Hi friends!
I want to build a pair of speakers and have my eyes set on the seas Thor transmission line speaker designed by dr d'Appolito himself ( http://www.seas.no/thor.htm ). The reason for my interest is the many good reviews their previous top-of-the-line DIY kit Odin got. ( i never liked the look of Odin and dont want a stand mount speaker. ) So when they launched a better design ( in their view ) that looked better i got interested.
Im a bit put of by it beeing a transmission line speaker. If it was a sealed box i would probably build it immediatly. Do you have any experiance with TL speakers?
I listen to most kinds of music, but dont need earth moving bass energy. I like the bass to be fast and tuneful. ( mind it, i dont MIND deep bass if the music is recorded like that. Its just not my top criteria for choosing a speaker )
I would much appreciate your view if you have experiance of your own.
I want to build a pair of speakers and have my eyes set on the seas Thor transmission line speaker designed by dr d'Appolito himself ( http://www.seas.no/thor.htm ). The reason for my interest is the many good reviews their previous top-of-the-line DIY kit Odin got. ( i never liked the look of Odin and dont want a stand mount speaker. ) So when they launched a better design ( in their view ) that looked better i got interested.
Im a bit put of by it beeing a transmission line speaker. If it was a sealed box i would probably build it immediatly. Do you have any experiance with TL speakers?
I listen to most kinds of music, but dont need earth moving bass energy. I like the bass to be fast and tuneful. ( mind it, i dont MIND deep bass if the music is recorded like that. Its just not my top criteria for choosing a speaker )
I would much appreciate your view if you have experiance of your own.
Transmission lines have a high cool factor in my book. There are two reasons I'm not currently running transmission lines--the size would be prohibitive in my case, and they're not a good choice for servo feedback loops, which happens to be something I want to play with for a bit.
I've used them off and on for years. Properly designed, they are a wonder.
Grey
I've used them off and on for years. Properly designed, they are a wonder.
Grey
Robert
Having built and used transmission lines at various times during the past 30 years, my advice is don't be put off. Properly designed, transmission lines offer the best subjective bass response of any type of enclosure. Tight, fast and deep. The only time I have not used a transmission line is when the available space has prohibited it.
Geoff
Having built and used transmission lines at various times during the past 30 years, my advice is don't be put off. Properly designed, transmission lines offer the best subjective bass response of any type of enclosure. Tight, fast and deep. The only time I have not used a transmission line is when the available space has prohibited it.
Geoff
If you hear a properly designed transmission line you will never go back to sealed box again. A TL has effectively NO RESONANCE! The bass is incredibly tight, and the woofer is finally given room to move. The bass will not be earth-shattering, it will be clean. I built a pair of transmission lines to go with my electrostatics. I will never own another sealed box.
I have published a preliminary report on a TL on my ESL website. You may find the construction pics and plans informative. Click on the "Hybrid ESL" tab.
http://home1.gte.net/res0f2t3/index.htm
Good Luck!
-M.A.
I have published a preliminary report on a TL on my ESL website. You may find the construction pics and plans informative. Click on the "Hybrid ESL" tab.
http://home1.gte.net/res0f2t3/index.htm
Good Luck!
-M.A.
M_Anker:
You're not quite right about TL's having 'NO RESONANCE'.
The whole principle of a TL is based on using one resonancefrequency, damped to a low Q, to increase bandwith. That's why the pipe is getting smaller, the further from the speaker You get.
But it does sound resonance-free, properly constructed.
Some years ago (@12) I did some research on damping materials for TL's, and came to the conclusion that 'long-haired sheep-wool' was best suited, since it had 6dB less attenuation at 30Hz than the best synthetic alternative, but had the same attenuation at higher frequencies.
I also found out that if I stuffed the 1/3 of the line closest to the speaker tight, the next 1/3 less tight, and the last 1/3 loosly, I'd get not only the best attenuation of undesired frequencies, but also the best sound.
Bedroc:
Go for it.
Put some effort into experimenting with the damping-material. It has a big influence on the final result.
Have a look at : http://www.t-linespeakers.org/index.html
I love TL's too.
You're not quite right about TL's having 'NO RESONANCE'.
The whole principle of a TL is based on using one resonancefrequency, damped to a low Q, to increase bandwith. That's why the pipe is getting smaller, the further from the speaker You get.
But it does sound resonance-free, properly constructed.
Some years ago (@12) I did some research on damping materials for TL's, and came to the conclusion that 'long-haired sheep-wool' was best suited, since it had 6dB less attenuation at 30Hz than the best synthetic alternative, but had the same attenuation at higher frequencies.
I also found out that if I stuffed the 1/3 of the line closest to the speaker tight, the next 1/3 less tight, and the last 1/3 loosly, I'd get not only the best attenuation of undesired frequencies, but also the best sound.
Bedroc:
Go for it.
Put some effort into experimenting with the damping-material. It has a big influence on the final result.
Have a look at : http://www.t-linespeakers.org/index.html
I love TL's too.
width of the line
It seems to be that there is no physicsal reason for the line to get
tighter at the output.
Years ago i buildt one with a linear cross-section and there was no differnce in frequency response.
In fact the resonant frequency is determined by the length of the tube.
The rest you'll get by proper damping.
regards
Arne
It seems to be that there is no physicsal reason for the line to get
tighter at the output.
Years ago i buildt one with a linear cross-section and there was no differnce in frequency response.
In fact the resonant frequency is determined by the length of the tube.
The rest you'll get by proper damping.
regards
Arne
The reason the TL appears to have no resonance is because of the tapered-pipe configuration. There are only two parallel surfaces(the sides), and the rest of the pipe is not parallel. This effectively removes any chance of standing waves; therefore, the TL has effectively no unwanted resonance. Hoffmeyer is correct, but what I am saying is that you will not experience any unwanted resonance coming from the enclosure. I am comparing the TL design with sealed box enclosures in which I've always had unwanted resonances.
If you taper both sides, or use a design that subdivides a rectangular cross section into triangular pipes, you can remove even that potential resonance.
One thing I've always wanted to try is to take a PVC pipe (yes, Bryan, Sonotube would work as well, and probably be cheaper), then roll a triangular piece of 1/2" foam into a tube and slide it inside. Since the triangle would taper the inside of the foam roll, you would achieve progressive damping, and due to the circular cross section and foam, there would be no chance for unwanted resonances at all. It would also be ridiculously easy and inexpensive to construct.
I think I'll build one this afternoon. (Just kidding...got heatsinks to fiddle with and amps to build, not to mention going to see my bees.)
Grey
One thing I've always wanted to try is to take a PVC pipe (yes, Bryan, Sonotube would work as well, and probably be cheaper), then roll a triangular piece of 1/2" foam into a tube and slide it inside. Since the triangle would taper the inside of the foam roll, you would achieve progressive damping, and due to the circular cross section and foam, there would be no chance for unwanted resonances at all. It would also be ridiculously easy and inexpensive to construct.
I think I'll build one this afternoon. (Just kidding...got heatsinks to fiddle with and amps to build, not to mention going to see my bees.)
Grey
Alternatives
Hi bedroc,
I have designed and built a couple of TLs over the years. Not bad technique, certainly better than BR. The last were a sub using 4 x 12" drivers each side, each with their own line. They were about 1000 litres each, per side. My friend still uses them.
I am into horns these days, but I have found a couple of items that might interest you, both by people who actually design gear for a living and have a long track record.
It's your choice, but I'm not impressed with the Thor. The LF and HF responses are too rolled off and it's only about 87dB midband efficiency. Can you say compression? Also the two midbasses have substantially different path lengths to the port. Tsk.
Lynn Olsen's Amity would be a better bet. Read the theory and development article (and the others in the library on the same site) to get the details: URL below. It's easily 75-20k FR, 94dB sensitivity, flat impedence, and there are a number of constructor groups around the world that can help too. It uses common, reasonably cheap drivers ( Vifa midbass, Scan tweet) and produces outstanding results, but is probably a bit more difficult to build than the Thor. Tonally accuracy and dynamics are it's strongest points.
http://www.aloha-audio.com/Ariel.html with v6 being the latest and best.
For some general reading on TL principles, take a look at John Risch's posts on AA. He is an engineer and designer for Peavey I beleive. Go to http://www.teresaudio.com/haven/ look at the subwoofer page (designed to match the Ariel: TL design too!) and next to the 4th pic there is a link.
For stuffing the line, carded and well supported long hair wool, or Acousta-stuff is best.
Hope this helps.
Cheers
Brett
Hi bedroc,
I have designed and built a couple of TLs over the years. Not bad technique, certainly better than BR. The last were a sub using 4 x 12" drivers each side, each with their own line. They were about 1000 litres each, per side. My friend still uses them.
I am into horns these days, but I have found a couple of items that might interest you, both by people who actually design gear for a living and have a long track record.
It's your choice, but I'm not impressed with the Thor. The LF and HF responses are too rolled off and it's only about 87dB midband efficiency. Can you say compression? Also the two midbasses have substantially different path lengths to the port. Tsk.
Lynn Olsen's Amity would be a better bet. Read the theory and development article (and the others in the library on the same site) to get the details: URL below. It's easily 75-20k FR, 94dB sensitivity, flat impedence, and there are a number of constructor groups around the world that can help too. It uses common, reasonably cheap drivers ( Vifa midbass, Scan tweet) and produces outstanding results, but is probably a bit more difficult to build than the Thor. Tonally accuracy and dynamics are it's strongest points.
http://www.aloha-audio.com/Ariel.html with v6 being the latest and best.
For some general reading on TL principles, take a look at John Risch's posts on AA. He is an engineer and designer for Peavey I beleive. Go to http://www.teresaudio.com/haven/ look at the subwoofer page (designed to match the Ariel: TL design too!) and next to the 4th pic there is a link.
For stuffing the line, carded and well supported long hair wool, or Acousta-stuff is best.
Hope this helps.
Cheers
Brett
Re: TLs (long)
I'm new to this forum, but this thread has been generating many visits to my website ( http://www.t-linespeakers.org ). I'd like to comment on a few points so far in this thread.
The 1 anecdotal report on the Thor i have heard by the builder was very positive. But the line is very classic looking ... I wonder if a more optimized TL exists.
Up until recently, the design of a TL was very much a "rule-of-thumb" and "cut-and-try" kind of deal. But with the recent works of Martin King and George Augspurger there are now tools which allow the designer to model & predict the performance of a TL before setting saw to wood. Martin's software is freely available on my website, George's is harder to get.
(For all i know Joe could have downloaded the model, and optimized the h%#$ out these drivers).
Martin's software is based on a mechanical analog, and George's on an electrical analog -- both agree very well, and were developed independently. These mark a new era in TL design.
An empty line resonants-- FR of unstuffed line . We want to keep the fundemental and block the rest. The damping & geometry of the line act as a low-pass filter. Augspurger showed a number of geometries that improved this LP.
Some of these are 1) use a 1/3 of the volume of the box as a pre-chamber (like a daline ), 2) move the driver from the end of line, 3) restrict the terminus, 4) taper the line, 5) combinations.
In classic design the taper was generally 1:1 up to 1.5:1. Augspurger suggests 3:1 to 4:1!
Martin's SW can also be used to models Voigt Pipes. Seems to me that the TL & the TQWT are part of a continuum of designs ranging from the steeply tapered TL to the steeply tapered pipe.
dave
Bedroc said:
I'm new to this forum, but this thread has been generating many visits to my website ( http://www.t-linespeakers.org ). I'd like to comment on a few points so far in this thread.
The 1 anecdotal report on the Thor i have heard by the builder was very positive. But the line is very classic looking ... I wonder if a more optimized TL exists.
Up until recently, the design of a TL was very much a "rule-of-thumb" and "cut-and-try" kind of deal. But with the recent works of Martin King and George Augspurger there are now tools which allow the designer to model & predict the performance of a TL before setting saw to wood. Martin's software is freely available on my website, George's is harder to get.
(For all i know Joe could have downloaded the model, and optimized the h%#$ out these drivers).
Martin's software is based on a mechanical analog, and George's on an electrical analog -- both agree very well, and were developed independently. These mark a new era in TL design.
An empty line resonants-- FR of unstuffed line . We want to keep the fundemental and block the rest. The damping & geometry of the line act as a low-pass filter. Augspurger showed a number of geometries that improved this LP.
Some of these are 1) use a 1/3 of the volume of the box as a pre-chamber (like a daline ), 2) move the driver from the end of line, 3) restrict the terminus, 4) taper the line, 5) combinations.
In classic design the taper was generally 1:1 up to 1.5:1. Augspurger suggests 3:1 to 4:1!
Martin's SW can also be used to models Voigt Pipes. Seems to me that the TL & the TQWT are part of a continuum of designs ranging from the steeply tapered TL to the steeply tapered pipe.
dave
Thanks for all the help!!
This really is a marvelous forum! I cant begin to tell you how happy I am with your time and effort. The result beeing im very eager to try a transmission line speaker. ( a bit cautious about the thor after Bretts word of warning. )
Ill post a new question on how to improve the Thor. Would much appreciate some views on that too. ( i kind of like the Thor. the design will go down very well with my spouse and a REALLY like making her happy! )
and again. THANKS!!!
This really is a marvelous forum! I cant begin to tell you how happy I am with your time and effort. The result beeing im very eager to try a transmission line speaker. ( a bit cautious about the thor after Bretts word of warning. )
Ill post a new question on how to improve the Thor. Would much appreciate some views on that too. ( i kind of like the Thor. the design will go down very well with my spouse and a REALLY like making her happy! )
and again. THANKS!!!
Dave,
Great site. It inspired me to try a Tempest TL.
Of course the wife made me take it apart after a month or so.
Pete
Great site. It inspired me to try a Tempest TL.
An externally hosted image should be here but it was not working when we last tested it.
Of course the wife made me take it apart after a month or so.
Pete
TLs are easy to design as you are (I do anyway) using a quarter wave to get the length. So the math is easy. The last ones I built have a 1' squared foot print and are 37" tall. My line was a bit short so I used some extra space as a resonence chamber and built 2 10" passive radiators per cabinet. Very good full range 2 ways that took 2 days to build. I am getting close to 30 Hz for bass and have only a 6 db high pass for the tweeter. I do need to do a bit more work on them but they will stack up against most good mid priced speakers on the market.
Optimised design....
Bedroc,
If you like the Thor and want to build it, go for it. But don't modify the design until you know what you are doing. Dr D'Appolito has a huge amount of experience and knowledge, and I doubt you will improve much on his design, within the parameters set for it.
If you want it to experiment, build one, and use it as a reference. I test almost all my speaker designs and mods in mono, and only when I feel I am getting close to optimum, do I mess with two. If they don't work in mono, they won't in stereo either. TL's are NOT that straightforward to get a good even response with.
For a novice experimenter (my assumption, sorry if I am wrong) build it by the plan, because you know it will work first time, or a construction error can be found more easily. Then live with it for a while, and experiment with placing, cables, amps etc until you have a real feel for the design. Then, build a modified box, x-over etc and compare it to your 'known quantity'. You will learn more, get faster and more consistent results this way. I mean to encourage you, but in a way that you will enjoy it and get useful results. Nothing worse than a garage full of experimental speakers, and you are no wiser as to what works than when you started, or maybe even more confused and dispirited. My voice of experience speaking here.
Dr D'Appolito has a book on testing loudspeakers that I'm sure would be of benefit, as would Vance Dickason's Cookbook. Version 6 has a new section on TL theory. Also try to get hold of a copy of Olsen's Acoustical Engineering.
Please also read Lynn Olson's description of the development process behind the Ariel, linked in my post above. There is a great deal of design wisdom contained there.
I would like to correct a comment of mine in my earlier post, about the drivers being a dissimilar length along the line. Re-reading, I see that was a design choice of Dr D'Appolito. I'm still not enamoured of the design (Thor) but I bow respectfully to his knowledge and choice of design compromises. I wonder though, if this design was commissioned by SEAS to showcase their Excel drivers.
I would have chosen different drivers, as IME metal cones have a very ragged response at the top of their band. The W18E001 is no exception. Look at the FR and distortion curves in the <a href="http://www.seas.no/excel_line/excel/E0018.PDF">datasheet (here)</a>. The FR rises a huge amount up until the driver resonance at about 4kHz (check the distortion!) which will only be minimally, if at all attenuated by the crossover. Some of the FR rise by the driver will be compensated for, by the TL at the low end (seen in overall system FR), but still not enough. Using room boundaries for LF reinforcement brings other problems, with imaging etc. I don't have my calculator with me, but the high pass to the tweet must be at around 1kHz so at any sort of decent level it's gonna compress and distort. Yuk!
The W18E001 also has a poxy small magnet (0.88T), so it's 10mm x-max useless as I would expect the IMD to be huge at even a fraction of those excursions. To borrow a description of djk's, at any reasonable level, voices would sound like they are "gargling". The more I examine this design the less I like it.
It's your choice, but I wouldn't waste my time or money on it.
Build the Ariel, if you want an MTM TL design.
Cheers
Brett
Bedroc,
If you like the Thor and want to build it, go for it. But don't modify the design until you know what you are doing. Dr D'Appolito has a huge amount of experience and knowledge, and I doubt you will improve much on his design, within the parameters set for it.
If you want it to experiment, build one, and use it as a reference. I test almost all my speaker designs and mods in mono, and only when I feel I am getting close to optimum, do I mess with two. If they don't work in mono, they won't in stereo either. TL's are NOT that straightforward to get a good even response with.
For a novice experimenter (my assumption, sorry if I am wrong) build it by the plan, because you know it will work first time, or a construction error can be found more easily. Then live with it for a while, and experiment with placing, cables, amps etc until you have a real feel for the design. Then, build a modified box, x-over etc and compare it to your 'known quantity'. You will learn more, get faster and more consistent results this way. I mean to encourage you, but in a way that you will enjoy it and get useful results. Nothing worse than a garage full of experimental speakers, and you are no wiser as to what works than when you started, or maybe even more confused and dispirited. My voice of experience speaking here.
Dr D'Appolito has a book on testing loudspeakers that I'm sure would be of benefit, as would Vance Dickason's Cookbook. Version 6 has a new section on TL theory. Also try to get hold of a copy of Olsen's Acoustical Engineering.
Please also read Lynn Olson's description of the development process behind the Ariel, linked in my post above. There is a great deal of design wisdom contained there.
I would like to correct a comment of mine in my earlier post, about the drivers being a dissimilar length along the line. Re-reading, I see that was a design choice of Dr D'Appolito. I'm still not enamoured of the design (Thor) but I bow respectfully to his knowledge and choice of design compromises. I wonder though, if this design was commissioned by SEAS to showcase their Excel drivers.
I would have chosen different drivers, as IME metal cones have a very ragged response at the top of their band. The W18E001 is no exception. Look at the FR and distortion curves in the <a href="http://www.seas.no/excel_line/excel/E0018.PDF">datasheet (here)</a>. The FR rises a huge amount up until the driver resonance at about 4kHz (check the distortion!) which will only be minimally, if at all attenuated by the crossover. Some of the FR rise by the driver will be compensated for, by the TL at the low end (seen in overall system FR), but still not enough. Using room boundaries for LF reinforcement brings other problems, with imaging etc. I don't have my calculator with me, but the high pass to the tweet must be at around 1kHz so at any sort of decent level it's gonna compress and distort. Yuk!
The W18E001 also has a poxy small magnet (0.88T), so it's 10mm x-max useless as I would expect the IMD to be huge at even a fraction of those excursions. To borrow a description of djk's, at any reasonable level, voices would sound like they are "gargling". The more I examine this design the less I like it.
It's your choice, but I wouldn't waste my time or money on it.
Build the Ariel, if you want an MTM TL design.
Cheers
Brett
Ariel
Aaron,
your post came whilst I was doing my last one. Build the Ariel. It's very, very good.
It looks complicated on paper, but it's not really. All the internal sections are the same width, so get these cut at a shop with a large precision saw table. Get them to do all the outside pieces at the same time. This was how I did mine. All the timber, plus all the pieces cut was about $A200 (~$US100).
Next, take the internal drawings (cross-section) and have them printed full size. Cost only a few bucks at my local Kinko's. Then you can trim all the internal pieces to size on the plan, and even transfer these drawings to the inner side wall piece of the Ariel. Now you know exactly where to mount the pieces, drill etc. Glue and screw them all together. Most of the internal pieces are not that difficult a shape, with only a few mitres to worry about.
It's really not that hard. I had some advice from my brother (cabinetmaker), and he took me to the place where I got the pieces cut, but otherwise I did it all myself, including working out how to lay out the pieces as described above. I am NOT a good carpenter, and am an EE by profession. It takes some time, and thought and planning, but what doesn't?
As for the driver prices (from Madisound's website)--
Thor
*T25CF 002 = $149.00 ea
*W18E-001 = $143.80 ea
Ariel
*Vifa P13WH-00 = $36.60 ea
*Scan-speak D2905/9500 = $84.00 ea
The crossovers in both designs are of similar complexity and should cost much the same depending on the quality and brand of the parts chosen. Build either into a box outside the speaker enclosure to minimise microphonics. If you build the Ariel, use the components Lynn has suggested exactly. HE voiced the speaker. I must disagree with Dr D'Appolito re the Thor crosssover: I would not use polyester caps. IME the parts quality in, and the design of, a passive crossover can be more important than the drivers used.
HTH
Cheers
Brett
PS: I will be finishing my Ariels late Feb / March as I am moving house very soon, and would be glad to help in any way I can. Feel fre to email me. The Ariel builders groups should also be a valuable source of information.
Aaron,
your post came whilst I was doing my last one. Build the Ariel. It's very, very good.
It looks complicated on paper, but it's not really. All the internal sections are the same width, so get these cut at a shop with a large precision saw table. Get them to do all the outside pieces at the same time. This was how I did mine. All the timber, plus all the pieces cut was about $A200 (~$US100).
Next, take the internal drawings (cross-section) and have them printed full size. Cost only a few bucks at my local Kinko's. Then you can trim all the internal pieces to size on the plan, and even transfer these drawings to the inner side wall piece of the Ariel. Now you know exactly where to mount the pieces, drill etc. Glue and screw them all together. Most of the internal pieces are not that difficult a shape, with only a few mitres to worry about.
It's really not that hard. I had some advice from my brother (cabinetmaker), and he took me to the place where I got the pieces cut, but otherwise I did it all myself, including working out how to lay out the pieces as described above. I am NOT a good carpenter, and am an EE by profession. It takes some time, and thought and planning, but what doesn't?
As for the driver prices (from Madisound's website)--
Thor
*T25CF 002 = $149.00 ea
*W18E-001 = $143.80 ea
Ariel
*Vifa P13WH-00 = $36.60 ea
*Scan-speak D2905/9500 = $84.00 ea
The crossovers in both designs are of similar complexity and should cost much the same depending on the quality and brand of the parts chosen. Build either into a box outside the speaker enclosure to minimise microphonics. If you build the Ariel, use the components Lynn has suggested exactly. HE voiced the speaker. I must disagree with Dr D'Appolito re the Thor crosssover: I would not use polyester caps. IME the parts quality in, and the design of, a passive crossover can be more important than the drivers used.
HTH
Cheers
Brett
PS: I will be finishing my Ariels late Feb / March as I am moving house very soon, and would be glad to help in any way I can. Feel fre to email me. The Ariel builders groups should also be a valuable source of information.
Cost of the Thor speaker
Arnach, i found a Thor kit at Madisound : http://www.madisound.com/thor.html
costing :
Per pair price with standard crossover and cabinets: $1550.00
Per pair price without cabinets: $960.00
Per pair price with premium crossover and cabinets: $1750.00
Per pair price without cabinets: $1160.00
But these prices seems excessive to me. The drivers and filters would cost 360$ in Sweden where i live.
But after the posts in this forum im enticed by the Ariel speaker. Im sure the Ariel sounds better after all the work and research put into that speaker. But the complexity of the Ariel... A real challange... I havent decided yet.
Arnach, i found a Thor kit at Madisound : http://www.madisound.com/thor.html
costing :
Per pair price with standard crossover and cabinets: $1550.00
Per pair price without cabinets: $960.00
Per pair price with premium crossover and cabinets: $1750.00
Per pair price without cabinets: $1160.00
But these prices seems excessive to me. The drivers and filters would cost 360$ in Sweden where i live.
But after the posts in this forum im enticed by the Ariel speaker. Im sure the Ariel sounds better after all the work and research put into that speaker. But the complexity of the Ariel... A real challange... I havent decided yet.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.