@ Ken, i should have guessed, seeing the stone wall in your picts 
@ fluid, thx !, nice offer....might want to take you up on some kind of guide if i can find a small enough tweeter (with big enough SPL). Search has not been encouraging so far...
@all...
So i was toying around with the 3fe25's, trying to see the tightest they could pack together. With the idea I'll look for a tweeter that fits in the resultant center space...just to keep the experiments going.
Well, when i tried to tight pack 4 of them into a square on a flat surface, one of them jumped over an inch out of the tight pack!
Then i found even a single pair noticeably pushes away from each other, just not strong enough to force movement.
So now I'm wondering if closely spaced drivers screw up the magnetic fields of each other, with regard to the field the coil operates in?
Out of my depth here ....
@ fluid, thx !, nice offer....might want to take you up on some kind of guide if i can find a small enough tweeter (with big enough SPL). Search has not been encouraging so far...
@all...
So i was toying around with the 3fe25's, trying to see the tightest they could pack together. With the idea I'll look for a tweeter that fits in the resultant center space...just to keep the experiments going.
Well, when i tried to tight pack 4 of them into a square on a flat surface, one of them jumped over an inch out of the tight pack!
Then i found even a single pair noticeably pushes away from each other, just not strong enough to force movement.
So now I'm wondering if closely spaced drivers screw up the magnetic fields of each other, with regard to the field the coil operates in?
Out of my depth here ....
Thanks Ken, good to know.
It is amazing to feel how much the drivers want to push away from each other...glad it doesn't matter.
Thanks Jack, I already had one in my PE cart.
It's about the only thing that will fit inside the tightest 4-pack of 3fe25's will reasonable SPL. (i'll have to cut a mounting tab off each 3fe25.)
Might get a handful of the little Peerless just to play with them alone..
It is amazing to feel how much the drivers want to push away from each other...glad it doesn't matter.
Thanks Jack, I already had one in my PE cart.
It's about the only thing that will fit inside the tightest 4-pack of 3fe25's will reasonable SPL. (i'll have to cut a mounting tab off each 3fe25.)
Might get a handful of the little Peerless just to play with them alone..
Just remembered that these Genelecs are advertised as point sources, doesn't get much smaller than this: The Ones - Genelec.com
Another is Danley Studio 1 DANLEY BREAKS THE MOLD ONCE AGAIN: INTRODUCING THE NEW STUDIO 1 AND STUDIO 2 NEARFIELD MONITORS AND STUDIO SUBWOOFER – MusicPlayers.com
Both using coaxial driver.
Mark, I like your guerilla style! Best way to learn is just doing it
Another is Danley Studio 1 DANLEY BREAKS THE MOLD ONCE AGAIN: INTRODUCING THE NEW STUDIO 1 AND STUDIO 2 NEARFIELD MONITORS AND STUDIO SUBWOOFER – MusicPlayers.com
Both using coaxial driver.
Mark, I like your guerilla style! Best way to learn is just doing it
Haha, guerilla style....
But you know, i keep finding the devil is in the construction details...it often comes down to theory be damned, what can i actually build.
And measurements are like being under live fire vs target practice at the range.
But seriously, yeah, Studio One and other small point sources have to rock.
Heard the Studio One..very nice. Although, only good to a certain SPL.
But you know, i keep finding the devil is in the construction details...it often comes down to theory be damned, what can i actually build.
And measurements are like being under live fire vs target practice at the range.
But seriously, yeah, Studio One and other small point sources have to rock.
Heard the Studio One..very nice. Although, only good to a certain SPL.
Not to put you off trying but a smaller tweeter that can't reach low and has no directivity doesn't seem like it will be anything other than a different compromise.
This is the same issue that MTM configurations suffer from, getting a tweeter that can go low enough for the distance to the mids.
Have you thought about a SEAS DXT or one of the combinations of waveguide and tweeter here Heissmann-Acoustics | Tweeter with Waveguide | reviews that can be bought off the shelf?
Agree
I used DXTs some time ago, but prefer a larger and more conventional WG, hence I studied the Heissmann-acoustics site, as well as some discussion of WGs and tweeters on an advice page on soundimports.eu, before selecting the combination I chose for my latest array (different tweeter, mine have ferrofluid, but similar geometry).
For a few days I've been musing on the following ideas:
What I dislike most: sound of a 1" dome crossing to a 6-8" mid on a narrow-flat-baffle speaker.
Flaws of that approach include the off-axis radiation and diffraction from the low-end of the tweeter. A WG helps with this, so for me, it's important to have control down to the low-end of the tweeter. That rules out the small tweeter in minimum ring approach that Mark wants to explore.
The consequential flaw with my latest array is that the inner ring is too large to meet the tweeter+WG coverage, and I see a dip in the off-axis response (in angles out to the next lobes which are somewhat blunted on the horizontal axis by the ring shape and orientation, and choice of crossover).
I suspect that this flaw - an off-axis dip - is less distracting to the brain trying to decode the sound than the usual flaw described above - that leads to strong reflections all round the room in a critical frequency band.
Even as I play around with the crossover, keeping the on-axis response relatively flat, but letting the off axis dip vary, I'm not aware of any substantial tonality changes or imaging changes. It's all of the type that my brain adjusts to in a moment, and the same goes for listening to varied music over many days.
Ken
I used DXTs some time ago, but prefer a larger and more conventional WG, hence I studied the Heissmann-acoustics site, as well as some discussion of WGs and tweeters on an advice page on soundimports.eu, before selecting the combination I chose for my latest array (different tweeter, mine have ferrofluid, but similar geometry).
For a few days I've been musing on the following ideas:
What I dislike most: sound of a 1" dome crossing to a 6-8" mid on a narrow-flat-baffle speaker.
Flaws of that approach include the off-axis radiation and diffraction from the low-end of the tweeter. A WG helps with this, so for me, it's important to have control down to the low-end of the tweeter. That rules out the small tweeter in minimum ring approach that Mark wants to explore.
The consequential flaw with my latest array is that the inner ring is too large to meet the tweeter+WG coverage, and I see a dip in the off-axis response (in angles out to the next lobes which are somewhat blunted on the horizontal axis by the ring shape and orientation, and choice of crossover).
I suspect that this flaw - an off-axis dip - is less distracting to the brain trying to decode the sound than the usual flaw described above - that leads to strong reflections all round the room in a critical frequency band.
Even as I play around with the crossover, keeping the on-axis response relatively flat, but letting the off axis dip vary, I'm not aware of any substantial tonality changes or imaging changes. It's all of the type that my brain adjusts to in a moment, and the same goes for listening to varied music over many days.
Ken
There was a Vanderkooy study where they varied the power response of a speaker and had listeners test the different responses. The conclusion was that dips in the power response were fairly benign but peaks were an issue which would seem to be in agreement with your anecdote.
Hadn't seen the Heismann site...thx.
The SEAS DXT looks very good. I had been looking at it, and the Morel CAT378, when i sprung for the SB26. Not sure why i picked the SB26...relooking i'd probably have tried the Morel for that first hail Mary ring build .....
Totally agree that putting a small tweeter in the middle of some 3 inches is nothing but a different type of compromise.
But i'm not trying to build a speaker right now. First ring attempt was as much about building a starting test bed as it was hoping it would possibly work.
Now, i'm trying to define ways to make measurements that teach me what c-to-c spacing is truly about.
You know, what really matters for coherent off-axis coupling...is it 1/4 wave, half wave, etc? is it the same for a ring as for a line? ........etc
As you say, it's the same issue for MTMs....heck it's the same issue for all speakers I guess... from lines to synergies to planars to everything in between, huh?
I guess this is where simulations are more appropriate...but like always, if i can experiment and learn on the cheap, just my preference.
Just remembered that these Genelecs are advertised as point sources, doesn't get much smaller than this: The Ones - Genelec.com
Another is Danley Studio 1 DANLEY BREAKS THE MOLD ONCE AGAIN: INTRODUCING THE NEW STUDIO 1 AND STUDIO 2 NEARFIELD MONITORS AND STUDIO SUBWOOFER – MusicPlayers.com
Both using coaxial driver.
Mark, I like your guerilla style! Best way to learn is just doing it
P-Audio :: P-Audio SN6-150CX PS70.46 IN STOCK (16 Sep 2020)
Wonder if its done with p audio 6.5"
Searching ...
Thanks - I wasn't aware of that study. The Vanderkooy paper I've read is with Lipshitz: J.AudioEng.Soc.,Vol.34,No.4, 198, on a related topic (and I keep having crossovers in quadrature in mind when making adjustments).
I looked at
John Vanderkooy's research works | University of Waterloo, Waterloo (UWaterloo) and other places
but the study you mention didn't jump out. Do you happen to recall where it's reported?
Ken
Thanks - I wasn't aware of that study. The Vanderkooy paper I've read is with Lipshitz: J.AudioEng.Soc.,Vol.34,No.4, 198, on a related topic (and I keep having crossovers in quadrature in mind when making adjustments).
I looked at
John Vanderkooy's research works | University of Waterloo, Waterloo (UWaterloo) and other places
but the study you mention didn't jump out. Do you happen to recall where it's reported?
Ken
I think it is this one
Power Response of Loudspeakers with Noncoincident Drivers-The Influence of Crossover Design
Vanderkooy, John; Lipshitz, Stanley P.
JAES Volume 34 Issue 4 pp. 236-244; April 1986
I don't have a copy and I haven't been able to find it easily online,
If you put Vanderkooy power response hole into google it should find some relevant threads where the conclusions have been discussed by some reputable people
Power Response of Loudspeakers with Noncoincident Drivers-The Influence of Crossover Design
Vanderkooy, John; Lipshitz, Stanley P.
JAES Volume 34 Issue 4 pp. 236-244; April 1986
I don't have a copy and I haven't been able to find it easily online,
If you put Vanderkooy power response hole into google it should find some relevant threads where the conclusions have been discussed by some reputable people
Fluid, thanks.
I can't find that paper either, but did follow your advice and found several threads mentioning the outcomes, including some I read here, most likely when they were active in 2011.
The general message seems to be along the lines:
Lipshitz and Vanderkooy observed that peaks in power response are audible if high enough and there's a range of power responses with dips that are least "disturbing". No power response curve has been established as correct.
I assume, from the title and partial abstract, that's mainly referring to the crossovers, so in the frequency range of interest in this thread.
"The acoustic power response of a loudspeaker having separated drivers is analyzed using mainly a simple monopole model. It is shown how different crossover types cause excess or suppressed power response at the crossover frequency. By proper choice of phase relationships between drivers one can achieve least coloration of the reverberant response,..."
there it is "coloration" which does hint at a subjective evaluation.
Ken
I can't find that paper either, but did follow your advice and found several threads mentioning the outcomes, including some I read here, most likely when they were active in 2011.
The general message seems to be along the lines:
Lipshitz and Vanderkooy observed that peaks in power response are audible if high enough and there's a range of power responses with dips that are least "disturbing". No power response curve has been established as correct.
I assume, from the title and partial abstract, that's mainly referring to the crossovers, so in the frequency range of interest in this thread.
"The acoustic power response of a loudspeaker having separated drivers is analyzed using mainly a simple monopole model. It is shown how different crossover types cause excess or suppressed power response at the crossover frequency. By proper choice of phase relationships between drivers one can achieve least coloration of the reverberant response,..."
there it is "coloration" which does hint at a subjective evaluation.
Ken
Hi guys,
Been feeling a need to follow up here...
I totally stopped with any speaker building efforts, and just started measuring multiple driver arrangements on a flat baffle. Pairs, tight pack quad, tight pack ring, etc. All done with same drivers already being used.
Idea was to test the rules of thumb for myself,.... on C-to-C spacing for combing, and also look into off-axis response.
Didn't really learn anything not commonly discussed, other than becoming increasingly disillusioned with multiple drivers covering the same freq range.
But i did see it could work quite well, if you have a lot of time to experiment and dial in.
Like Ken did ! (I might need a decade or so too !)
Moving back into theoretical design.
Like said earlier, my interest in the ring arrays was to see if a "flattened synergy" could be made, maybe with superior band blending, and C-to-C spacing.
It seems now, that flattening a cone can solve any front to back depth spacing between driver centers, but at the expense of widening C-to-C centers horizontally and vertically.
Flattening forces outward radial expansion of acoustic centers.
Maybe the best vector compromise without time delay adjustment would be a 90 deg cone?
But with time delay adjustment, maybe optimum becomes tighter if you can stand the the narrower pattern.
Dunno, thinking out loud....
Anyway, I've been moved back in to some further synergy development.
Especially since i tend to prefer a tighter pattern.
Despite being happy as can be with last syn7 build, i've got some refinement ideas from these ring-array experiments i want to try.
Been feeling a need to follow up here...
I totally stopped with any speaker building efforts, and just started measuring multiple driver arrangements on a flat baffle. Pairs, tight pack quad, tight pack ring, etc. All done with same drivers already being used.
Idea was to test the rules of thumb for myself,.... on C-to-C spacing for combing, and also look into off-axis response.
Didn't really learn anything not commonly discussed, other than becoming increasingly disillusioned with multiple drivers covering the same freq range.
But i did see it could work quite well, if you have a lot of time to experiment and dial in.
Like Ken did !
(I might need a decade or so too !)Moving back into theoretical design.
Like said earlier, my interest in the ring arrays was to see if a "flattened synergy" could be made, maybe with superior band blending, and C-to-C spacing.
It seems now, that flattening a cone can solve any front to back depth spacing between driver centers, but at the expense of widening C-to-C centers horizontally and vertically.
Flattening forces outward radial expansion of acoustic centers.
Maybe the best vector compromise without time delay adjustment would be a 90 deg cone?
But with time delay adjustment, maybe optimum becomes tighter if you can stand the the narrower pattern.
Dunno, thinking out loud....
Anyway, I've been moved back in to some further synergy development.
Especially since i tend to prefer a tighter pattern.
Despite being happy as can be with last syn7 build, i've got some refinement ideas from these ring-array experiments i want to try.
Mark,
I've not been visiting my thread, hence the delay.
I keep going round a circle of thoughts - all in the context of speakers that must work close-against (but not in) a wall - requiring wide, shallow speakers. This is the best use of my listening space.
What are the options?
A suitable large, shallow waveguide is impractical/unobtainable (if a good solution even exists, about which I'm not convinced). 90-degree WGs (or synergies) for use in corners could work, but I can't do that either - the room is not a good shape for that either.
Another option might be a 15" co-axial, but there would still be problems to deal with (their depth is challenging, diffraction at internal discontinuities, ...).
My last arrays might be expected to be poor in terms of horizontal dispersion, but every time I compare with the earlier version which measures flatter horizontal angles >45 degrees, I prefer them on a wide range of music.
It's hard to make measurements to justify my observation, but my impression is that the defects in the room - which make the usual case of excess power response around 1-2 kHz sound particularly bad - complement the dip in power response. After all, the response is good out to >30 degrees with no eq on the 3FE25s.
What might I do different?
There could be a compromise with closer horizontal packing of the mids, between the 1:1 H:V ratio of the last design and the ~4:1 of the AMT. That would need a tweeter or tweeter+WG with the right properties, and - ruling out diffraction slots from past experience - I've not found a better solution.
Have fun with your synergies!
Ken
I've not been visiting my thread, hence the delay.
I keep going round a circle of thoughts - all in the context of speakers that must work close-against (but not in) a wall - requiring wide, shallow speakers. This is the best use of my listening space.
What are the options?
A suitable large, shallow waveguide is impractical/unobtainable (if a good solution even exists, about which I'm not convinced). 90-degree WGs (or synergies) for use in corners could work, but I can't do that either - the room is not a good shape for that either.
Another option might be a 15" co-axial, but there would still be problems to deal with (their depth is challenging, diffraction at internal discontinuities, ...).
My last arrays might be expected to be poor in terms of horizontal dispersion, but every time I compare with the earlier version which measures flatter horizontal angles >45 degrees, I prefer them on a wide range of music.
It's hard to make measurements to justify my observation, but my impression is that the defects in the room - which make the usual case of excess power response around 1-2 kHz sound particularly bad - complement the dip in power response. After all, the response is good out to >30 degrees with no eq on the 3FE25s.
What might I do different?
There could be a compromise with closer horizontal packing of the mids, between the 1:1 H:V ratio of the last design and the ~4:1 of the AMT. That would need a tweeter or tweeter+WG with the right properties, and - ruling out diffraction slots from past experience - I've not found a better solution.
Have fun with your synergies!
Ken
A bit of a long one if there is any interest. I have continued to spend (probably too much) time on arrays that meet the criteria that the midrange goes up to an XO where Ka = 1 (pure pistonic operation), and the spacing between the mid and tweeter is </= 1/4 Lambda at the crossover.
As we discussed before it is very difficult, even with the Bliesma T34B that can be used reliably down to 1.3kHz (it has an impressive Xmax of 1.3mm for a dome tweeter). It is also very wide, and unlike most dome tweeters the faceplate isn't removable, hence difficult to machine away (seems like it would be pretty risky to do it intact!).
The Accuton C51, crossing at 800Hz is interesting as a possibility in the tweeter position, the breakup almost usefully high, though there would be top octave rolloff and high directivity.
One of the problems with the approach is that the array in a large baffle would radiate at 180 degrees throughout. The directivity would be uniform until the tweeter begins to beam, but wide.
I recently found an article that shows the directivity of two drivers with spacing by lambda. This may provide a path to limited directivity in the vertical plane without lobes and relax the spacing criteria usefully.
Then again, maybe best to wait until @mabat develops the optimum waveguide for his Faital 1440f and cross it as low as possible.
Bill
As we discussed before it is very difficult, even with the Bliesma T34B that can be used reliably down to 1.3kHz (it has an impressive Xmax of 1.3mm for a dome tweeter). It is also very wide, and unlike most dome tweeters the faceplate isn't removable, hence difficult to machine away (seems like it would be pretty risky to do it intact!).
The Accuton C51, crossing at 800Hz is interesting as a possibility in the tweeter position, the breakup almost usefully high, though there would be top octave rolloff and high directivity.
One of the problems with the approach is that the array in a large baffle would radiate at 180 degrees throughout. The directivity would be uniform until the tweeter begins to beam, but wide.
I recently found an article that shows the directivity of two drivers with spacing by lambda. This may provide a path to limited directivity in the vertical plane without lobes and relax the spacing criteria usefully.
Then again, maybe best to wait until @mabat develops the optimum waveguide for his Faital 1440f and cross it as low as possible
.Bill
Attachments
Bill,
for the Horbach-Keele idea to work (links in post #1), the target regime is d/\lambda greater than or equal to ~0.6.
In the overlapping band lobing (big source) and non-lobing (small source) sum to give roughly constant directivity, at least as measured out to some relatively large angle. That needs an unconventional cross-over (H-K or BButterfield's or similar).
It's weird to measure the individual bands (outer ring, inner ring) and get "horrible" results, which then smooth out when the next higher band is added in.
I first found the compromise shallow WG while examining measurements that Fluid had also spotted - see post #87 for links. Better WG solutions appear to generally be too deep for my "against the wall" requirement.
Thanks to everyone who has posted, points and questions have helped clarify my thinking and understand what I hear.
Ken
for the Horbach-Keele idea to work (links in post #1), the target regime is d/\lambda greater than or equal to ~0.6.
In the overlapping band lobing (big source) and non-lobing (small source) sum to give roughly constant directivity, at least as measured out to some relatively large angle. That needs an unconventional cross-over (H-K or BButterfield's or similar).
It's weird to measure the individual bands (outer ring, inner ring) and get "horrible" results, which then smooth out when the next higher band is added in.
I first found the compromise shallow WG while examining measurements that Fluid had also spotted - see post #87 for links. Better WG solutions appear to generally be too deep for my "against the wall" requirement.
Thanks to everyone who has posted, points and questions have helped clarify my thinking and understand what I hear.
Ken
Yes,
Your approach as described is the interesting component of the thread (as well as what "Follgott" achieved).
My ideas are different for sure, trying to get a more predictable result with less experimentation, so not terribly relevant to this thread. I would absolutely love to ensconce myself in a lab-type experimental situation where I could investigate all the possibilities. Maybe some day when a less full work situation and an "empty nest".
Best wishes,
Bill
Your approach as described is the interesting component of the thread (as well as what "Follgott" achieved).
My ideas are different for sure, trying to get a more predictable result with less experimentation, so not terribly relevant to this thread. I would absolutely love to ensconce myself in a lab-type experimental situation where I could investigate all the possibilities. Maybe some day when a less full work situation and an "empty nest"
.Best wishes,
Bill