Transmision Line ?

If you stay away from sporting events Philly is a fairly nice place. All too many of the sports fans here are rude drunken boors.:yes:

The attendees at Jazz and classical events are IME polite and quite friendly.

One thrown bottle stopped an Aerosmith concert! What a bunch of pansies!:D
 
Misconception of transmission line loudspeaker

Really? I thought that Wikipedia entry was chock full of errors, misunderstandings and bizarre explanations.

Other than that, it was great..... :)

-k

You description of the Wiki article on transmission line loudspeaker is right on and also applicable to many of the links in that article. Transmission line is a generic term and can mean different thing in different situation. But the Bailey "non-resonant Loudspeaker Enclosure Design" published in 1965 means something very specific. Bailey had very clearly defined objective, had a thorough understanding of how acoustic wave propagate in a fibrous material filled tube and sound engineering practices to develop the experiments to support the design. 11 years later, Professor Bradbury develop the mathematical model that describes the acoustics accurately as published in the Journal of Audio Engineering Society. But there are several things that works against the Bailey design.

One of the unfortunate incident is an typo in the Bradbury paper that practically provented others to duplicate his results. The typo is actually very easy to spot if the reader follows his derivation of the equations. I will come back in a few weeks to discuss the typo to see if others also found it.

The things that work against the Bailey design are:
  1. It needs the long fiber wool to work properly. Bailey and Bradley spelled out the definition of this "long fiber wool" very clearly, but it was often ignored by builder or some self proclaimed experts. It is a low cost, stiff, thick wool used for blanket and, in the very early days, sound control in submarine. It may not be a problem for folks in UK or Austratralia. But I can imagine a US buyer may easily be convinced by a wool mill into buying expensive, fine fiber wool that does not work for the Bailey design.
  2. The propagation of sound in the long fiber wool filled tube is non-linear. Both the sound velocity and its absorption are frequency (thus, speed) dependent. It does not lean it to any kind of T-S type filter simulation that use only simple algebra.
  3. It is conceptually difficult for layman to accept Bailey design as a half wave length line without reading the Bailey and Bradbury articles carefully.
  4. The linear algebra model by Bradley does not lean itself to easy integration with the simple speaker driver model.

It is almost 40 years after the Bradbury paper and there is still no extension of his work. But the internet is flooded with simple linear analysis that ignores the effect of acoustic propagation in fibrous material. It is like cutting the foot to fit the shoes. A very sad situation.

But if you spend the time to find the right kind of "long fiber wool" and build a Bailey non-resonant speaker enclosure, you will be richly rewarded. Yes, I am a transimission line speaker fanboy.

For the record, I am an aerospace engineer, not an acoustic engineer. My specialty is in fluid dynamics and thermodynamics, not in sound. Audio is just a hobby.
 
Last edited:
MLTL=BR, elongated. The Bailey line is not a reflex. You folks who need to throw out the effect of damping, or dampening for some, would NEVER drive a car without a damped suspension, yet you blather on about the problem with long fiber wool. Einstein: simple, not too simple. Try honoring complexity, otherwise known as diversity. Rant over.
 
What Bailey and Bradbury missed was the impact the tapered TL had on the tuning frequency. There was no significant reduction in the speed of sound due to the fiber stuffing. I have done the math starting from Bailey and Bradbury models and they do not work, predictions and measurements do not agree. Many others have arrived at the same conclusion.

Both George Augspurger and I have generated computer models based on two totally different analyses without any significant slowing of the speed of sound as a result of fiber motion and the correlations between predicted and measured responses is excellent. Many DIY and commercial TL designs have been built based on our two computer models, they behave as predicted.

You can hang onto the Bailey and Bradbury work if you like but most of the modern TL designs are based on newer computer simulations that have been proven over and over again to be much more accurate. I have never seen a computer simulation based on Bailey and Bradbury models producing predictions that match measurements.
 
TL implies certain type of non-resonant driver loading

What Bailey and Bradbury missed was the impact the tapered TL had on the tuning frequency. There was no significant reduction in the speed of sound due to the fiber stuffing. I have done the math starting from Bailey and Bradbury models and they do not work, predictions and measurements do not agree. Many others have arrived at the same conclusion.

Both George Augspurger and I have generated computer models based on two totally different analyses without any significant slowing of the speed of sound as a result of fiber motion and the correlations between predicted and measured responses is excellent. Many DIY and commercial TL designs have been built based on our two computer models, they behave as predicted.

You can hang onto the Bailey and Bradbury work if you like but most of the modern TL designs are based on newer computer simulations that have been proven over and over again to be much more accurate. I have never seen a computer simulation based on Bailey and Bradbury models producing predictions that match measurements.
The nonlinear algebra model that Bradbury used were based on many military researches done in the 50's and 60's era with sound physics behind it. When you start talking about "TL tuning frequency" and "sound speed slowing by fiber motion", it shows your lack of understanding of the Bailey design. It is a "non-resonant" design to simulate infinite baffle. There were attempts to use adiabatic mixing to predict the performance of a fiber stuffed line. Bailey soundly rebuffed that in his letter to the editor. The long fiber wool works because it is stiff and heavy enough not to move, thus forces the air to move around a cylinder of the fiber that resulted in the change of sound speed. All these physics were clearly explained in the Bailey and Bradbury papers with solid research works behind them.

I do not see how you can state that Bradbury's model does not work. I cannot find any problem with his model, except the typo that I mentioned earlier. It was not Bradbury's mistake.

Neither Bailey or Bradbury were original researchers in the acoustics of a fiber filled tube. They just found those researches interesting and applied them to an audio design with great success. I know because I built one many years ago and I trust my ears.

Before you declare that Bailey's experiment was wrong, repeat his test and show how your data different from his. Also, compare the dimension of your "long fiber wool" to what the Bailey design calls out.

If you and George Augspurger like tuned vented boxes, fine. But is it necessary to call your design "transmission line" too while they are clearly NOT "non-resonant" and does not work as a phase inverter?
 
I have been through Bailey's and Bradbury's articles very carefully, rederived all of the equations and replicated the numerical results. I spent probably 10 years on and off unsuccessfully trying to make them work and match test data. One of the key assumptions in Bradbury's method was that the air and the fibers coupled and moved together at low frequencies producing a dramatic reduction in the speed of sound. Bob Bullock wrote a paper for the AES where he tried to correlate the Bradbury math against a TL that he had built and measured, he could not get things to work. I replicated his work and got the same result. Augspurger also wrote two articles for the AES and came up with the same conclusions, fibers did not slow the speed of sound.

Eventually I built a test TL and measured the electrical impedance with and without fiber stuffing. I used polyester and long fiber wool at three different packing densities. Again, Bradbury's math model did not work. I saw no evidence of a slowing of the speed of sound or a phase inversion. All of these results are available for review on my website, my method has been around for about 15 years and nobody has sent me any data that would invalidate the results and conclusions. Augspurger's AES articles have also been around for about the same amount of time and to the best of my knowledge nobody has found fault with his methods or results.

If you have your own set of models based on Bailey's and Bradbury's work and then testing that correlates with the calculations I would love to see them presented. In particular, electrical impedance and SPL output from the driver and the open end of the TL with and without fiber stuffing. If you have that kind of data it would be great and we can discuss the results in detail. If you do not have the data then continued discussion is kind of pointless and we should just agree to disagree.
 
I have been through Bailey's and Bradbury's articles very carefully, rederived all of the equations and replicated the numerical results. I spent probably 10 years on and off unsuccessfully trying to make them work and match test data. One of the key assumptions in Bradbury's method was that the air and the fibers coupled and moved together at low frequencies producing a dramatic reduction in the speed of sound. Bob Bullock wrote a paper for the AES where he tried to correlate the Bradbury math against a TL that he had built and measured, he could not get things to work. I replicated his work and got the same result. Augspurger also wrote two articles for the AES and came up with the same conclusions, fibers did not slow the speed of sound.

Eventually I built a test TL and measured the electrical impedance with and without fiber stuffing. I used polyester and long fiber wool at three different packing densities. Again, Bradbury's math model did not work. I saw no evidence of a slowing of the speed of sound or a phase inversion. All of these results are available for review on my website, my method has been around for about 15 years and nobody has sent me any data that would invalidate the results and conclusions. Augspurger's AES articles have also been around for about the same amount of time and to the best of my knowledge nobody has found fault with his methods or results.

If you have your own set of models based on Bailey's and Bradbury's work and then testing that correlates with the calculations I would love to see them presented. In particular, electrical impedance and SPL output from the driver and the open end of the TL with and without fiber stuffing. If you have that kind of data it would be great and we can discuss the results in detail. If you do not have the data then continued discussion is kind of pointless and we should just agree to disagree.

No, I have no data of my own. I am an engineer in flow dynamics and thermodynamics. I recognize a good experiment and good analysis when I see one. The acoustics in a fiber filled tube falls into my area of interest. And Bailey and Bradbury's work are done on sound engineering basis to extend the application for domestic audio. (If you want to see my technical publications, google "kei lau boeing". They are not in audio.)

I have asked you a simple question about how the dimensions (diameter and length) of the fiber that you used compared to the spec called out by Bailey and Bradley, it was never answered. When you mentioned polyester fiber, it is not even remotely related to the researches on fiber filled tube in the 40. 50 and 60's. Why did you do that?

The difference between Bailey's transmission line enclosure and your vented boxes are the problem of quality vs. quantity. A vented boxed can get more extended base on the same volume, but the driver is not optimally damped to minimize overshoot. It is a principle of physics on a resonant box. Thiele and Small wrote a series of good papers to help us understand their pros and cons of different damping. And later extended that to closed box design too.

For those interested, Bailey 2 papers and one letter to the editor used to be found at jordan-usa.com. (http://documents.jordan-usa.com/Famous-Articles/) But the web site seemed to have been removed. They are not available from the Wireless World web page. The Bradbury paper is copyrighted and available for download from the Audio Engineering Society for a fee.

http://www.aes.org/e-lib/browse.cfm?elib=2632
 
No, I have no data of my own. I am an engineer in flow dynamics and thermodynamics. I recognize a good experiment and good analysis when I see one. The acoustics in a fiber filled tube falls into my area of interest. And Bailey and Bradbury's work are done on sound engineering basis to extend the application for domestic audio. (If you want to see my technical publications, google "kei lau boeing". They are not in audio.)

I am also an engineer with over 35 years of experience in dynamics and vibrations. Bailey and Bradbury wrote good papers and they look very credible when you read them and follow the math. The problem was that the results do not make any sense when you look closely and ask critical question. For example, Bradbury coupled the air and the fibers so that at low frequency the fibers moved adding an additional mass term to the wave equations, but no stiffness or damping which seems like a significant ommision. If you compress a volume of fibers and release then they slowly expand back to the original volume, that is not a mass driven process. Why no results for the empty TL to be able to separate the impact of the fibers from a baseline case? The results from Bailey and Bradbury can not be used to design a TL speaker.

I have asked you a simple question about how the dimensions (diameter and length) of the fiber that you used compared to the spec called out by Bailey and Bradley, it was never answered. When you mentioned polyester fiber, it is not even remotely related to the researches on fiber filled tube in the 40. 50 and 60's. Why did you do that?

Over 15 years ago I tested several types of fiber, one being polyester and several different samples of wool. The results were similar and there was no reduction in the speed of sound. I stuck with polyester because it is cheap, uniform, and easy to obtain. It is probably not as good at damping waves as wool or fiber-glass but on the up side it does not smell or itch. It works and correlates great in my TL designs.

The difference between Bailey's transmission line enclosure and your vented boxes are the problem of quality vs. quantity. A vented boxed can get more extended base on the same volume, but the driver is not optimally damped to minimize overshoot. It is a principle of physics on a resonant box. Thiele and Small wrote a series of good papers to help us understand their pros and cons of different damping. And later extended that to closed box design too.

I have built several classic TLs and the calculations have matched the measured results closely. They are no longer on my website because the drivers are not available any more.

I suggest you set up your own test speaker and do the design calculations to see how well the predictions match the measurements. Without doing that you are just talking about a topic when you have not done the necessary background work. Proof is in the doing. I will be happy to answer questions but I am not going to debate the topic with you any more until you have done the experiments, you are not prepared for that depth of discussion.
 
The Bailey designs were almost completely stuffed with wool and did not allow a free flow of air to line terminus.
There many articles in Speaker Builder magazine that stated this over the years before the publication ceased.

The I.M. Fried and R.Wright designs were advertised as using free flow resistive acoustic filters. The use of low density polyurethane open cell foam accomplished this..
The purpose of the foam was to dissipate the driver back wave and to dampen line resonances.

I know from building a few of the Fried Products 1/4W T-line designs that there is a free flow of air to line terminus in these type of designs.

I wonder what T-line designs Keilau one of our new forum members has heard?

BTW welcome to AK Keilau. :)
 
The Bailey designs were almost completely stuffed with wool and did not allow a free flow of air to line terminus.
There many articles in Speaker Builder magazine that stated this over the years before the publication ceased.

I wonder what T-line designs Keilau one of our new forum members has heard?

BTW welcome to AK Keilau. :)
I built the original Bailey design more than 30 years ago following the recommendation by Jastak in the Audio Amateur (1973, 1:3). I used the Philips 10100 10" woofer. The HF section was two Janszen 2-panel electrostatic, crossover at 1K Hz. An execellent system for my small apartment room and very affordable for a graduate student. It replaced the Larger Advent's that I had before them. The cabinet went into storage when the Philips driver deteriorated due to age and I could not find a suitable replacement.

You described it correctly, "The Bailey designs were almost completely stuffed with wool and did not allow a free flow of air to line terminus". And the non-linear effect of flow resistance (frequency dependent) threw a lot of people off balance. The mathematic is only based on linear algebra which most upper class undergraduate students should have taken. Prof. Bradbury at University of Surrey (UK) published an elegant paper on this topic and proved the foresight of Dr. Bailey.

The Bailey design requires 1/2 pound per cubic foot of long fiber wool to provide the resistance to the air flow. The fiber must be thick enough to create the drag coefficient to attenuate the sound wave at higher frequencies and slows down the air speed to about half at the low end. That effectively makes a 6 feet transimission line enclosure a half wave tube at the 25 Hz, making the TL openning a phase inverter. Bradbury measured the wool fiber diameter at 0.028 mm, and showed both analytically and experimentally that it works exactly as Bailey designed. Bradbury also investigated the fiber glass which had a diameter of 0.005 mm, or about 1/6 of the wool fiber's. The fiber glass specific gravity is also only 1/4 of the wool used. It demonstrated that low specific gravity and thin fiber does not work the same and thus not suitable for transmission line enclosure. I know that people who treats the TL as a vented box and did not duplicate the Bradbury experiment would not be convinced.

I did not replicate the Bradbury experiment, but I was able to replicate his equation derivation (found a typo) and recreate his calculation. I found absolutely no basis to question the integraty of both Bradbury's analytical and experimental work. I just do not understand how anyone can claim that the Bradbury model does not work.

I wish someday, someone will pick up and extend Bradbury's work to combine it with a woofer driver model. It will take some hardwork, but is entirely possible.
 
I wish someday, someone will pick up and extend Bradbury's work to combine it with a woofer driver model. It will take some hardwork, but is entirely possible.

Several people have done exactly what you describe, myself included programming in BASIC on an old PC about 25 years ago and more recently over the past 15 years in MathCad. In every case I am aware of the Bradbury methods did not work, there was no significant reduction in the speed of sound. There is a paper in the AES by Bullock that documents his efforts if you do not believe my work.

If you are an engineer, then you should be able to run the experiments at home and do the math on a PC to prove to yourself if the methods work or not. It is not that hard for somebody with a decent math background and some free test software.
 
Several people have done exactly what you describe, myself included programming in BASIC on an old PC about 25 years ago and more recently over the past 15 years in MathCad. In every case I am aware of the Bradbury methods did not work, there was no significant reduction in the speed of sound. There is a paper in the AES by Bullock that documents his efforts if you do not believe my work.

If you are an engineer, then you should be able to run the experiments at home and do the math on a PC to prove to yourself if the methods work or not. It is not that hard for somebody with a decent math background and some free test software.

Bradbury showed excellent comparison between his test results and analytical calculations. I am talking about Figures 5 and 6 on page 166 of Journal of The Audio Engineering Society, Vol 24, #3, April 1976. Yes, the comparison were NOT perfect. The experimental data points did not lay on top of the line. But in engineering, these are execellent comparison in trend and in magnitude.

You keep bad mouthing the work by Bradbury. What problem did you find with his test methodology? You stated that you "rederived all of the equations and replicated the numerical results". Can you point out where the typo is in the Bradbury paper? Did you calculation showed same result as Bradbury once you corrected the typo? You do not need a PC to do this. It can be done on a piece of paper. You do need to have some working knowledge of linear algebra and complex number.

I never said that extending Bradbury's work is easy. It is not a question of programing. The challenge is in integrating the equations for the non-linear fiber filled tube with a linear speaker driver model. I do not see that you understand the nature of this problem.

I am defending the work by Bailey and Bradbury with passion because I am an engineer. Audio is my hobby. I do not have the resources to repeat the Bradbury experiment at home. I did follow the mathematics of his model for the fun of it and found it a excellent piece of work. I do not have the vested interest of a revenue generating website. But I am passionate about the integrity of engineering.

Yes, I am a bonafide engineer. You can google "kei lau boeing" if you want to know what I do for a day job.
 
Last edited:
MJKing, a follow-on question: How did you measure the speed of sound in a fiber filled tube when you did your experiment? I know that it was a long time ago, but you do remember, right?
 
...But if you spend the time to find the right kind of "long fiber wool" and build a Bailey non-resonant speaker enclosure, you will be richly rewarded. Yes, I am a transimission line speaker fanboy...

Well that makes two of us!!!

Seriously gentlemen, I have really appreciated both of your passionate positions and it is, as someone said, pretty special when we have both Ken and Martin making their wealth of knowledge known.

I do not have this kind of background, but I do have a pair of TDL Studio-4's that I briefly yanked the tweeter out of. And if the type of stuffing in that area of the enclosure is anything like what is in the Bass Enclosure of the TL, then I think Keilau has a point! :yes:

And I'd also have to say John Wright probably had a pretty good handle on what Bailey was trying to get across! :thmbsp:

IMG_0600.JPG
 
Last edited:
I found this article from PMC interesting. It gets into the absorption material issues to a certain degree!

What is really interesting from my perspective is this comment:

"There are therefore two distinct forms of bass loading employed in a TL, which historically and
confusingly have been amalgamated in the TL description. Separating the upper and lower bass
analysis reveals why the TL has so many advantages over reflex and infinite baffle designs. The upper
bass is completely absorbed by the line allowing a clean and neutral response. The lower bass is
extended effortlessly and distortion is lowered by the line’s control over the drive unit’s excursion.
One of the exclusive benefits of the TL design is its ability to produce very low frequencies even at
low monitoring levels."


I have heard this low frequency even at low monitoring levels! And I believe that's what us TL Heads call... "the magic!" :thmbsp:
 
Last edited:
Bradbury showed excellent comparison between his test results and analytical calculations. I am talking about Figures 5 and 6 on page 166 of Journal of The Audio Engineering Society, Vol 24, #3, April 1976. Yes, the comparison were NOT perfect. The experimental data points did not lay on top of the line. But in engineering, these are execellent comparison in trend and in magnitude.

You keep bad mouthing the work by Bradbury. What problem did you find with his test methodology? You stated that you "rederived all of the equations and replicated the numerical results". Can you point out where the typo is in the Bradbury paper? Did you calculation showed same result as Bradbury once you corrected the typo? You do not need a PC to do this. It can be done on a piece of paper. You do need to have some working knowledge of linear algebra and complex number.

I never said that extending Bradbury's work is easy. It is not a question of programing. The challenge is in integrating the equations for the non-linear fiber filled tube with a linear speaker driver model. I do not see that you understand the nature of this problem.

I am defending the work by Bailey and Bradbury with passion because I am an engineer. Audio is my hobby. I do not have the resources to repeat the Bradbury experiment at home. I did follow the mathematics of his model for the fun of it and found it a excellent piece of work. I do not have the vested interest of a revenue generating website. But I am passionate about the integrity of engineering.

Yes, I am a bonafide engineer. You can google "kei lau boeing" if you want to know what I do for a day job.

Why do you keep repeating, over and over again, that you are an engineer? You already said that. Several of us around here are engineers. Definitely including Mr. King.

-k
 
MJKing, a follow-on question: How did you measure the speed of sound in a fiber filled tube when you did your experiment? I know that it was a long time ago, but you do remember, right?

My work on TL's, measurements and theory, along with my opinion on where Bailey ad Bradbury made a findumental mistake can be found on my website www.quarter-wave.com under the TL theory link. If you read the articles I wrote you can then decide which theory makes more sense to you and follow that path. My method and models match the ones George Augspurger published in the JAES at about the same time, coincidentily we worked the problem in parallel with two completely different modeling methods and got identical results.
 
If you are an engineer, then you should be able to run the experiments at home and do the math on a PC to prove to yourself if the methods work or not. It is not that hard for somebody with a decent math background and some free test software.
Why do you keep repeating, over and over again, that you are an engineer? You already said that. Several of us around here are engineers. Definitely including Mr. King.

-k

I said that in response to MJKing's question. The key to the Bradbury model is the derivation of the drag parameters of various kinds of fiber materials. Bradbury conducted the proper experiment to validate his model. MJKing may be an engineer, but he demonstrated that he does not understand the fluid dynamics of air moving in a fiber filled tube. He kept saying that the Bradbury model does not work, and sound speed was not reduced in long fiber wool filled tube. But he never showed evidence that he conducted sound speed measurements, nor show proof that Bradbury did not derive the right model to calculate drag parameters.

All these work by Bailey and Bradbury showed that why ONLY the proper dimensioned long fiber wool provides the right kind of acoustic loading. But mathematical knowledge of linear algebra, complex number etc. is not necessary to design good loudspeakers. But trying to make a simple filter model to descirbe the transmission line enclosure like MJKing did just does not work. I suspect that people built loudspeaker enclosures based on MJKing method will never enjoy the deep and tight bass that the Bailey design can provide. But they should get the bass of a vented, large sized enclosure.

BTW, I am currently using a pair of NHT Super One as the front speakers and argumented by a pair of Wharfedale for the low end. Wharfedale called their bi-amp Monitor M-138 model "transmission line" too, but they are not. It was an aperiodic loaded enclosure with twin 7" woofers. The bass does not go as deep, nor as tight as the Bailey TL. But they are better than many commercial subwoofer boxes on the market. I am an audio hobbyist who likes to experiment with DIY equipments, but my engineer nature sometimes gets in the way. (Sorry that I said I was an engineer again.)

Ken, I respect you because of the sound of the Super One that I hear. It is NOT easy to do it right and keep the cost down.
 
Last edited:
Back
Top Bottom