Hi All,
I've just started to look into TL's as a possible mid-bass configuration and I had a few questions since you can't use "TL" as a search criteria.
Is the responce of a driver in a TL box simular to that of an Infinite Baffle?
How do the T/S parameters differ from the Closed Box or Ported Box alignments?
Is the transient response generally better or worse then a Closed Box or Ported Box?
Thanks for your assistance
Rodd Yamashita
I've just started to look into TL's as a possible mid-bass configuration and I had a few questions since you can't use "TL" as a search criteria.
Is the responce of a driver in a TL box simular to that of an Infinite Baffle?
How do the T/S parameters differ from the Closed Box or Ported Box alignments?
Is the transient response generally better or worse then a Closed Box or Ported Box?
Thanks for your assistance
Rodd Yamashita
Rodd,
I am currently working a little TL for the Tangband W3-879 which will be used as a mid-tweeter.
A TL can be tuned to roll-off anywhere from 2nd order to 4th order and the phase, transient reponse, bass augmentation, etc will all be affected by your choices hen designing the line. (pretty open-ended answer ).
For a midrange thou you usually aren't trying for bass extension but to gain the advantages of flattening the resonant peak, and minimizing any time smear/reflections from coming back thru the cone (what i refer to as sucking the back-wave down the pipe). In this case you will no doubt what to push the TL towards the aperiodic end of the spectrum which results in FR similar to a sealed box (perhaps a bit better), with a flattened impedance curve, and excellent transient response.
dave
I am currently working a little TL for the Tangband W3-879 which will be used as a mid-tweeter.
A TL can be tuned to roll-off anywhere from 2nd order to 4th order and the phase, transient reponse, bass augmentation, etc will all be affected by your choices hen designing the line. (pretty open-ended answer
).For a midrange thou you usually aren't trying for bass extension but to gain the advantages of flattening the resonant peak, and minimizing any time smear/reflections from coming back thru the cone (what i refer to as sucking the back-wave down the pipe). In this case you will no doubt what to push the TL towards the aperiodic end of the spectrum which results in FR similar to a sealed box (perhaps a bit better), with a flattened impedance curve, and excellent transient response.
dave
Exactly. More specifically, I am looking for a way to extend the usable low end of drivers like the TAD 1101H or the JBL 2123H that are made to xover to the woofer at >200Hz. I want to be able to use these drivers down to 70 or 80Hz if possible and xover to a sub-bass.planet10 said:
What type of TL should I use? How long is long enough? Is there a way of predicting the results of a design? After some searching this evening, I noticed that there isn't much quantitative information available beyond wavelength calculations.
Thanks,
Rodd Yamashita
roddyama said:
A TL won't give you an octave & a bit -- maybe a half octave, and then only if you tune for bass efficiency.
The best available quantitative tool is Martin King's quarterwave modeling software. (Augspurger's is very close but just not very available)
It started out on my site http://www.t-linespeakers.org/projects/martin/index.html, and now has his own site http://www.quarter-wave.com/. There is also some excellent studies done by Bob Brines.
As well, but not out yet, are some alignments being generated by Rick Shultz (exolinear). Rick has been exploring the huge space of quarter-wave designs that Augspurger & King's models have opened up for exploration and what he is doing (my interpretation) is to set some strict parameters (ie line taper, terminus size, driver placement) and for this subset of quarter-wave space, come up with a brain dead easy way of plugging the T/S parameters in and getting a quarter-wave line out. alphaTL is the 1st (& should be published soon in aXp), i know he is working on more.
dave
Hi,
One of my first big projects was the Rogers Monitor. It produced a lot of dry bas, but I never got it right and it ended in a disappointment for me. Later on the closed box with a Bessel alignment was my favourite, easy to build and a to tune. But in recent years I have heard well designed BR’s that gave cleaner bas response than a CB. To my idea the non-linear spring action of the enclosed air of a CB and the large excursion of the LS unit at lo frequencies are responsible for a load of distortion, harmonic as well as IM.
I am up for building a new set of speakers in the near future. After all those years of DIY speaker building I am still puzzled what topology to use. A TML has still a lot of attraction, because of the low pressures involved in the cabinet, smooth roll-off and low distortion. But I don’t want to go into a long trajectory of “Trial and Error”.
For a TML, I have an eye on the Audax HM170 or the HM210, or maybe one of the top 8” ones from Focal. I am planning a 3-way system with a x-over at app. 125Hz on the low end.
I have ploughed completely trough the articles on http://www.quarter-wave.com/, Good reading, but I have an uneasy feeling this mathematical way is not the way to go. Baily’s original articles are in itself very straightforward and I have the idea that there is no real progress in a methodology of designing a TML, except the work of Augspurger.
Chaio,
Peter
One of my first big projects was the Rogers Monitor. It produced a lot of dry bas, but I never got it right and it ended in a disappointment for me. Later on the closed box with a Bessel alignment was my favourite, easy to build and a to tune. But in recent years I have heard well designed BR’s that gave cleaner bas response than a CB. To my idea the non-linear spring action of the enclosed air of a CB and the large excursion of the LS unit at lo frequencies are responsible for a load of distortion, harmonic as well as IM.
I am up for building a new set of speakers in the near future. After all those years of DIY speaker building I am still puzzled what topology to use. A TML has still a lot of attraction, because of the low pressures involved in the cabinet, smooth roll-off and low distortion. But I don’t want to go into a long trajectory of “Trial and Error”.
For a TML, I have an eye on the Audax HM170 or the HM210, or maybe one of the top 8” ones from Focal. I am planning a 3-way system with a x-over at app. 125Hz on the low end.
I have ploughed completely trough the articles on http://www.quarter-wave.com/, Good reading, but I have an uneasy feeling this mathematical way is not the way to go. Baily’s original articles are in itself very straightforward and I have the idea that there is no real progress in a methodology of designing a TML, except the work of Augspurger.
Is there some more reading about this on the internet Dave?
Chaio,
Peter
Pjotr said:
Well given that Augspurger's work and Martin King's work are equivalent to each other i don't know why you would say this. Martin's is a mechanical analog and George's is an electrical analog, but they both generate the same lines from the same data (Rick & Martin have confirmed this). Martin's model doesn't have as much limitation at the higher frequencies as George's.
So if you want to build a TL with confidence, download Martin's software & play. Augspurger's papers are a required read (3 part article in Speaker Builder or the revised version in the AES Journal) because they add some context to Martin's articles.
The bulk is at the 3 sites i mentioned.
dave
Thanx Dave,
Well, I just started investigating building a TML a few months ago. Your site and Bob Brines site looks nice as a good start. I have no recent fundamental experience at all with TML’s. It is new adventure.
The reason that I am reserved about doing simulations on such a complex system as a TML is the many variables that are unknown numerically. Like the exact properties of the stuffing and the number of bends and the way the bends are implemented. Simulations never give you an idea how the final thing will sound except you have already a lot of experience with the subject. But you are right, tasting the pudding is in the eating.
On the other hand the rules look simple, the length of the line should be ¼ wave acoustically at the tuning frequency and preferably with no bends at all. To taper or not to taper seems a design choice to me. I will dig up again the JAES articles of Augspurger in the library this week.
Chaio
Well, I just started investigating building a TML a few months ago. Your site and Bob Brines site looks nice as a good start. I have no recent fundamental experience at all with TML’s. It is new adventure.
The reason that I am reserved about doing simulations on such a complex system as a TML is the many variables that are unknown numerically. Like the exact properties of the stuffing and the number of bends and the way the bends are implemented. Simulations never give you an idea how the final thing will sound except you have already a lot of experience with the subject. But you are right, tasting the pudding is in the eating.
On the other hand the rules look simple, the length of the line should be ¼ wave acoustically at the tuning frequency and preferably with no bends at all. To taper or not to taper seems a design choice to me. I will dig up again the JAES articles of Augspurger in the library this week.
Chaio
Thanks for the replies navin. I will probably have other questions for you regarding aperiodic enclosures, but I'll save those for a new thread. I'm leaning towards a aperiodic or no-terminus-output TL.
By the way I'm proposing a new, less cumbersome name for the latter: NOTL or No Output Transmission Line. Unless there is already a name for them that I don't know of.
By the way I'm proposing a new, less cumbersome name for the latter: NOTL or No Output Transmission Line. Unless there is already a name for them that I don't know of.
Just a couple of quick comments on what is wrtten above. Most of the info presented is excellent but I do have one or two nits :
George Augsburger's model breaks the TL into approximatly 800 (if I remember correctly) little discrete elements of mass, stiffness, and damping using electrical circuit elements. This type of model was first documented by Bert Locanthi in an AES paper many years ago which can be found in the JAES Anthology. This discrete style of model is a numerical approximation that is accurate at the lower frequencies and becomes less accurate at the higher frequencies. It is a numerical solution of the 1D wave equation.
My model uses a closed form solution of the 1D wave equation that is accurate over all of the frequencies. So you can model a TL using one section and get valid results at every frequency calculated by the program. Also my model applies a frequency dependent B.C. at the open end which does a much better job of modeling the acoustic damping provided by this opening at the higher frequencies. The same solution of the wave equation is used to model horns but without the fiber damping term.
Rick Schultz and I have run several similar problems with both models and achieved very good correlation. This is an excellent double check for the two independently formulated computer models. Both models start to break down when the enclosure dimensions start to generate standing half waves that occur across the area of the line. This typically is not seen until you reach higher frequencies.
My model is based on one type of fiber material and George Augsburger has included several types of fiber materials. I need to expand my fiber data to provide more options.
Last point, be careful with the standard rule of thumb equation for calculating the first tuning frequency. The formula
f = c / (4 x L)
is only valid for straight TLs. If you add a taper then the frequency will be lower, if you add an expansion like a TQWT then the frequency will be higher. Any kind of nonunifor section in the line will also impact this calculation.
I believe that both of the computer models are very accurate if applied correctly. I have made mine available for people to use, but everybody is free to use whatever method suits their own needs and capabilities.
OK, my very sick (terminal I am afraid) computer is about at the end of its up cycle so I have to power down. Everyday has been an adventure with this computer, the new one is on its way. 8^)
Good luck,
Martin
www.quarter-wave.com
George Augsburger's model breaks the TL into approximatly 800 (if I remember correctly) little discrete elements of mass, stiffness, and damping using electrical circuit elements. This type of model was first documented by Bert Locanthi in an AES paper many years ago which can be found in the JAES Anthology. This discrete style of model is a numerical approximation that is accurate at the lower frequencies and becomes less accurate at the higher frequencies. It is a numerical solution of the 1D wave equation.
My model uses a closed form solution of the 1D wave equation that is accurate over all of the frequencies. So you can model a TL using one section and get valid results at every frequency calculated by the program. Also my model applies a frequency dependent B.C. at the open end which does a much better job of modeling the acoustic damping provided by this opening at the higher frequencies. The same solution of the wave equation is used to model horns but without the fiber damping term.
Rick Schultz and I have run several similar problems with both models and achieved very good correlation. This is an excellent double check for the two independently formulated computer models. Both models start to break down when the enclosure dimensions start to generate standing half waves that occur across the area of the line. This typically is not seen until you reach higher frequencies.
My model is based on one type of fiber material and George Augsburger has included several types of fiber materials. I need to expand my fiber data to provide more options.
Last point, be careful with the standard rule of thumb equation for calculating the first tuning frequency. The formula
f = c / (4 x L)
is only valid for straight TLs. If you add a taper then the frequency will be lower, if you add an expansion like a TQWT then the frequency will be higher. Any kind of nonunifor section in the line will also impact this calculation.
I believe that both of the computer models are very accurate if applied correctly. I have made mine available for people to use, but everybody is free to use whatever method suits their own needs and capabilities.
OK, my very sick (terminal I am afraid) computer is about at the end of its up cycle so I have to power down. Everyday has been an adventure with this computer, the new one is on its way. 8^)
Good luck,
Martin
www.quarter-wave.com
Dave, Martin,
Clearly, my first hurried look at Dave's site was insufficient to see what has been done so far on the understanding of TL’s. I must thank you and the others referred to on the http://www.t-linespeakers.org/ site for providing this information. I guess it is now up to me to determine if I am willing to invest the time and effort to understand more fully what your web sites have to offer.
My goal is to make a smooth transition from a cone type direct radiator mid-bass speaker to a horn mid-range speaker. One of the aspects of this transition is the dynamic character of the sound of a horn speaker compared to that of the cone speaker. The transmission line may or may not be the solution for me, but I will at least try to put the effort in to decide one way or the other.
Thanks again,
Rodd Yamashita
Clearly, my first hurried look at Dave's site was insufficient to see what has been done so far on the understanding of TL’s. I must thank you and the others referred to on the http://www.t-linespeakers.org/ site for providing this information. I guess it is now up to me to determine if I am willing to invest the time and effort to understand more fully what your web sites have to offer.
My goal is to make a smooth transition from a cone type direct radiator mid-bass speaker to a horn mid-range speaker. One of the aspects of this transition is the dynamic character of the sound of a horn speaker compared to that of the cone speaker. The transmission line may or may not be the solution for me, but I will at least try to put the effort in to decide one way or the other.
Thanks again,
Rodd Yamashita
navin said:
yep, just sucks up the entire back wave and doesn't let it get out.
dave
An externally hosted image should be here but it was not working when we last tested it.
Hi Rod,
You can easily try an aperiodic loading, and I think this may work best for you to get maximum dynamics.
Simply mount your driver in a tube approx 1/3 to 1/2 wavelength of your lowest needed frequency, then heavily stuff the rear. This gives you no compression loading of the driver, but absorbs most of the rearward radiation.
You might also try open baffle, this could work well for you as well
You can easily try an aperiodic loading, and I think this may work best for you to get maximum dynamics.
Simply mount your driver in a tube approx 1/3 to 1/2 wavelength of your lowest needed frequency, then heavily stuff the rear. This gives you no compression loading of the driver, but absorbs most of the rearward radiation.
You might also try open baffle, this could work well for you as well
There are a couple of articles by a fellow named Jon Risch that explains the difference between a Closed Box and a closed end Transmission Line. Amazingly, they don't seem to be available at his website. Or at least, I couldn't find them. But I have read them.
LOL, Roddy seems to be one of the few Americans I have ever seen use the term Infinite Baffle. For some reason you don't see the term much over here, just like we don't use the term "total enclosure" for Closed or Sealed Box.
Just for the record, an Infinite Baffle, (IB), means a Closed Box with a volume roughly equal to or greater than the speaker's Vas. And Acoustic Suspension, (or Air Suspension), is a Closed Box with a volume somewhat less than the speaker's Vas.
LOL, Roddy seems to be one of the few Americans I have ever seen use the term Infinite Baffle. For some reason you don't see the term much over here, just like we don't use the term "total enclosure" for Closed or Sealed Box.
Just for the record, an Infinite Baffle, (IB), means a Closed Box with a volume roughly equal to or greater than the speaker's Vas. And Acoustic Suspension, (or Air Suspension), is a Closed Box with a volume somewhat less than the speaker's Vas.
Found 'em. Either both were posted at Planet 10's T-Line website or one was at his site and one was posted here. Not sure which. Anyway, here they are:
http://www.t-linespeakers.org/design/classic.html
http://www.audioasylum.com/audio/speakers/messages/11476.html
There is also a detailed article by Jopn Risch somewhere where he builds one with a specific driver. Has graphs, etc. I'll try to dig that one up.
http://www.t-linespeakers.org/design/classic.html
http://www.audioasylum.com/audio/speakers/messages/11476.html
There is also a detailed article by Jopn Risch somewhere where he builds one with a specific driver. Has graphs, etc. I'll try to dig that one up.
OK, so as I see it there are three classes of TL with variations in between them. Just so we are all talking about the same things, this is a preliminary description:
1) True or classic transmission line: small amount of in-phase terminus output, described by Jon Risch in the links Kelticwizard posted.
2) Bass-reflex or augmented transmission line: greater amount of terminus output, I believe this is the kind most commonly seen today.
3) Labyrinth transmission line: ideally no output from terminus, line usually made as long as possible and really stuffed. Has two categories:
a) Closed-end like the Nautilus, etc. Usually little stuffing?
b) With terminus, what I was describing earlier as the No Output Transmission Line. Tapering the line towards the terminus is believed to aid in reduction of terminus output.
P.S. Are we wandering off the original post topic?
P.P.S Yes Navin, it is -13 C here right now, with 3 inches of snow on the ground.
1) True or classic transmission line: small amount of in-phase terminus output, described by Jon Risch in the links Kelticwizard posted.
2) Bass-reflex or augmented transmission line: greater amount of terminus output, I believe this is the kind most commonly seen today.
3) Labyrinth transmission line: ideally no output from terminus, line usually made as long as possible and really stuffed. Has two categories:
a) Closed-end like the Nautilus, etc. Usually little stuffing?
b) With terminus, what I was describing earlier as the No Output Transmission Line. Tapering the line towards the terminus is believed to aid in reduction of terminus output.
P.S. Are we wandering off the original post topic?
P.P.S Yes Navin, it is -13 C here right now, with 3 inches of snow on the ground.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.