Few said:
My goal of building my ESL's was maximum performance with a simple design. I knew I was going to build laser beams of sound by using a single panel. I also wanted to use the ESL for the bulk of the musical spectrum. I don't listen at ear-bleeding levels, and I'm really sold on the ESL thing for all portions of the audio range that matter. So I wanted to make the cone drivers play as low as possible. It's really a max SPL vs. crossover point tradeoff. I can de-bone a chicken with the panels and my 20 watt tube amps, so I was pretty happy with the max SPL portion, and pushed the XO down until I had all panel playing the range that I consider music. the woofer is just thumping along to the music.
I keep toying with the idea of building ribbons but with all the designs I have seen or considered, they don't play low enough, and you are really listening to cone drivers with a ribbon helping out in the high end.
I suspect that if you build a reasonable panel and a good woofer enclosure,you can run the bottom octave woofers you are planning right up to the panel crossover point. Unless you also want to de-bone chickens with the panels when not using them for music. If I read your description right, you are going to have a sealed enclosure for the bottom couple of octaves (20-80 hz?), and then are going to run a big line array of mid-woofers for the next two+? (80-240 hz?) It seems that if you keep your sealed woofers close to the panels (at the bottom works well), that you could get away without the line array pretty easily. Certainly it's worth a test without them before you pay for all the drivers and the amplifier channels.
I like the idea of the wire stators, and I'm seriously considering that as my next ESL experiment (it's the next logical increase in complexity in my design). I suspect that with added dispersion, the dynamics will suffer, but that way be worth it.
Sheldon
PS I just re-read this and it's a rambling turd of a post, but it may be useful so I'm hitting the send button.
Well, one person's rambling turd is another person's bed of roses...or something like that. In any case, I find it very useful to bounce the ideas around rather than only listen to the voices in my head. People who do too much of the latter seem always to end up in trouble.
I'm currently running big panels down to about 200-300 Hz, and then using a pair of NHT 1259's in sealed boxes below that point. I knew going in that the woofers would really be pushed to a higher than ideal crossover point, but I just wanted to get the panels up and listenable. I've now been listening to this scheme for several years, and have always felt that the midbass is really the weak point---that and the laser beam problem. I'd like to make the next panels smaller, not larger, so moving the crossover point lower doesn't look very promising. I have to admit, though, I've noted with interest that your "dot spacers" between the diaphragm and the stators seems to keep the resonance frequency quite low compared to the "stripe spacers" I used in my panels. Nonetheless, I'd really like to try a smaller diaphragm/stator spacing in the next generation of panel, and that's inconsistent with a low crossover point as well.
I've heard many people complain about the difficulty of mating cones to ESLs, and then parrot the explanation that it's because the cone drivers are too "slow." I think that explanation is complete BS, so I started looking for other causes of the perceived trouble. My current hypothesis, and the one I was hoping to test by using the dipole woofer array idea, is that the disparity between the radiation patterns of a single woofer in a box and a dipolar panel is the real cause of the disconnect everyone complains about. Going to a dipolar woofer arrangement might therefore solve the problem.
Also, I would think a decent woofer, or even better, an array of several them, would behave quite pistonically, and probably not be driven into heavy distortion, if used from 80 - 400 Hz or so. Do you think there's a fundamental difference in the sound generated by an ESL panel versus an array of decent woofers in that range? (This is an actual question, not just rhetoric.) I'm more and more inclined to think two types of dynamic woofer, both with fairly low distortion, would sound pretty much the same in that frequency range if their responses were equalized flat, even if they're made with different cone materials. By extension, I can't help wondering whether dipolar woofers working in their happiest range might not sound at least as good as an ESL pushed into large excursions. Here I'm defining "happiest range" as the loudness and frequency range within which the cones' excursions are kept well below x-max, and the cones vibrate pistonically. I could be all wrong about this (and lots of other stuff) but it'd be interesting to find out. Maybe I'm just getting caught up in the latest fad of using dynamic drivers without enclosures. I never heard the Martin-Logan Statement E2, but those that heard them seemed to be impressed. I wouldn't aim to duplicate the grand scale (and expense) of those speakers, but perhaps a more down to earth version would still be worthwhile.
Sorry if this has ended up being a follow-up rambling turd, but I thought I'd spew some of the reasoning that's been causing me to consider the woofer array plus ESL panel combo. I found it useful to write it, even if nobody else finds it useful to read it!
Few
I'm currently running big panels down to about 200-300 Hz, and then using a pair of NHT 1259's in sealed boxes below that point. I knew going in that the woofers would really be pushed to a higher than ideal crossover point, but I just wanted to get the panels up and listenable. I've now been listening to this scheme for several years, and have always felt that the midbass is really the weak point---that and the laser beam problem. I'd like to make the next panels smaller, not larger, so moving the crossover point lower doesn't look very promising. I have to admit, though, I've noted with interest that your "dot spacers" between the diaphragm and the stators seems to keep the resonance frequency quite low compared to the "stripe spacers" I used in my panels. Nonetheless, I'd really like to try a smaller diaphragm/stator spacing in the next generation of panel, and that's inconsistent with a low crossover point as well.
I've heard many people complain about the difficulty of mating cones to ESLs, and then parrot the explanation that it's because the cone drivers are too "slow." I think that explanation is complete BS, so I started looking for other causes of the perceived trouble. My current hypothesis, and the one I was hoping to test by using the dipole woofer array idea, is that the disparity between the radiation patterns of a single woofer in a box and a dipolar panel is the real cause of the disconnect everyone complains about. Going to a dipolar woofer arrangement might therefore solve the problem.
Also, I would think a decent woofer, or even better, an array of several them, would behave quite pistonically, and probably not be driven into heavy distortion, if used from 80 - 400 Hz or so. Do you think there's a fundamental difference in the sound generated by an ESL panel versus an array of decent woofers in that range? (This is an actual question, not just rhetoric.) I'm more and more inclined to think two types of dynamic woofer, both with fairly low distortion, would sound pretty much the same in that frequency range if their responses were equalized flat, even if they're made with different cone materials. By extension, I can't help wondering whether dipolar woofers working in their happiest range might not sound at least as good as an ESL pushed into large excursions. Here I'm defining "happiest range" as the loudness and frequency range within which the cones' excursions are kept well below x-max, and the cones vibrate pistonically. I could be all wrong about this (and lots of other stuff) but it'd be interesting to find out. Maybe I'm just getting caught up in the latest fad of using dynamic drivers without enclosures. I never heard the Martin-Logan Statement E2, but those that heard them seemed to be impressed. I wouldn't aim to duplicate the grand scale (and expense) of those speakers, but perhaps a more down to earth version would still be worthwhile.
Sorry if this has ended up being a follow-up rambling turd, but I thought I'd spew some of the reasoning that's been causing me to consider the woofer array plus ESL panel combo. I found it useful to write it, even if nobody else finds it useful to read it!
Few
Hi,
I am with You Few
Imho the ´eplanation´ fast ESL versus slow cone is utterly BS
I too agree that the problem of mating ESLs and cones is solely due two factors:
- change of radiaton pattern
- differences in acoustical filter-response
I´m working with dipole-woofers for years now and always made the experience that even a single source dipole-woofer mated well with ESLs. Things improve further when You use strip-like ESLs and line-array-dipole-(sub)woofers (may I call them LADS) since then the SPL/distance-distribution becomes also the same. Theoretically and in measurement the dipole-basses loose on output below a certain freq, which depends on the rooms dimensions (astonishing though, that it doesn´t seem to do so sonically). As a ´solution´ it´s suggested to add a more classical subwoofer for the lowest app. two octaves. That is exactly what ML has done with its Statement.
In case You don´t have the resources needed for a Statement You could in first place omit the subwoofer and redesign the rest on a more compact scale.
That´s what I have done with my system (even though I haven´t read ML´s infos about their Statement at that time)
You may have seen the pic before
The panel is ~ the size of the older ML Sequel (125x25cm), but having smaller holes and higher openness than those. I work with a thinner film than ML, which leads to a lower resonance frequency (170Hz vers. 250Hz; can´t pull the film as much as the 4 times thicker ML-film), but to more pronounced highs and more resolution.
The panel is crossed over with a 2nd order HP and a Notch (placed on the Fs) at ~300Hz. The acoustical filter-function is rather 6th order.
The woofer is an array of 8 18cm drivers in an A-shaped dipole-cabinet. The A-shape lowers the drivers free-air-resonance significantely (in this case ~5Hz). It´s dimensions are so small that the first ´dipole-peak´ lies just above 300Hz. Together with the crossover this results in an acoustical filter-function that is equally 6th-order.
So both of the above mentioned factors are fulfilled.
Since dipole-woofers are further known to play very precise and without ´boooom´, this combination plays imo exceptional well and homogenous.
Dynamics are way above what one is used from typical ESLs and reach nearly horn-like levels.
I think I can imagine what a system like the ML Statement could do when cost is no object but hardly could think of any other speaker that could come even close to that.
jauu
Calvin
I am with You Few
Imho the ´eplanation´ fast ESL versus slow cone is utterly BS
I too agree that the problem of mating ESLs and cones is solely due two factors:
- change of radiaton pattern
- differences in acoustical filter-response
I´m working with dipole-woofers for years now and always made the experience that even a single source dipole-woofer mated well with ESLs. Things improve further when You use strip-like ESLs and line-array-dipole-(sub)woofers (may I call them LADS) since then the SPL/distance-distribution becomes also the same. Theoretically and in measurement the dipole-basses loose on output below a certain freq, which depends on the rooms dimensions (astonishing though, that it doesn´t seem to do so sonically). As a ´solution´ it´s suggested to add a more classical subwoofer for the lowest app. two octaves. That is exactly what ML has done with its Statement.
In case You don´t have the resources needed for a Statement You could in first place omit the subwoofer and redesign the rest on a more compact scale.
That´s what I have done with my system (even though I haven´t read ML´s infos about their Statement at that time)
You may have seen the pic before
An externally hosted image should be here but it was not working when we last tested it.
The panel is ~ the size of the older ML Sequel (125x25cm), but having smaller holes and higher openness than those. I work with a thinner film than ML, which leads to a lower resonance frequency (170Hz vers. 250Hz; can´t pull the film as much as the 4 times thicker ML-film), but to more pronounced highs and more resolution.
The panel is crossed over with a 2nd order HP and a Notch (placed on the Fs) at ~300Hz. The acoustical filter-function is rather 6th order.
The woofer is an array of 8 18cm drivers in an A-shaped dipole-cabinet. The A-shape lowers the drivers free-air-resonance significantely (in this case ~5Hz). It´s dimensions are so small that the first ´dipole-peak´ lies just above 300Hz. Together with the crossover this results in an acoustical filter-function that is equally 6th-order.
So both of the above mentioned factors are fulfilled.
Since dipole-woofers are further known to play very precise and without ´boooom´, this combination plays imo exceptional well and homogenous.
Dynamics are way above what one is used from typical ESLs and reach nearly horn-like levels.
I think I can imagine what a system like the ML Statement could do when cost is no object
but hardly could think of any other speaker that could come even close to that.jauu
Calvin
What's the point about different radiation patterns?
As you sit on your chair you will hear amplitude and phase, not radiation pattern, since the latter is a function of distance.
Besides, the theoretical radiation pattern is ruined by the boundaries of the limited size of the listening so it is less a problem (if it is a problem at all) than predictied by theory.
From my experinece with the sequel2, the problem is more in the added colouration (and distortion?) of the sealed design. The whole speaker shakes and resonanates because of the sealed woofer which is connected with wooden parts with the ESL-frame. Just put your hand on the panels while playing music!
Dipole configuration is one way to solve this.Ï've heard commercial ones employing two woofers, but peronally I would use more than two like the array Calvin has build (nice one!), to get more output at low bass
First I'll try a kind of transmission line to blend with the curved wire esl. If this works I will report here. Might take a lot of months/years because the lack of time.........
As you sit on your chair you will hear amplitude and phase, not radiation pattern, since the latter is a function of distance.
Besides, the theoretical radiation pattern is ruined by the boundaries of the limited size of the listening so it is less a problem (if it is a problem at all) than predictied by theory.
From my experinece with the sequel2, the problem is more in the added colouration (and distortion?) of the sealed design. The whole speaker shakes and resonanates because of the sealed woofer which is connected with wooden parts with the ESL-frame. Just put your hand on the panels while playing music!
Dipole configuration is one way to solve this.Ï've heard commercial ones employing two woofers, but peronally I would use more than two like the array Calvin has build (nice one!), to get more output at low bass
First I'll try a kind of transmission line to blend with the curved wire esl. If this works I will report here. Might take a lot of months/years because the lack of time.........
Thanks for the contribution, Calvin. I have indeed seen the photograph of your system before, but I can always use another shot of inspiration so I'm glad you included it. Can you tell me what you mean by "A-shaped" dipole cabinet?
In response to MJD, here's the reasoning that has led to my interest in the radiation pattern. Much of the sound reaching my ears is radiated off-axis by the speaker and then reflected toward my ears by a room boundary. I'd like those reflections to have a similar sonic character to the on-axis sound, and this requires that the speaker's directivity be considered.
Also, if the sound from a tall narrow ESL drops off roughly as 1/distance, but the sound from a single sealed woofer, for example, drops off as 1/distance squared, then you can only match the levels of the two responses for one particular listening distance. At shorter distances the woofer will be too loud and at longer distances it will be too quiet. You might argue that if there's only one listener, and that listener is always in the same spot (not the case for me) then it isn't an issue, but the room reflections once again complicate the issue. The reflected sound will follow a longer path before reaching the listener and so once again the woofer response will not properly match that of the ESL. I certainly agree that the room reflections will tend to cause deviations from this simple picture, but I'm not confident that they'll automatically make things better. I'd rather ensure that there's no disparity between the direct and reflected sounds. There seems to be mounting evidence, from the work of Floyd Toole and others, that better sound results when the radiation pattern is given due consideration.
So, if you listen in an anechoic chamber, or outdoors, then only the on-axis sound may be of any importance, but in a more conventional listening environment the indirect sounds matter as well. The bottom line, it seems to me, is that the presence of the room boundaries is the reason that radiation pattern matters, not the reason it can be ignored. Then again, I'm not the one whose done all the studies, so I don't have first-hand experience to back all this up.
I do agree with you that more woofers makes sense in a dipole configuration. I want the height of the woofer array to match the height of the ESL, and I don't like the idea of an 8" woofer having to move +/-10 mm in order to keep up---even if the manufacturer claims it's up to the task. Too many compromises must be made to achieve a huge x-max; I'd rather pay for more woofer cone area (multiple drivers) than invest in one woofer that can move a country mile.
Few
In response to MJD, here's the reasoning that has led to my interest in the radiation pattern. Much of the sound reaching my ears is radiated off-axis by the speaker and then reflected toward my ears by a room boundary. I'd like those reflections to have a similar sonic character to the on-axis sound, and this requires that the speaker's directivity be considered.
Also, if the sound from a tall narrow ESL drops off roughly as 1/distance, but the sound from a single sealed woofer, for example, drops off as 1/distance squared, then you can only match the levels of the two responses for one particular listening distance. At shorter distances the woofer will be too loud and at longer distances it will be too quiet. You might argue that if there's only one listener, and that listener is always in the same spot (not the case for me) then it isn't an issue, but the room reflections once again complicate the issue. The reflected sound will follow a longer path before reaching the listener and so once again the woofer response will not properly match that of the ESL. I certainly agree that the room reflections will tend to cause deviations from this simple picture, but I'm not confident that they'll automatically make things better. I'd rather ensure that there's no disparity between the direct and reflected sounds. There seems to be mounting evidence, from the work of Floyd Toole and others, that better sound results when the radiation pattern is given due consideration.
So, if you listen in an anechoic chamber, or outdoors, then only the on-axis sound may be of any importance, but in a more conventional listening environment the indirect sounds matter as well. The bottom line, it seems to me, is that the presence of the room boundaries is the reason that radiation pattern matters, not the reason it can be ignored. Then again, I'm not the one whose done all the studies, so I don't have first-hand experience to back all this up.
I do agree with you that more woofers makes sense in a dipole configuration. I want the height of the woofer array to match the height of the ESL, and I don't like the idea of an 8" woofer having to move +/-10 mm in order to keep up---even if the manufacturer claims it's up to the task. Too many compromises must be made to achieve a huge x-max; I'd rather pay for more woofer cone area (multiple drivers) than invest in one woofer that can move a country mile.
Few
Hi,
with A-shaped I mean that the cabinet is wider at the front than at the back (in this case the front is 200mm wide, while the backside is only 132mm wide). This has two effects, one positive and one negative.
The negative one is that there will be a pronounced peak of several dB at ~200-400Hz -depending on the cabinet´s size. As long as this peak is well above Your desired frequency range You can cope with that problem quite easy by electrical filtering (a simple lowpass might be sufficient, a notch might be needed)
The positive effect is that the drivers freeair resonance is lowered more than a simple OpenBaffle or a U-shaped cabinet would allow for.
The 16 drivers had a mean Fs value of >40Hz and a built in Fs of 34-35Hz, which I consider very alright for a 6.5" driver system. Too does this cabinet shape look more pleasing to my eyes.
Even though the drivers were quite cheap, they feature a nice stiff membrane and magnetical shielding and a really nice set of parameters for dipole use.
With regard to radiation pattern:
I have very serious room resonances and couldn´t listen to double base and such stuff. The dipole lessend this problem alot but just when I built the tower-dipole I nearly completely got rid of the strong vertical room mode too (the tower´s radiation pattern becomes rather a dipolar cylinder -like the ESL panel- than just two dipolar lobes or even worse a monopole, thereby the vertical room modes too are less critical).
jauu
Calvin
with A-shaped I mean that the cabinet is wider at the front than at the back (in this case the front is 200mm wide, while the backside is only 132mm wide). This has two effects, one positive and one negative.
The negative one is that there will be a pronounced peak of several dB at ~200-400Hz -depending on the cabinet´s size. As long as this peak is well above Your desired frequency range You can cope with that problem quite easy by electrical filtering (a simple lowpass might be sufficient, a notch might be needed)
The positive effect is that the drivers freeair resonance is lowered more than a simple OpenBaffle or a U-shaped cabinet would allow for.
The 16 drivers had a mean Fs value of >40Hz and a built in Fs of 34-35Hz, which I consider very alright for a 6.5" driver system. Too does this cabinet shape look more pleasing to my eyes.
Even though the drivers were quite cheap, they feature a nice stiff membrane and magnetical shielding and a really nice set of parameters for dipole use.
With regard to radiation pattern:
I have very serious room resonances and couldn´t listen to double base and such stuff. The dipole lessend this problem alot but just when I built the tower-dipole I nearly completely got rid of the strong vertical room mode too (the tower´s radiation pattern becomes rather a dipolar cylinder -like the ESL panel- than just two dipolar lobes or even worse a monopole, thereby the vertical room modes too are less critical).
jauu
Calvin
. hi
the 4 tranny setup does work pretty good .
toroids seem to me as more linear . with better bass
iron trannys sound thinner . these are not big diferences
i use 4 ohm in series to condition the load to the amp and dump transformer ringing .
adding more trannys same way stresses the insulation of the ending pair making them hiss and sound like music .
the 4 tranny setup does work pretty good .
toroids seem to me as more linear . with better bass
iron trannys sound thinner . these are not big diferences
i use 4 ohm in series to condition the load to the amp and dump transformer ringing .
adding more trannys same way stresses the insulation of the ending pair making them hiss and sound like music .
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.