HiFiNutNut said:
I noticed you've updated your NaO system. Is it totally fulfilling now or are you still looking for that Esotar NaO configuration?
🙂
If I remember correctly you have pretty large room. How adequate NaO system is in your environment? Woofer setup is very different than what I have in mind but how the top end MTM handles it? 120Hz cross to woofers? 4th order?
Steves BoB and NaO are two good examples how this business isn't that exact as some theorys suggest. Like the theorys I've mensioned here. Both have low crossoverpoint to woofers, both have MTM vertical directivity and no waveguide for the tweeter. And still very few or none complain about their sound.
I have considered the rear tweeter in this setup. Have you experimented how it's action changes the whole setup? I remember NaO rear tweeter has a level adjustment and was there a on/off switch as well?
I know my waveguide setup with dipole mids front radiation matched radiation pattern causes small drop in power response. Is that audible? At least the nonwaveguided solutions don't have this drop but they have uneven front radiation pattern instead. In theory rear tweeter with mids rear radiation pattern matched behavior would solve the problem. But the panel isn't that well behaved above 1Khz at 180 degree angle anyway so I don't know does it make that much difference.
First hand experiences would be very welcome.
Jussi
Hello Jussi,
I have to say this is an interesting design. I too am designing my own dipole speakers and hope your experiences will be of use and vice versa.
Question:
Will you be documenting this project on an external website in addition to the forum?
I have to say this is an interesting design. I too am designing my own dipole speakers and hope your experiences will be of use and vice versa.
Question:
Will you be documenting this project on an external website in addition to the forum?
Hi Jussi,
first, you've done a lot of thinking on this and in many ways I recognize my own meanderings in my 6 yr+ dipole journey... So I'll try to give a few comments about my experiences.
Some history: I started with a fullrange dipole in '99 or so, augmented by 2x 10" woofers on a slim and shallow W-F-W U-shaped panel. It all started with SL's website, but at the time he only had the Phoenix and I had other ideas about form factor and wanted to keep the mids-highs full range, etc. So I designed my own. Later I added a tweeter. Then I rebuilt the panel into a 24" wide 48" high 7.5" deep WWMT U-frame with a SS8543 6.5" mid and Seas 27TBFCG tweeter, and push-pull 10" Vifa woofers on the bottom. This is what it is now, though I keep on fiddling with the x-o points (100 and 1700 Hz as of now) which is easy with my MOX variable X-O, and generally yields only minor audible differences.
Surprisingly, while the system did improve over the various iterations in construction and optimization, even the drastic changes in driver choice over time did not alter the general characteristics or balance. It keeps on getting more precise and more polished, but it was pleasurable even at a time when I did not fully understand or correct some issues. There is still isues - more output capability on the low end, more modern woofers for instance - so there is still room for more optimization. One possible conclusion is that suboptimal dipoles will sound good enough to lure many DIYers into never tying to really reach their full potential, in the way SL's Orion is optimized to the last little detail. Conclusion 2 IMO is that *all* of SL's x-o and filtering techniques are required for serious consideration, because the address real issues that other, different dipole systems, will also have. Ignore at your peril, SL does know his stuff.
Now for the details, in random order
- rear tweeter: I tried that briefly. Intriguing at first (more depth), deemed unnecessary and distracting after a week or so of listening. But did not try to optimize it.
- waveguide: likely the power response deviation has more drawbacks than the gain in radiation pattern matching. It just doesn't seem to be much of an issue if M and T differ in 2 or 3 dB at 45 or 60 degrees. The global power response accuracy however is a beautiful thing. My dipoles always astonished me in this respect. Off axis and in a different room, that's when they sound at their most puzzlingly realistic. I can't explain why, I find this highly ironic. Ha. So, my conclusion is that power response actually matters more than on axis smoothness.
- never noticed a major difference between WFW and WWMT in their vertical patterns in-room. Vertically accurate dispersion does not seem to matter all that much to me.
- however, floor bounce does show up strongly in measurements.
- however again, this is not something the ear seems to notice as a spectral imbalance - notching it out does make sound things thinner, but not more accurate. The typical delay of say, 6-10 ms of the floor bounce, is enough for the ear to recognize it as a reflection, I wouldn't bother making major design decisions dependent on the floor bounce issue.
- the opposite is true for the 200-100 floor gain: don't forget SL's 200-100 shelving HP. Without it, the bass *does* sound bloated, even though it *measures* quite flat (!) (at least, when *I* measure it). This, I can't explain it.
- On the low end I just can't hear much difference between setting the rolloff at 20-sh , vs say, at 40-ish Hz. There does not seem to be enough content below say, 35-40 Hz, to bother and optimize everything around optimal output in the bottom octave
- however, low distortion and high ouput capability seems to matter a lot in the 80-200 range
- for instance I just measured my woofer's distortion and realized these "lowly" older Vifas are cleaner up until 140 Hz, than the expensive SS8543's (which are really great from 200 up though). Unless I made some mistake. The larger drivers are just less stressed in this range, and this translates into less HD.
- in other words, it seems a good idea to run the large drivers as high as possible, even though they will be less accurate than the smaller drivers in terms of energy storage.
- when I measured my woofers I found *more* 2nd and *less* 3rd HD in push pull, compared to a single woofer at the same SPL. I can't explain it. Push pull may not always give the expected results.
- one thing you should address with accuracy is the dipole peak before rolloff. IME it exists and does matter for detail, though it seems benign on measurements, and in the 200-400 range where it occurs for typical baffles it is hard to measure because floor bounce also falls in this region. Once you manage to pinpoint it and notch it out, things do improve quite a bit in clarity.
- metal baffles: not sure if the rigidity advantage offsets possible problems with ringing, not to mention difficulty in construction and therefore difficulty to remedy mistakes.
Ahem, running out of energy now. It's late here. I'll write more if more comes to my mind ...
Disclaimer: all statements are just my personal experience and opinion. I don't claim to know all about speakers, dipoles, or anything else for that matter.
first, you've done a lot of thinking on this and in many ways I recognize my own meanderings in my 6 yr+ dipole journey... So I'll try to give a few comments about my experiences.
Some history: I started with a fullrange dipole in '99 or so, augmented by 2x 10" woofers on a slim and shallow W-F-W U-shaped panel. It all started with SL's website, but at the time he only had the Phoenix and I had other ideas about form factor and wanted to keep the mids-highs full range, etc. So I designed my own. Later I added a tweeter. Then I rebuilt the panel into a 24" wide 48" high 7.5" deep WWMT U-frame with a SS8543 6.5" mid and Seas 27TBFCG tweeter, and push-pull 10" Vifa woofers on the bottom. This is what it is now, though I keep on fiddling with the x-o points (100 and 1700 Hz as of now) which is easy with my MOX variable X-O, and generally yields only minor audible differences.
Surprisingly, while the system did improve over the various iterations in construction and optimization, even the drastic changes in driver choice over time did not alter the general characteristics or balance. It keeps on getting more precise and more polished, but it was pleasurable even at a time when I did not fully understand or correct some issues. There is still isues - more output capability on the low end, more modern woofers for instance - so there is still room for more optimization. One possible conclusion is that suboptimal dipoles will sound good enough to lure many DIYers into never tying to really reach their full potential, in the way SL's Orion is optimized to the last little detail. Conclusion 2 IMO is that *all* of SL's x-o and filtering techniques are required for serious consideration, because the address real issues that other, different dipole systems, will also have. Ignore at your peril, SL does know his stuff.
Now for the details, in random order
- rear tweeter: I tried that briefly. Intriguing at first (more depth), deemed unnecessary and distracting after a week or so of listening. But did not try to optimize it.
- waveguide: likely the power response deviation has more drawbacks than the gain in radiation pattern matching. It just doesn't seem to be much of an issue if M and T differ in 2 or 3 dB at 45 or 60 degrees. The global power response accuracy however is a beautiful thing. My dipoles always astonished me in this respect. Off axis and in a different room, that's when they sound at their most puzzlingly realistic. I can't explain why, I find this highly ironic. Ha. So, my conclusion is that power response actually matters more than on axis smoothness.
- never noticed a major difference between WFW and WWMT in their vertical patterns in-room. Vertically accurate dispersion does not seem to matter all that much to me.
- however, floor bounce does show up strongly in measurements.
- however again, this is not something the ear seems to notice as a spectral imbalance - notching it out does make sound things thinner, but not more accurate. The typical delay of say, 6-10 ms of the floor bounce, is enough for the ear to recognize it as a reflection, I wouldn't bother making major design decisions dependent on the floor bounce issue.
- the opposite is true for the 200-100 floor gain: don't forget SL's 200-100 shelving HP. Without it, the bass *does* sound bloated, even though it *measures* quite flat (!) (at least, when *I* measure it). This, I can't explain it.
- On the low end I just can't hear much difference between setting the rolloff at 20-sh , vs say, at 40-ish Hz. There does not seem to be enough content below say, 35-40 Hz, to bother and optimize everything around optimal output in the bottom octave
- however, low distortion and high ouput capability seems to matter a lot in the 80-200 range
- for instance I just measured my woofer's distortion and realized these "lowly" older Vifas are cleaner up until 140 Hz, than the expensive SS8543's (which are really great from 200 up though). Unless I made some mistake. The larger drivers are just less stressed in this range, and this translates into less HD.
- in other words, it seems a good idea to run the large drivers as high as possible, even though they will be less accurate than the smaller drivers in terms of energy storage.
- when I measured my woofers I found *more* 2nd and *less* 3rd HD in push pull, compared to a single woofer at the same SPL. I can't explain it. Push pull may not always give the expected results.
- one thing you should address with accuracy is the dipole peak before rolloff. IME it exists and does matter for detail, though it seems benign on measurements, and in the 200-400 range where it occurs for typical baffles it is hard to measure because floor bounce also falls in this region. Once you manage to pinpoint it and notch it out, things do improve quite a bit in clarity.
- metal baffles: not sure if the rigidity advantage offsets possible problems with ringing, not to mention difficulty in construction and therefore difficulty to remedy mistakes.
Ahem, running out of energy now. It's late here. I'll write more if more comes to my mind ...
Disclaimer: all statements are just my personal experience and opinion. I don't claim to know all about speakers, dipoles, or anything else for that matter.
Hara said:
Good. This hobby is all about learning. 🙂
I don't know yet. I'll consider it.
MBK said:
Good. Alternative perspectives are always welcome. I'll try to comment something back...
I agree. I guess it has something to do with smooth power response which leads to much smoother operation in random room that pretty much any box speaker can ever achieve. Naturally a box speaker can be optimized to a room and get very good performance out of them. And there are certain differences between them. Personally I don't know are they pretty much room caused differences in positioning and such, propably. Then there are some issues that follow the system all the way to anechoic chamber.
Seems like Siegfried ended up to same conclusion.
Sure it does. I'm just wondering when directivity is high enough to give more weight to front radiation over power response. But naturally both should be matched for optimum result. In my case sidewalls run pretty close so I consider my front radiation matching more important than it would be in wider space.
I don't consider that 3dB drop at power response that big potential problem. Normal box speakers have much more dramatical changes in their power balance. Full bafflestep compensation alone causes some roll off towards midrange and even tweeters matched radiation pattern it just keeps on rolling off towards top octave.
Overally flat power response isn't that good solution. I have heard few systems where constant directivity is fixed to it's maximum and most recordings just don't support that balance. It sounds way too bright without balls. Modern studio operation is causing this while recordings are monitored with normal power response speakers and balance is EQd to match them. In this aspect perfection isn't a good thing since rest of the world is less perfect. But on the other hand unedited recordings where just a single stereopair mics are used work like magic. Unfortunately these are rare, most are done in the 50's and 60's. Naturally these recordings lack treble and upper midrange in modern power response systems.
That sounds a bit optimistic delay for floor bounce. 2,5m listening distance with 0,9m tweeter height create about 36 degree floor bounce angle. And again with 2,5m listening distance this means bounce goes about 50cm longer distance to listeners ears (also 0,9m from floor). 50cm with 340m/s means about 1,5ms delay which is definitely heard as same sound with the direct radiation. And without any vertical directivity bounce isn't even down that much so effect is also pretty dramatical.
With dipole panel radiation from sidewalls isn't that much a problem since system has directivity. Even in my case while sidewalls are pretty close to the panels I'll have 1st reflection from sidewall about 80 degrees off axis from speakers view (while speakers are pointed towards listening position). This creates >10dB directivity from there. Or it doesn't if tweeter isn't waveguided. In vertical plane things are more difficult. Walls can be treated, heavy acoustic treatment to floor is difficult to arrange. And the bounce comes with very short delay...
Overall balance might be ok, even with the biggest dip around 250Hz on it but fixing it should give more natural presentation and more dynamics when floor reflection isn't cancelling some of the sound.
But like you know there are other aspects along with floor bounce improvement. More displacement capacity, less range for individual drivers, more efficiency, lower distorsion.
I'll make a variable 100-200HP circuit. But at least Orions anechoic measurements didn't show any of 100-200HP circuits effect.
This might have something to do with room balance, dipole cancellation and inadequate displacement. One local builder did dipole subwoofers with 150Hz 4th order cross using just single 10" XLS on each side in modest size U-baffle. Radiation pattern is still dipole since U is only 6" deep. He balanced them down to 15Hz flat but there isn't that much bass below 30Hz or so. Cones sure move like heck but there isn't enough displacement to play frequences below that.
I know playing <30Hz from dipole is very stubborn business all together. Pretty much all advantages are gone and the whole thing turns against itself. But there isn't that much <30Hz material so at least I'm willing to make that compromisse. It's harder to match a monopole in mid-upperbass region than using dipoles all the way.
Althought I have heard very promissing results from some Magnepan owners. Seems like adding monopole subwoofer to a dipole mainpanel is easier task than mixing monopole to monopole. At least using monopole a bit over the dipoles range should create some cardioid radiation pattern while two monopoles just boost one another.
This might be intresting issue to experiment. Broad overlap between dipole panels and monopole subwoofer. Maybe even single monopole might be ok if it's placed between the mains, propably along the frontwall. Monopole gives adequate grunt 15-30Hz, dipole comes to play, cooperation to upperbass range and then dipole takes over.
As it does to any other range as well. Has anyone heard such bigger openbaffles as I'm building here? Compared Beethovens to present Orion models for example? How do they compare at that say 80-300Hz range where those lowmids operate?
I know Orions have edge with the midrange since bafflearragement and spine are better and Orion uses at least somewhat better drivers.
Yep. Perfect driver to every range makes the system pretty multiway. But there is also some advantages. Many multiway systems with plenty of drivers have this effortless handmark on their reproduction. Sound doesn't stress, easy and nontiring to listen, lifelike dynamics. And multiway allows lower crossover orders than a system where every driver is runned to its maximum.
I'll have these properly measured when the time comes. On and off axis, vertical and horizontal, distorsions with proper equipment, baffle resonances with accelerometer, the works. So the fine tuning shouldn't be that big problem. Or at least it doesn't depend on measurements and measuring limitations.
Compromisses compromisses. I don't have any chance to make these as slim and elegant as they are without using something different. Still it's just guessing but I'll have them measured so I don't have the chance that there are some factor X caused fault that just makes me scratch my head without progress.
Jussi
MBK said:
May I ask how you got this 6-10ms?
Even the 6ms delay means 2m extra distance. With 2m listening height (me on a small stool) and 5m listening distance I get about 38 degree bounce angle with 1,5m extra distance which leads to under 5ms delay. At least my listening position isn't even close like that. With 2m listening height there is also a bigger problem from the ceiling unless the room is over 4m high.
With 1,5m distance from speaker to front wall I can have 3m extra distance and therefore 9ms delay for dipoles rear radiation. This should be heard as reflection, not direct sound.
I'd consider 1st reflections from sidewalls, ceiling and mostly the floor most harmful to speakers operation in room. And I think pretty large amount of acoustics and studio experts agree with me on this. And handling those problems isn't easy. Acoustic treatment is one. Diffusing propably better than damping (which kills the sound). But in many domestic environments heavy and broadband diffusing isn't possible so at least I'll have to settle for directivity.
Here is a simple example of this. Two point sources at same location. Network 1 demonstrating direct sound, network 2 floor reflection. Later one has 500mm extra "Z" dimension on it so it comes that calculated ~1,5ms later on scene. With zero vertical directivity response (ideally) looks like this.
Jussi
Attachments
And here a simple example of -6dB vertical directivity. Or basically this same example works in horizontal plane as well. As you see there is pretty significant improvement in overall response. For really good result directivity should be something like -20dB which a good electrostatic, long linesource or a long ribbon can create.
Naturally these are just fast made examples. Vertical directivity isn't that even, different surfaces and their reflections even up the result and so on. But even if the response looks good it doesn't proof good results. Covering one reflections mess up with other reflection doesn't make overall result any better. It's like you're using brush and a set of paint to clean up window. After all the hard work you might have a landscape on your window thats similar to actual view (depending how good painter you are) but it's still far from chrystal clear window you can look through.
And I think this applies to many different things in audio. For example you have overbright speakers. Ok. You can neither invest neutral speakers or you can buy an amplifier that has shy treble. Both fix the audible problem but only one of them truely bites the problem, other one just covers it. I think this is partially the way people build their systems in audio / high end world. I think it's not the right way thought.
There are many examples of this behavior in commercial and why not in some DIY systems as well. For example speakerspikes. Why do you use them? To isolate the speakers from the floor. Why? Ok, speaker is shaking and it can't deliver it's vibrations to floorboards. Ok. Spikes fix in nicely. Problem solved. Or is it? I'd consider using properly build box without any resonances what so ever much better solution and it really gets into the problem and fixes it. Unless you want to dance with your speakers which naturally requires serious shakes.
Other example could be room problems overall. Mostly in bassrange. Most people overlooked parametric EQs until Velodyne made one with "interface for dummies". After that it's been praised as the most important thing in subwoofers for decades. And what I've heard Velodynes EQ/crossover isn't even that high tech. Has some electronical problems. Ok. Solves some problems. Or does it? There still are room resonances, room modes and the Eq is there only to ease the headache. Electronical gadgets don't reach acoustic problems. Why not do it better? Make a system that just doesn't wake up room modes. This means directivity. Where do you need the eq anymore? And the result is far more uniform and usable around the room that one spot optimized parametric eq can ever be.
Strange business. But it seems to be quite enough for most people. Sadly this leads commercial companys to slack. They don't even have to invent anything new, make any real R&D. Just make more of the good old gadgets, put some new exotic materials on them, create catchy marketing phrases for them and sell away.
Jussi
Naturally these are just fast made examples. Vertical directivity isn't that even, different surfaces and their reflections even up the result and so on. But even if the response looks good it doesn't proof good results. Covering one reflections mess up with other reflection doesn't make overall result any better. It's like you're using brush and a set of paint to clean up window. After all the hard work you might have a landscape on your window thats similar to actual view (depending how good painter you are) but it's still far from chrystal clear window you can look through.
And I think this applies to many different things in audio. For example you have overbright speakers. Ok. You can neither invest neutral speakers or you can buy an amplifier that has shy treble. Both fix the audible problem but only one of them truely bites the problem, other one just covers it. I think this is partially the way people build their systems in audio / high end world. I think it's not the right way thought.
There are many examples of this behavior in commercial and why not in some DIY systems as well. For example speakerspikes. Why do you use them? To isolate the speakers from the floor. Why? Ok, speaker is shaking and it can't deliver it's vibrations to floorboards. Ok. Spikes fix in nicely. Problem solved. Or is it? I'd consider using properly build box without any resonances what so ever much better solution and it really gets into the problem and fixes it. Unless you want to dance with your speakers which naturally requires serious shakes.
Other example could be room problems overall. Mostly in bassrange. Most people overlooked parametric EQs until Velodyne made one with "interface for dummies". After that it's been praised as the most important thing in subwoofers for decades. And what I've heard Velodynes EQ/crossover isn't even that high tech. Has some electronical problems. Ok. Solves some problems. Or does it? There still are room resonances, room modes and the Eq is there only to ease the headache. Electronical gadgets don't reach acoustic problems. Why not do it better? Make a system that just doesn't wake up room modes. This means directivity. Where do you need the eq anymore? And the result is far more uniform and usable around the room that one spot optimized parametric eq can ever be.
Strange business. But it seems to be quite enough for most people. Sadly this leads commercial companys to slack. They don't even have to invent anything new, make any real R&D. Just make more of the good old gadgets, put some new exotic materials on them, create catchy marketing phrases for them and sell away.
Jussi
Attachments
Hi Jussi,
First off, sorry for getting you into all these floor bounce calculations. The 6-10 ms was a mistake of the late hour. I dimly remembered I got >1.5 ms (which is enough to get out of the "blending" range of the ear) and must have made a mistake on the calculator when I tried to verify it. Anyway for myself I have concluded I won't touch the floor bounce, though your remark on using vertical directivity to limit it got me thinking.
Interesting remarks, again you touch on many thoughts I've had, for instance on mixing monopole and dipole over a wide range, say with a 1st order x-o. My idea came to me because I measured off axis response of my mid and realized that it became true dipole only below 250 Hz or so. Above that, response is hardly down at all until about 1.2 kHz, at either 45 or 60 degrees. The 7,5" deep U-frame must be making it going towards cardioid over most of its operating range. So I thought, the tweeter is closer to a cardioid as well with my wide baffles (not much of the sound wraps around >1.6 kHz as I use it), and so that leaves the bass to transform into cardioid as well if one wanted to be truly picky. So I thought, either add a monopole and overlap dipole and monopole bass, or try SL's suggestions of acoustic resistor. Haven't tried any of these yet though. Interesting remark again on how an accurate directivity at playback may do more harm than good because it is already mixed using monitoring systems with uneven directivity...
As for horizontal reflections, again, if you can get them above a few ms delay, I wouldn't touch them, quite the contrary. In my experience this is what adds the air and presence to the sound. Studios control them to hear just the direct sound, which is necessary for accurate monitoring, but at home you want a "real" feeling, not a microscope the highlights what happened at 5 cm of the singer's mouth 🙄 .
The 100-200 filter does not quite take away 6 dB, it is closer to 4 dB. What i noticed was that when I set my system flat from say 200 down to 50 Hz, measuring at floor level to catch maximum floor effects, it will *still* sound bass bloated! So I just put the filter back in without thinking any further, because it made it sound "right" regardless of my measurements.
BTW lucky if you have anechoic of otherwise siphisticated measurement capabilities. I produce a lot of artefacts in my measurements, my outdoors is too noisy, my indoors forces me to either go too close to the speaker to represent true baffle effects, or so far that room effects mask most of the direct sound. I have become very cautious in interpreting my own measurements. But anyway, I add a few curves, measured indoors at conditions that Ii believe add few artefacts. Most curves octave averaged, which seems a lot, but gives me the correct shape of the responce w/o distracting wiggles due to reflections and diffraction. Just for illustration. First the mids off axis.
First off, sorry for getting you into all these floor bounce calculations. The 6-10 ms was a mistake of the late hour. I dimly remembered I got >1.5 ms (which is enough to get out of the "blending" range of the ear) and must have made a mistake on the calculator when I tried to verify it. Anyway for myself I have concluded I won't touch the floor bounce, though your remark on using vertical directivity to limit it got me thinking.
Interesting remarks, again you touch on many thoughts I've had, for instance on mixing monopole and dipole over a wide range, say with a 1st order x-o. My idea came to me because I measured off axis response of my mid and realized that it became true dipole only below 250 Hz or so. Above that, response is hardly down at all until about 1.2 kHz, at either 45 or 60 degrees. The 7,5" deep U-frame must be making it going towards cardioid over most of its operating range. So I thought, the tweeter is closer to a cardioid as well with my wide baffles (not much of the sound wraps around >1.6 kHz as I use it), and so that leaves the bass to transform into cardioid as well if one wanted to be truly picky. So I thought, either add a monopole and overlap dipole and monopole bass, or try SL's suggestions of acoustic resistor. Haven't tried any of these yet though. Interesting remark again on how an accurate directivity at playback may do more harm than good because it is already mixed using monitoring systems with uneven directivity...
As for horizontal reflections, again, if you can get them above a few ms delay, I wouldn't touch them, quite the contrary. In my experience this is what adds the air and presence to the sound. Studios control them to hear just the direct sound, which is necessary for accurate monitoring, but at home you want a "real" feeling, not a microscope the highlights what happened at 5 cm of the singer's mouth 🙄 .
The 100-200 filter does not quite take away 6 dB, it is closer to 4 dB. What i noticed was that when I set my system flat from say 200 down to 50 Hz, measuring at floor level to catch maximum floor effects, it will *still* sound bass bloated! So I just put the filter back in without thinking any further, because it made it sound "right" regardless of my measurements.
BTW lucky if you have anechoic of otherwise siphisticated measurement capabilities. I produce a lot of artefacts in my measurements, my outdoors is too noisy, my indoors forces me to either go too close to the speaker to represent true baffle effects, or so far that room effects mask most of the direct sound. I have become very cautious in interpreting my own measurements. But anyway, I add a few curves, measured indoors at conditions that Ii believe add few artefacts. Most curves octave averaged, which seems a lot, but gives me the correct shape of the responce w/o distracting wiggles due to reflections and diffraction. Just for illustration. First the mids off axis.
Attachments
Jussi,
The NaO sounds good but I can't say it is totally fulfilling for my particular needs. I would like to play at live level for certain types of music and the NaO just doesn't cut it. Please read under the context - I have a large room larger than the average. For most music at moderate to loud level the NaO is fine. But if I want to play organ music then I am out of luck. The 12" XLS in the U-frame can be bottomed if the music has much content under 40Hz. Also the mid-woofers can be bottomed as well. When playing Johnny Cash a bit loud in my room I found distortions coming from the lower midrange / upper bass - the mid-woofers simply run of linear excursion.
Well, I must make it clear that is not to blame John K's work. Indeed I have to say the NaO is exceptionally good as long as you don't place it in a large room like mine and you don't desire the SPL at crazy level like mine. For a small to medium size room I can highly recommend the NaO.
Mine is the original NaO that uses a single 12" XLS crossed at 100Hz. John K since replaced the 12" XLS with one or two 10" XLS and moved the XO point to 120Hz. This was one stone to kill two birds: (1) with a high XO point the mid-woofer is less likely to be bottomed so the overall SPL of the system can be higher; (2) With a single or dual 10" woofer(s) the U-frame was made shorter which helps to raise the first resonance frequency of the U-frame. Because mine is the 12" woofer with a long U-frame I found the first resonance frequency can not be damped fully. I reduced the setting over that range and this does not make the music 100% smooth when crossing between the woofers and the midranges.
I guess the NaO II has addressed the problems I have with the original NaO, but still, mine is not a NaO II and I don't expect the NaO II can play at the SPL level I want. Interestingly, John K said that he never found the U-frame bottoming out on his single 10" XLS and I was the first one in the world to suggest dual 12" for the NaO.
I believe John K has published a lot of information about the NaO including the above before so I think the above does not disclose (enough) technical information about the NaO to make John K upset about it.
That was why I asked if you have considered the 26W8861. Of course at the time I had not read your entire text and thought you were going to use them as woofers. By the way, I have read all of your posts now and thank you so much for sharing your knowledge and experience with us.
I have not connected the rear tweeters yet so I can't comment on them.
I have to say the passive XO of the original NaO was only averaged to good but the new passive XO is superb! It is the low loss and very smooth integration of the XO give the NaO outstanding sound. I tuned the MTM part of my original WWMTMWW speakers for two years to make it near perfection and I have to say the NaO is certainly not lesser. Therefore, I have basically abandoned my thoughts on using the Esotar T330D on the NaO because I don't want to destroy the best part of the NaO. It appears the Excel T25 may be sonically more transparent than the Esotar and it is the last driver to be bottomed in the NaO so I think I will keep the arrangement as it is. Not that the Esotar is less better given its practically nearly unlimited power handling capability I found musicality and smoothness in the Esotar possibly better than the T25. The NaO panel above 150Hz, I would say it is very difficult to beat and it can play fairly loud. What I want is simply to remove any resonance in the U-frame and improve the SPL capability under 150Hz substantially.
Having had the WWMTMWW I have to say I love speakers capable of doing high SPL. Whether I play them loud or not is a different story but they just give you the uncompressed sound and can deliver as much power as you want at your wish. When an amplifier clips we can easily tell. When the drivers run of linear excursion the sound just appears to be compressed.
With regards to your directivity / waveguide, I do find the tweeter energy is stronger at my sitting position. But then which speakers don't have similar problems? At high frequencies it beams and that is physics.
As a matter of fact I am going backward and will experiment using a sealed box for 100Hz down. This is because my room has the first room mode at 2X Hz so a dipole or monopole should make not much difference. Of course, if I mount 4 x 12" U-frame per side and make the U-frames shorter and raise the XO point then it might just do it for me, just. Shortening the U-frame may decrease the SPL by, say, 6dB while using 4 only gives a 12dB gain so at the end I only get 6dB gain, which is not much. I won't be crazy enough to use 8 x 12" per side. Oh, this is the only evil of dipole bass.
Some people commented that low excursion stiff cone drivers sound much more musical in bass so I guess I will try a simple sealed box with 2 x 26W8861 doing push-pull in 9x litres with a Q of about or a bit below 0.7. I think I will use active EQ to boost the bottom a bit. This should give me sufficient SPL I want. What do you think?
More importantly (please correct me if I am wrong), with a U-frame the first resonance is at 1/4 wavelength. With a sealed box it is at 1/2 wavelength. In other words, currently my U-frame with an XO at 100Hz has a first resonance at about between 150Hz to 200Hz (I don't want to disclose John's design) that can only be partially damped. If I use a sealed box, the first resonance will be around or above 300Hz. XOed at 100Hz I believe the sealed box will give a very clean sound comparing to the U-frame, especially given that my room mode is as low as 23Hz, for a sealed box should work.
Sorry I have stolen your space to ask you questions. But I guess we are sharing our experience with our speakers that have basically the same design. Of course, you are far more advanced than me.
Regards,
Bill
The NaO sounds good but I can't say it is totally fulfilling for my particular needs. I would like to play at live level for certain types of music and the NaO just doesn't cut it. Please read under the context - I have a large room larger than the average. For most music at moderate to loud level the NaO is fine. But if I want to play organ music then I am out of luck. The 12" XLS in the U-frame can be bottomed if the music has much content under 40Hz. Also the mid-woofers can be bottomed as well. When playing Johnny Cash a bit loud in my room I found distortions coming from the lower midrange / upper bass - the mid-woofers simply run of linear excursion.
Well, I must make it clear that is not to blame John K's work. Indeed I have to say the NaO is exceptionally good as long as you don't place it in a large room like mine and you don't desire the SPL at crazy level like mine. For a small to medium size room I can highly recommend the NaO.
Mine is the original NaO that uses a single 12" XLS crossed at 100Hz. John K since replaced the 12" XLS with one or two 10" XLS and moved the XO point to 120Hz. This was one stone to kill two birds: (1) with a high XO point the mid-woofer is less likely to be bottomed so the overall SPL of the system can be higher; (2) With a single or dual 10" woofer(s) the U-frame was made shorter which helps to raise the first resonance frequency of the U-frame. Because mine is the 12" woofer with a long U-frame I found the first resonance frequency can not be damped fully. I reduced the setting over that range and this does not make the music 100% smooth when crossing between the woofers and the midranges.
I guess the NaO II has addressed the problems I have with the original NaO, but still, mine is not a NaO II and I don't expect the NaO II can play at the SPL level I want. Interestingly, John K said that he never found the U-frame bottoming out on his single 10" XLS and I was the first one in the world to suggest dual 12" for the NaO.
I believe John K has published a lot of information about the NaO including the above before so I think the above does not disclose (enough) technical information about the NaO to make John K upset about it.
That was why I asked if you have considered the 26W8861. Of course at the time I had not read your entire text and thought you were going to use them as woofers. By the way, I have read all of your posts now and thank you so much for sharing your knowledge and experience with us.
I have not connected the rear tweeters yet so I can't comment on them.
I have to say the passive XO of the original NaO was only averaged to good but the new passive XO is superb! It is the low loss and very smooth integration of the XO give the NaO outstanding sound. I tuned the MTM part of my original WWMTMWW speakers for two years to make it near perfection and I have to say the NaO is certainly not lesser. Therefore, I have basically abandoned my thoughts on using the Esotar T330D on the NaO because I don't want to destroy the best part of the NaO. It appears the Excel T25 may be sonically more transparent than the Esotar and it is the last driver to be bottomed in the NaO so I think I will keep the arrangement as it is. Not that the Esotar is less better given its practically nearly unlimited power handling capability I found musicality and smoothness in the Esotar possibly better than the T25. The NaO panel above 150Hz, I would say it is very difficult to beat and it can play fairly loud. What I want is simply to remove any resonance in the U-frame and improve the SPL capability under 150Hz substantially.
Having had the WWMTMWW I have to say I love speakers capable of doing high SPL. Whether I play them loud or not is a different story but they just give you the uncompressed sound and can deliver as much power as you want at your wish. When an amplifier clips we can easily tell. When the drivers run of linear excursion the sound just appears to be compressed.
With regards to your directivity / waveguide, I do find the tweeter energy is stronger at my sitting position. But then which speakers don't have similar problems? At high frequencies it beams and that is physics.
As a matter of fact I am going backward and will experiment using a sealed box for 100Hz down. This is because my room has the first room mode at 2X Hz so a dipole or monopole should make not much difference. Of course, if I mount 4 x 12" U-frame per side and make the U-frames shorter and raise the XO point then it might just do it for me, just. Shortening the U-frame may decrease the SPL by, say, 6dB while using 4 only gives a 12dB gain so at the end I only get 6dB gain, which is not much. I won't be crazy enough to use 8 x 12" per side. Oh, this is the only evil of dipole bass.
Some people commented that low excursion stiff cone drivers sound much more musical in bass so I guess I will try a simple sealed box with 2 x 26W8861 doing push-pull in 9x litres with a Q of about or a bit below 0.7. I think I will use active EQ to boost the bottom a bit. This should give me sufficient SPL I want. What do you think?
More importantly (please correct me if I am wrong), with a U-frame the first resonance is at 1/4 wavelength. With a sealed box it is at 1/2 wavelength. In other words, currently my U-frame with an XO at 100Hz has a first resonance at about between 150Hz to 200Hz (I don't want to disclose John's design) that can only be partially damped. If I use a sealed box, the first resonance will be around or above 300Hz. XOed at 100Hz I believe the sealed box will give a very clean sound comparing to the U-frame, especially given that my room mode is as low as 23Hz, for a sealed box should work.
Sorry I have stolen your space to ask you questions. But I guess we are sharing our experience with our speakers that have basically the same design. Of course, you are far more advanced than me.
Regards,
Bill
MBK said:
I found some room acoustic lecture material that suggests all below 35ms arrived sounds are blended into a direct sound in human ear. This sounds a bit much on most environments. But then again I've heard plenty of recommendations for dipolepanel to be at least 2m from frontwall and even prefering >4m distances. 4m would give rear radiation 24ms delay before reaching listeners ears. Below 2m distances are caused effects where rear radiation is "glued" on to the direct sound and mess it up.
There are plenty of practical examples of this. Imagine yourself listening a live chorus. Your listening position isn't perfect and so you aren't in focus with all singers. With large chorus even few meters distance difference are very possible. You can't pinpoint individual singers or groups of singers just by the fact that some of them are a bit futher away from you than others. Unless one of them is singing a solo... But you also get blurred singing. Even large chorus recordings suffer from this. Chorus sounds united and whole but words aren't distinct. Ear can't separate singers but their different timing smeares the sound and change the tone.
I also found another lecture material that says human can spot sounds more precisely with direct sound and reflections below 1ms delay. With room + speaker this means diffractions from cabinet edges which is the first problem. After that ear summs first reflections and direct sound up to 40ms delay and only after that ear can really determine which sound is reflected and which is direct sound. In most rooms reverberation time lasts up to 500ms or so. Within that 1-40ms time sensed main direction for sound is still there but the tone changes. Ear mixes them all up, different phases, different frequences from different angles, different reflective materials reflect different frequences and all this is multiplied if speakers radiation pattern against first reflections isn't even. In most cases dipole speakers rear radiation is also included in <40ms delaytime unless the room is huge. This is why I'd use some heavy diffusing for the rear radiation. Just to make the room "bigger". Unfortunately this isn't possible. Broadband diffusor that would really do some good is basically size of a bookshelf and I can't have one on my front wall. But I have it at the rear wall which is the other problem.
Actually listener position is also crusial in this aspect. Listening seat near the rear wall and first reflections from there arrive your ear with very short delay. In my case I can use Linkwitz recommended dipole positioning advice and have equal space from listening spot to rear wall and from speakers to front wall. Still both direct sound bounced off the rear wall and rear radiation bounced from front wall come to my ears with aprox 9ms delay. Hope that's enough. At least it's helluva lot more than reflections from sidewalls or the floor. Therefore I need the directivity to minimize the problem there...
Against floor bounce dipole directivity just isn't enough. Even without any baffle (vertically floorstanding panel is pretty "wide" and therefore not that much of a dipole) dipoles natural -3dB at 45 degrees isn't nearly enough. And with normal listening distances and heights 1st reflection arrives below 40 degree angle. Even my +-45 degree simulations are too optimistic.
But don't take this too seriously. I've heard few systems with poor positioning (in these aspects) and have listener just along the wall and they sound fine. But I don't think that's even the question here. The question is: can it be better? Of cource. How? Proper positioning, directivity and diffusive material at critical spots.
In actual live situation this doesn't mather that much. You can consider live concert to be a piece of art itself so diffractions and all are part of the performance. This also applies to two people talking in poor acoustic environment. Even while mouth radiation pattern isn't that ideal we can still hear what people are saying. But with recorded material where primary objective is to deliver direct sound to listener as much as possible this isn't ideal solution. Far from it.
But then again recordings aren't made as 100% eardrum material. Most recordings rely on the fact that there are first reflections (below 40ms barrier) which keeps the music alive. Unideal I know but so is two channel stereo as well. 🙂
U-baffle isn't that good cardioid. But then again I don't know how good cardioid it should be. I've heard one very directive speaker where floor reflection is optimized and overall radiation pattern is hypercardioid. And I can say they are extremely critical for recordings. Most albums just aren't meant for such playback. This is also one reason why I selected basically less ideal open baffle with its problems over technically better solution. I want to enjoy the music.
That's based on experience. But depends on what you're listening to. With overall balance accurate directivity isn't the problem. System can still be directive. But its radiation pattern should allow gentle roll off in power response towards the treble which leads to gently rolling off room balance. Overally system can still be constantly directive. Fullrange monopole is one sample of constant directivity. Increasing directivity from there lets more direct sound to your ears but going too far kills the music or calls for more playback channels to work.
Personally I think the reverberation field makes the live and airy sound. First reflections just mess up the accuracy. This is why diffusing is a better why improving things than damping what they do in most studios just to get that ultra-dry monitoringsound. Diffusing just diffuses reflections around the room and they arrive to listeners ears later creating illusion of bigger room.
But there isn't much chances to place a bookshelf or such on the floor. To both sides it's possible. And the only solution is vertical directivity. Once again. Horizontal problems are easier to manage and treat.
Yep. But it still doesn't explain why this didn't show up in anechoic measurements with adequate measuring distance.
Good looking measurements. Not that much directivity thought. Have you tried to measure them from 180 degree angle to see how well behaved cardioid they are?
Jussi
HiFiNutNut said:
Ok. Sounds like a problem of a big room. But you can also consider yourself lucky. A big system in a big room can sound superb. There are some unsolved problems in smaller rooms. Mainly created by the nearby surfaces.
Naturally. I guess NaO is designed for smallish to midsize rooms.
Sounds like NaO limitations are just where I'd expect them to be. Lower midrange (just above lower crossoverpoint) and naturally the bass (the headache in most systems).
So just how much difference for example addition of 2x10" below 300Hz and adequate displacement woofersystem create here?
Know the feeling. Easy, uncompressed, nontiring and effortless at any level I can think of. Cool. Even without loud levels. With modest levels overkilled systems just play what they are fed hands down.
Actually a bit under a year ago I gained experience of actual compressing. I have these albums that I use to evaluate systems. I have listened them a lot so I pretty much know how they sound in different systems. I visited Finnish studiomanufacturer Genelec and tried their biggest model 1036A. The place is surreal. 60-70m2 space with adequate acoustics, build for those speakers alone. After listening I got thinking how damn much normal home systems actually compress dynamics. Even 1036A is pretty reserved while distorsion is very low and room acoustics rather dry but they don't hold back. Good old Jurassic Lunch was pretty terrifying. That T-rex really bites!
This is just one funny thing in this business. People complain a lot about compressed recordings and still most systems are incapable of reproducing them if there is some. I did some analysis on Gustav Mahlers 6th symphony part four. Lowest sound the program found is around -70dB level and usable dynamics range is about 55dB. Now that's heavy. With normal room humming (30dB) this means about or even over 100dB SPL levels to get those smallest details without sinking them into the humming. And darn that symphony is pretty brutal to listen as well. I don't know can speakers ever reproduce it but I'd like to try. Just to honor symphonic masterpieces.
Actually uncompressed soloinstruments have pretty large dynamics range alone. Here is one very useful site about this:
http://www.moultonlabs.com/more/loudspeaker_as_musical_instrument/P2/
Around center of the page is a chart with different instruments. Solopiano for example has 40-100dB range at 3m distance. And like you know live piano is pretty aggressive instrument. But how many times you hear something like that from CD? I know it isn't CDs fault since it's capable of storing piano and such without compression. But is it the album buyers. Again you can't listen to uncompressed solopianoalbum from your kitchen radio and cook dinner at the same time. It needs attension.
That beaming problem can be fixed with certain type of waveguide. But like said earlier, most recordings don't support such constant directivity with flat power response. It's just way too bright. With those 50's and 60's recordings it works. Depends on what you're listening.
Yep. Down to 30Hz it's pretty easy but then it takes the plunge.
Sounds reasonable. But is SPL capacity adequate? 660cm2, your WWMTMWW had 880cm2. Was that adequate?
Sounds correct.
No problem. I don't want to limit this thread into my project only. Different projects can offer new perspectives.
Jussi
About the reflection vs. diffusion vs. reverb times:
There are several things happening with reflections. At first, the ear is not capable of separating source and delayed sound at all. This causes smear and goes until 0.5-1 ms. From about 1.5 to 10 ms we "know" that a reflection is there, but it is subconscious - in this range, reflections are desired and add air. In the 10-50 ms early reverb range we still can't separate the direct sound from the relfected one, but too much reflections/reverb here does disturb intelligibility. Above 150 ms the reverb again contributes positively, echoes are perceived as echoes, and this range is important for bass envelopment. There was some discussion on this before and I wrote some more on it
here
I learned much from the Griesinger articles on SL's site and others on Griesinger's own site I suppose you know those. I had another article that detailed the figures I mentioned above but I can't find it now...
As for placement 1.2 m from live side wall is the absolute minimum in my experience. More would be better I suppose. And to take advantage of the side reflections you are right, diffusors are the best option, rather than absorbers.
As for my measurements I have to do a full 180 degrees one day to check what is really going on. The total power response must be good, because the off axis tonal balance is good and doesn't change when walking around - except directly behing the speaker the tweeter is significantly cut. I am quite confident it won't be a "real" cardioid, but as you saw in my figures there is surprisingly little horizontal falloff in the midrange, and what's more, only at wavelengths much longer than the U depth does the sound "ignore" the U (say >1 m wavelength is needed to ignore my ca. 20 cm U depth).
There are several things happening with reflections. At first, the ear is not capable of separating source and delayed sound at all. This causes smear and goes until 0.5-1 ms. From about 1.5 to 10 ms we "know" that a reflection is there, but it is subconscious - in this range, reflections are desired and add air. In the 10-50 ms early reverb range we still can't separate the direct sound from the relfected one, but too much reflections/reverb here does disturb intelligibility. Above 150 ms the reverb again contributes positively, echoes are perceived as echoes, and this range is important for bass envelopment. There was some discussion on this before and I wrote some more on it
here
I learned much from the Griesinger articles on SL's site and others on Griesinger's own site I suppose you know those. I had another article that detailed the figures I mentioned above but I can't find it now...
As for placement 1.2 m from live side wall is the absolute minimum in my experience. More would be better I suppose. And to take advantage of the side reflections you are right, diffusors are the best option, rather than absorbers.
As for my measurements I have to do a full 180 degrees one day to check what is really going on. The total power response must be good, because the off axis tonal balance is good and doesn't change when walking around - except directly behing the speaker the tweeter is significantly cut. I am quite confident it won't be a "real" cardioid, but as you saw in my figures there is surprisingly little horizontal falloff in the midrange, and what's more, only at wavelengths much longer than the U depth does the sound "ignore" the U (say >1 m wavelength is needed to ignore my ca. 20 cm U depth).
MBK said:
This sounds like a good argument why you need to do something against the floor reflection. It isn't easy thought. Even and broadrange vertical directivity would require tall system with plenty of drivers. With 2,5m listeningdistance and 0,9m tweeter / ear height floor reflection arrives just 1,5ms later than direct sound and without any directivity to that angle it changes quite a bit. But like said, it doesn't sound as bad as it looks.
I agree. Unfortunately I can't have that much space to sides. Or maybe if listening direction is turned 90 degrees but that moves speakers much closer to frontwall and listener just along the rear wall. Both are concrete while present frontwall is mainly window and a door and rear wall thin with 1/3rd wide doorway so I think it's still working much better in this direction. 30 degree toe in (panels towards listener) moves first sidewalls reflections way off axis from the speaker and directivity should help there.
But there is a significant difference between U-baffle structure and dipole structure in this aspect. U-baffle and such need sidewall reflections to sound "live". Dipole has the same from its rear radiation. Both can have plenty of delay to their reflections but positioning differs significantly. Dipole works better closer to sidewalls but needs plenty of space behind it. U-baffle can be closer to frontwall but needs space to its sides.
Fullrange cardioid or hypercardioid is very uniform in different directions. Basically the balance can be very even even at large angles. And listening to them 180 degrees angle sound appears to come from wall behind the listening spot. There is enough 180 degree directivity that the front radiation bounced from the wall sounds much more powerful than the 180 degree radiation from the speaker itself.
Jussi
Jussi,
Thank you very much for your point about cardioid vs dipole in relation to rear and side wall positioning. I hear a distinct difference and kept thinking it was the cab, not placement.
One thing that seems left out as those, like yourselves, venture into more complex open driver systems is addressing problems using the shape of the construction itself to taylor response. For example, resonances in HiFiNut's U-baffles can be cured to a great extent by the shape of the U-baffle and its terminus. Also, the ripple effects of edge diffraction seem to be ignored in all open designs even though they are much greater than with monopole.
I've tried to stimulate discussion elsewhere regarding edge diffraction, but haven't found the right group. Yes, the frontal size and shape of the baffle can smooth the response, but I believe this only averages the ripples together by spreading them across a wider frequency range to generate a smoother overall response. I've obtained quite audible sonic gains by changing the shape and dimensions of the baffle edge, without changing the front view shape. I don't want to average a bunch of garbage together to get a smooth landfill. I want to eliminate the garbage completely, and I find Olsen's work in this area very helpful.
Thank you very much for your point about cardioid vs dipole in relation to rear and side wall positioning. I hear a distinct difference and kept thinking it was the cab, not placement.
One thing that seems left out as those, like yourselves, venture into more complex open driver systems is addressing problems using the shape of the construction itself to taylor response. For example, resonances in HiFiNut's U-baffles can be cured to a great extent by the shape of the U-baffle and its terminus. Also, the ripple effects of edge diffraction seem to be ignored in all open designs even though they are much greater than with monopole.
I've tried to stimulate discussion elsewhere regarding edge diffraction, but haven't found the right group. Yes, the frontal size and shape of the baffle can smooth the response, but I believe this only averages the ripples together by spreading them across a wider frequency range to generate a smoother overall response. I've obtained quite audible sonic gains by changing the shape and dimensions of the baffle edge, without changing the front view shape. I don't want to average a bunch of garbage together to get a smooth landfill. I want to eliminate the garbage completely, and I find Olsen's work in this area very helpful.
johninCR said:
Yep. But I don't know is the effect that similar to dipole and cardioid sources. More diffusing sound from sidewalls + ceiling and floor and less from front wall against less diffusing sound from sidewalls + ceiling and floor and more from front wall. Seems like many dipole systems (planars, dynamic open baffles etc) get praised from their imaging depth. But is cardioid type better in horizontal level and have more lively width in its imaging?
Oh well. That's just a thought based on reflection direction and common sense. I guess speaker positioning has a big effect in these things. Different sound sources just need different positioning to work optimally.
I agree. Maybe it has something to do with mensioned "good enough" parameter. While open baffle sounds better or at least different than any monopole it creates a state of happiness and doesn't encourage to push it even further. Perhaps.
At least this kind of R&D would require a lot of prototypes and high quality measurements. I know few projects that have really put some effort to get the perfect open baffle done.
For smoothest response and most uniform radiation pattern in open baffle the best solution would be no baffle at all. Just driver hanging in air. Maybe Orion revision style magnetic installation could be used. Finnish Gradient had their 1.1, 1.2 and 1.3 models with such midrange. They feature 8-10" closed woofer, 10-12" fully open baffle dipole midrange and a four dometweeter array. They went out of production sometime in the 90's. Here is some info and few pics from that design:
http://www.regonaudio.com/Gradients.html
Those speaker have some kind of cult status here in Finland. People just fix them if surrounds or such are broken and just keep on listening even while the oldest models are about twenty years old.
Basically open baffle system is easier to design that a cardioid system. There are much more diffraction problems, resonances and such in the cardioid to take care of. But if done properly results are very good. Dipole works like trains toilet, easy and guaranteed. Naturally there are certain flaws that can't be ignored. One is the 1/4th wave resonance and the dipole nulls that appear somewhere in the midrange depending on baffle size. Without any baffle there isn't such problems, there is just a smooth transition from dipole directivity to cone area directivity. But fully open drivers efficiency really sucks. We measured a single 7" Peerless HDS without any baffle and it has about 77dB efficiency at 300Hz. So it needs larger drivers or more than one to work. But even the dipole nulls doesn't seem to be that big problems in actual systems.
Some time ago I actually draw such fully open WMTMW configuration. It would require separated woofers to actually play fullrange but in radiation pattern vice it should be pretty good. Check the attachment. Looks exotic and good to be but I never got the physical stability problems solved without running into visual appearance problems. So I didn't see sense to build even a prototype if the system is so ugly there isn't a future for it. Shame.
Jussi
Attachments
Jussi,
Funny that you mentioned no baffles at all. Here's a test I ran 2 weeks ago. Quite a few laughed at the idea, but with active XO & EQ it worked surprisingly well. 15" drivers is all it takes. I used a coax on top and woofer on the bottom. The real flaw was that early reflection off of the support structure. I'm just about finished with a magnet mount support structure and will try some ideas wrt diffraction rings to see if I can make an 8" driver work with no baffle.
Here's what is in process. The 2 supports at the bottom are for the 15" woofer whose structure is isolated.
Funny that you mentioned no baffles at all. Here's a test I ran 2 weeks ago. Quite a few laughed at the idea, but with active XO & EQ it worked surprisingly well. 15" drivers is all it takes. I used a coax on top and woofer on the bottom. The real flaw was that early reflection off of the support structure. I'm just about finished with a magnet mount support structure and will try some ideas wrt diffraction rings to see if I can make an 8" driver work with no baffle.
An externally hosted image should be here but it was not working when we last tested it.
Here's what is in process. The 2 supports at the bottom are for the 15" woofer whose structure is isolated.
An externally hosted image should be here but it was not working when we last tested it.
Jussi,
I think your attached image is absolutely great! Being a fan of focused line array implementations, your concept looks like a killer set-up. With the correct time alignment and decent drivers, it should probably have stunning clarity and definition. The use of an open baffle also lends itself to an open, natural and live presentation.
Can you build a mockup with a wood frame and threaded rods and bolts to get the acoustic setback/ time alignments? I would think that once you have your array "dialed in" you could take measurements and contruct a permanent and attractive structure.
JohninCR,
Your pictures are an inspiration! My eyes aren't what they were 30 years ago, so tell me: Is the mounting block surrounding the magnet a through bolted split block? It seems that you've hit on a very simple, yet articulate system for adjusting your drivers' position. Well done!
It may be that I'll have to try something like you've done, if for no other reason than it looks like a fun project.
Let's see now, where did I put that box of spare drivers?
Both of you have given me something new to think about. Thanks!
Best Regards,
TerryO
I think your attached image is absolutely great! Being a fan of focused line array implementations, your concept looks like a killer set-up. With the correct time alignment and decent drivers, it should probably have stunning clarity and definition. The use of an open baffle also lends itself to an open, natural and live presentation.
Can you build a mockup with a wood frame and threaded rods and bolts to get the acoustic setback/ time alignments? I would think that once you have your array "dialed in" you could take measurements and contruct a permanent and attractive structure.
JohninCR,
Your pictures are an inspiration! My eyes aren't what they were 30 years ago, so tell me: Is the mounting block surrounding the magnet a through bolted split block? It seems that you've hit on a very simple, yet articulate system for adjusting your drivers' position. Well done!
It may be that I'll have to try something like you've done, if for no other reason than it looks like a fun project.
Let's see now, where did I put that box of spare drivers?
Both of you have given me something new to think about. Thanks!
Best Regards,
TerryO
TerryO said:
Thanks Terry.
I wanted the mount to be adjustable for different drivers and/or baffles. For different drivers I just need new half circles. Cut a rectangle out of the diameter equal to the size of the magnet, drill the hole, and then (yes) cut the half circle in half to form the correct sized clamp. I cut the bases for the main driver and woofer from the same piece of wood for a good fit and matching grain, so they should look really nice once I rub some oil into them.
johninCR said:
Why did you used the woofer in "no baffle" environment as well? Advantage is in the midrange and I think the woofers panel should be as wide (or deep or what ever depending on frame shape) as needed top extension allows.
My rendering is a 3-way top for a 4-way system. Those 10" woofers can't play bass and they are supposed to work with upper mids with gentle slope somewhere around 300-500Hz.
TerryO said:
I did some scetches with spine and floorplate but they didn't seem even nearly promissing. Such system is just so high which makes it physically unstable and with huge and stable spine it just ruins its advantages.
It would require separated woofers for the actual bass, drivers that allow gentle crossover slopes and propably a nonsymmetric waveguide for the tweeter to match it in vertical plane. Perhaps my 25Ws would have worked as lower mids, I asked Audio Technology for a high efficiency short throw midrange drivers but I didn't have to guts to start executing it.
And further more there appeared some WAF problems with that design. Seems like my speakers can't draw too much attension. I don't know why. Personally I don't like box like ordinary speakers while there is a chance to make something different. Perhaps something like a sculpture.
Jussi
- Status
- Not open for further replies.