Transmission Line Computer Models :
The first mathematical model formulated took the equivalent circuits used by
Thiele and Small and replaced the circuit elements modeling the boxes with a
transmission line acoustic or electrical impedance. Figure 1.1 shows the acoustic
equivalent circuit, using the impedance analogy, while Figure 1.2 shows the electrical
equivalent circuit. An excellent discussion of this kind of equivalent circuit modeling can
be found in the referenced acoustics text by Beranak(6). I am assuming that everybody is
familiar with the Thiele / Small parameters for a driver or that you can read the
references to bring yourself up to speed. All of the circuit elements in Figure 1.1 and 1.2
can be derived from the Thiele / Small driver parameters, as shown in the figures, except
the equivalent electrical impedance Zel and the equivalent acoustic impedance Zal of the
transmission line. Expressions for the transmission line electrical and acoustic
impedance are required to completely solve each of the circuits.
My first mathematical models were based on Bradbury’s(8) paper published by
the Audio Engineering Society. In his paper, Bradbury presents an elegant derivation of
the wave equation applied to sound waves passing through a fibrous tangle. At low
frequencies, the air and the fibers are coupled by a viscous damping coefficient that
drags the fibers along with the air. As sound waves pass through the fibrous tangle, the
waves are attenuated and the speed of sound is significantly reduced due to the added
mass of the moving fibers. As the frequency of the sound wave increases, this coupling
decreases and a transition is made to a stationary fibrous tangle that only attenuates the
sound waves without any reduction in the speed of sound. This particular model of
sound waves passing through a fibrous tangle is very popular and I have seen a number
of attempts to apply it to the analysis of transmission line enclosures.
I spent a long time deriving and experimenting with Bradbury’s equations. From
this effort, I extracted an expression for the acoustic impedance of the transmission line,
as seen by the driver, and an expression for the velocity of the air at the open end, or
terminus, as a function of the driver velocity. Bradbury’s equations are for a constant
cross-section transmission line with constant stuffing density. The acoustic impedance
was inserted into the equivalent circuits shown in Figures 1.1 and 1.2. After solving the
acoustic equivalent circuit in Figure 1.1 for Ud, the air velocity at the terminus can be
calculated as shown in the last four equations at the bottom of the figure. Using the
velocities of the driver and the air at the terminus, the sound pressure for each can also
be calculated. These two sound pressures can then be summed, taking into account the
relative phase angle, and the total system sound pressure level SPL and phase plotted.
There were two sources of data available to me for correlating this model.
Bullock and Hillman(9) wrote a paper in 1986 that used Bradbury’s equations to model a
test transmission line. A second source for data was a contact I made over the Internet.
My electronic contact was kind enough to provide a number of measurements for
several transmission line and driver combinations. We compared my computer model
predictions against his test data. For both sets of results, the correlation showed promise
but was clearly not accurate enough to design an enclosure. The general shapes of the
impedance and SPL response plots were close but the locations of the peaks and
valleys were shifted in frequency. This was particularly evident in the frequency range
below 200 Hz.
Page 2 of 5
Section 1.0 : Introduction
By Martin J. King, 07/05/02
Copyright © 2002 by Martin J. King. All Rights Reserved.
Although the computer model, based on Bradbury’s equations, did not correlate
as well as I would have liked, I saw enough potential to build my own test line and start
formulating a new mathematical model. My goal at this point in the project was to derive
and correlate a better mathematical model and use it to design a transmission line
system. The next step was to select a driver and fabricate a simple transmission line for
testing and measurement.
Before I could start testing, I needed to select and purchase a mid-bass driver.
Since I wanted to eventually build a two-way system for my first transmission line
project, I researched six and eight inch diameter mid-bass drivers from several highly
rated manufacturers. Eventually, I selected the Focal 8V 4412 mid-bass driver. I had
used Focal drivers in a ported three-way design several years ago and had been happy
with their performance. The drivers were purchased through Zalytron.
A test line was built and measured, as described in the next section, and data
was generated and used to correlate the simple transmission line models depicted in
Figures 1.1 and 1.2. These simple computer models were used to design the Focal two-
way transmission line, which in turn produced more data and even more questions. The
simple computer model predictions and the Focal two-way system measurements were
in reasonable agreement, but some additional work was still needed to fully understand
the measured results.
Continued Evolution of the Transmission Line Computer Models :
There have been two or three major revisions to my MathCad transmission line
computer models over the past two years. Some of the changes that were made
corrected errors in the derived equations while some of the changes extended the
capabilities of the models to represent other types of enclosures. Since September of
2000, versions of these models have been available for downloading from the Internet
and they have received wide use by transmission line and TQWT DIY speaker building
enthusiasts. To the best of my knowledge, the speakers built based on designs from
these MathCad worksheets have been very successful.
Last fall, I made a major change in the calculation algorithm. This change
increased the flexibility allowing MathCad to model many more complex transmission
line enclosures. It also became apparent that this new calculation scheme could be used
to accurately model sealed, bass-reflex, back-loaded horns, front-loaded horns, and
isobaric enclosures. In addition, just over two months ago I was able to iron out all of the
small mistakes and tricky details in my derivation of the transmission line equations and
produce what I believe to be a clean start to finish description of my latest method of
analysis. It was time to document this new body of work.
The following sections are intended to provide a smooth path to follow through
my method for modeling fiber filled transmission line loudspeakers. I welcome any
feedback or comments that will help me improve the methods or make the
documentation more understandable. Over the past two years, the correspondence I
have received in the form of comments and questions has been invaluable in helping me
further my MathCad computer models and increase my understanding of the basic one-
dimensional wave equation. I consider this document to be a dynamic work that will be
modified and improved as I learn more about transmission line loudspeakers.
Page 3 of 5
Section 1.0 : Introduction
By Martin J. King, 07/05/02
Copyright © 2002 by Martin J. King. All Rights Reserved.
Figure 1.1 : Acoustic Equivalent Circuit for a Simple Transmission Line Speaker
Ratd Mad Cad
Ud
pg Zal
where :
pg = pressure source
= (eg Bl) / (Sd Re)
Rad = driver acoustic resistance
= (Bl2 / Sd2) [Qed / ((Rg + Re) Qmd)]
Ratd = total acoustic resistance
= Rad + (Bl)2 / [Sd2 ((Rg + Re) + jω Lvc)]
Cad = driver acoustic compliance
= Vd / (ρair c2)
Mad = driver acoustic mass
= (fd2 Cad)-1
Zal = transmission line acoustic impedance
Ud = driver volume velocity
= Sd ud
ud = driver cone velocity
then :
uL = terminus air velocity
= ε ud
ε = uL / ud
Page 4 of 5
Section 1.0 : Introduction
By Martin J. King, 07/05/02
Copyright © 2002 by Martin J. King. All Rights Reserved.
Figure 1.2 : Electrical Equivalent Circuit for a Transmission Line Speaker
Rg + Re Lvc
eg Lced Cmed Red Zel ed
where :
eg = voltage source
= 2.8284 volt
Rg+Re = electrical resistance of the amplifier, cables, and voice coil
Lvc = voice coil inductance
Lced = inductance due to the driver suspension compliance
= [Cad (Bl)2] / Sd2
Cmed = capacitance due to the driver mass
= (Mad Sd2) / (Bl)2
Red = resistance due to the driver suspension damping
= Re (Qmd / Qed)
Zel = transmission line equivalent electrical impedance
= (Bl)2 / (Sd2 Zal)
ed = Bl ud
Page 5 of 5
