Nearing the end of this one's evolution
So a while ago, I figured out I generally like no baffle sound, but acknowledged it needs help in deep bass.
I'm happy with the blend these get with a previous Open top H Frame.(open top, but steeply tilted back, to limit midbass 1/2 space reinforcement from the floor).
So, after trying 98 db/ watt 15" trace Elliot (Celestion) woofers, I eventually got unhappy with the beaming.
This is my attempt to address beaming, and choose the boundary interactions that wouldn't be obstructive to a good listen.
From the top down, 16 ohm Peerless Alnico tweeters, firing in push pull, wired in parallel.
I'm hoping this provides vertical limiting on dispersion, limiting ceiling and floor reflections. I'm dubious this lowers distortion at frequencies this high(actively crossed at 2500 hz), but they do seem to sound sweeter this way.
The four midrange drivers have split the band in two-
Vertically stacked push pull, wired in series for an electrical resistance of 14 ohms, medium qts Peerless/M&K 5 1/4" , for vertical control of mid-high frequencies, passively crossed to the push pull 10", vertically stacked, series wired as well, for an 14.28 electrical resistance.
The 10" and 5 1/4" centre point of cross over is a little further apart than a first order calculator suggested, this dips the output for an acceptable response after dipole front wall reinforcement.
So this configuration gives controlled vertical dispersion similar to a line array, reduces side wall and ceiling reflections, provides push pull distortion cancelling benefits, passive eq'ing of the mid band, keeps the upper midrange away from floor and ceiling reflections, allows lower mid bass to couple with the floor for good 1/2 space reinforcement, and uses progressively larger drivers to move more air, while maintaining similar radiation patterns. As an added bonus, the Peerless 5 1/4" drivers have less Doppler distortion than my previous wide open tries.
These sound very even, top to bottom, and need no eq'ing from my now two way active crossover.
I've chosen to rigidly fix the tweeters to the speaker frame, as the decoupled woofers are effectively isolated from the frame.
Time, now, to make these look pretty, maybe maple and cherrywood, or possibly oak.
So a while ago, I figured out I generally like no baffle sound, but acknowledged it needs help in deep bass.
I'm happy with the blend these get with a previous Open top H Frame.(open top, but steeply tilted back, to limit midbass 1/2 space reinforcement from the floor).
So, after trying 98 db/ watt 15" trace Elliot (Celestion) woofers, I eventually got unhappy with the beaming.
This is my attempt to address beaming, and choose the boundary interactions that wouldn't be obstructive to a good listen.
From the top down, 16 ohm Peerless Alnico tweeters, firing in push pull, wired in parallel.
I'm hoping this provides vertical limiting on dispersion, limiting ceiling and floor reflections. I'm dubious this lowers distortion at frequencies this high(actively crossed at 2500 hz), but they do seem to sound sweeter this way.
The four midrange drivers have split the band in two-
Vertically stacked push pull, wired in series for an electrical resistance of 14 ohms, medium qts Peerless/M&K 5 1/4" , for vertical control of mid-high frequencies, passively crossed to the push pull 10", vertically stacked, series wired as well, for an 14.28 electrical resistance.
The 10" and 5 1/4" centre point of cross over is a little further apart than a first order calculator suggested, this dips the output for an acceptable response after dipole front wall reinforcement.
So this configuration gives controlled vertical dispersion similar to a line array, reduces side wall and ceiling reflections, provides push pull distortion cancelling benefits, passive eq'ing of the mid band, keeps the upper midrange away from floor and ceiling reflections, allows lower mid bass to couple with the floor for good 1/2 space reinforcement, and uses progressively larger drivers to move more air, while maintaining similar radiation patterns. As an added bonus, the Peerless 5 1/4" drivers have less Doppler distortion than my previous wide open tries.
These sound very even, top to bottom, and need no eq'ing from my now two way active crossover.
I've chosen to rigidly fix the tweeters to the speaker frame, as the decoupled woofers are effectively isolated from the frame.
Time, now, to make these look pretty, maybe maple and cherrywood, or possibly oak.
Attachments
Nearing the end of this one's evolution
So a while ago, I figured out I generally like no baffle sound, but acknowledged it needs help in deep bass.
I'm happy with the blend these get with a previous Open top H Frame.(open top, but steeply tilted back, to limit midbass 1/2 space reinforcement from the floor).
So, after trying 98 db/ watt 15" trace Elliot (Celestion) woofers, I eventually got unhappy with the beaming.
This is my attempt to address beaming, and choose the boundary interactions that wouldn't be obstructive to a good listen.
From the top down, 16 ohm Peerless Alnico tweeters, firing in push pull, wired in parallel.
I'm hoping this provides vertical limiting on dispersion, limiting ceiling and floor reflections. I'm dubious this lowers distortion at frequencies this high(actively crossed at 2500 hz), but they do seem to sound sweeter this way.
The four midrange drivers have split the band in two-
Vertically stacked push pull, wired in series for an electrical resistance of 14 ohms, medium qts Peerless/M&K 5 1/4" , for vertical control of mid-high frequencies, passively crossed to the push pull 10", vertically stacked, series wired as well, for an 14.28 electrical resistance.
The 10" and 5 1/4" centre point of cross over is a little further apart than a first order calculator suggested, this dips the output for an acceptable response after dipole front wall reinforcement.
So this configuration gives controlled vertical dispersion similar to a line array, reduces side wall and ceiling reflections, provides push pull distortion cancelling benefits, passive eq'ing of the mid band, keeps the upper midrange away from floor and ceiling reflections, allows lower mid bass to couple with the floor for good 1/2 space reinforcement, and uses progressively larger drivers to move more air, while maintaining similar radiation patterns. As an added bonus, the Peerless 5 1/4" drivers have less Doppler distortion than my previous wide open tries.
These sound very even, top to bottom, and need no eq'ing from my now two way active crossover.
I've chosen to rigidly fix the tweeters to the speaker frame, as the decoupled woofers are effectively isolated from the frame.
Time, now, to make these look pretty, maybe maple and cherrywood, or possibly oak.
So a while ago, I figured out I generally like no baffle sound, but acknowledged it needs help in deep bass.
I'm happy with the blend these get with a previous Open top H Frame.(open top, but steeply tilted back, to limit midbass 1/2 space reinforcement from the floor).
So, after trying 98 db/ watt 15" trace Elliot (Celestion) woofers, I eventually got unhappy with the beaming.
This is my attempt to address beaming, and choose the boundary interactions that wouldn't be obstructive to a good listen.
From the top down, 16 ohm Peerless Alnico tweeters, firing in push pull, wired in parallel.
I'm hoping this provides vertical limiting on dispersion, limiting ceiling and floor reflections. I'm dubious this lowers distortion at frequencies this high(actively crossed at 2500 hz), but they do seem to sound sweeter this way.
The four midrange drivers have split the band in two-
Vertically stacked push pull, wired in series for an electrical resistance of 14 ohms, medium qts Peerless/M&K 5 1/4" , for vertical control of mid-high frequencies, passively crossed to the push pull 10", vertically stacked, series wired as well, for an 14.28 electrical resistance.
The 10" and 5 1/4" centre point of cross over is a little further apart than a first order calculator suggested, this dips the output for an acceptable response after dipole front wall reinforcement.
So this configuration gives controlled vertical dispersion similar to a line array, reduces side wall and ceiling reflections, provides push pull distortion cancelling benefits, passive eq'ing of the mid band, keeps the upper midrange away from floor and ceiling reflections, allows lower mid bass to couple with the floor for good 1/2 space reinforcement, and uses progressively larger drivers to move more air, while maintaining similar radiation patterns. As an added bonus, the Peerless 5 1/4" drivers have less Doppler distortion than my previous wide open tries.
These sound very even, top to bottom, and need no eq'ing from my now two way active crossover.
I've chosen to rigidly fix the tweeters to the speaker frame, as the decoupled woofers are effectively isolated from the frame.
Time, now, to make these look pretty, maybe maple and cherrywood, or possibly oak.
With a lower Qts of .48 the 255-8 doesn't look as suitable for OB.
The Eminence Alpha 15a, OTOH has a Qts of 1.26.
Qts of 0.48 is quite high. Just look at Linkwitz LX521 bass and lower midrange drivers - their Qts params are really low. Low Qts drivers usually have better transient response.
I am also using low Qts drivers in my OB speakers - Deltalite II 2515 with Qts of 0.38 (<110hz) and 18sound 12NMB420 with Qts of 0.28 (110-500hz). Together with NEO10 and NEO3 PDR they sound very good.
I am also using low Qts drivers in my OB speakers - Deltalite II 2515 with Qts of 0.38 (<110hz) and 18sound 12NMB420 with Qts of 0.28 (110-500hz). Together with NEO10 and NEO3 PDR they sound very good.
That driver has been used successfully on a couple designs at Audio circle. Thats where the idea came from to try it. I had them already anyway. Very glad I did. I just run them as three ways at the moment and they already sound great. I may try 4 way at some point to see if I can squeak a bit more improvement out of them, but at this point any improvement i believe is going to be marginal.
Clearly, now
One thing I especially like about no baffle speakers, is one may prototype many efforts, simply by hanging new drivers in a frame.
These were a little thin, with no baffle. Eq'ing them actively robbed my meagre 7 watts worth of midrange dynamics.
Enter some presized 11x14 Lexan sheets, very affordable at my local Home Depot. They're a little hard to see, even with photoshopping, but it's here in the right channel.
I was also unable to get a really good blend with any sub, using active crossovering at 4th order.
Happened to have 4 47uF caps about, so added these to the Neo 10, no active highpass filtering.
Wanted the same electrical phase to the actively crossed tweeter, so added a 2uf cap to the speaker circuit, this corrects a little roll off on the Neo 3.
Bass was still a little lumpy, added a 12mH P-Core inductor, and this now features the improvements of tandem active and passive filtering.
So I get a little help from a baffle, and I don't need to see it.
One thing I especially like about no baffle speakers, is one may prototype many efforts, simply by hanging new drivers in a frame.
These were a little thin, with no baffle. Eq'ing them actively robbed my meagre 7 watts worth of midrange dynamics.
Enter some presized 11x14 Lexan sheets, very affordable at my local Home Depot. They're a little hard to see, even with photoshopping, but it's here in the right channel.
I was also unable to get a really good blend with any sub, using active crossovering at 4th order.
Happened to have 4 47uF caps about, so added these to the Neo 10, no active highpass filtering.
Wanted the same electrical phase to the actively crossed tweeter, so added a 2uf cap to the speaker circuit, this corrects a little roll off on the Neo 3.
Bass was still a little lumpy, added a 12mH P-Core inductor, and this now features the improvements of tandem active and passive filtering.
So I get a little help from a baffle, and I don't need to see it.
Attachments
Tried various OBs designs inspired by the looks and concept Jamo R909, and more loosely by various online published research and productions (esp.S.Linkwitz, John K. and Gainphile).
Observations collected from these OB prototypes as follows:
1st model: rectangle flat baffle 45x120cm, per side 2 x Beta 15A + Excel W18EX + SS HDS front firing per side and seas TDF rear firing.
2nd model: trapezoid shape flat baffle 20cm at top widening to 40 cm at base housing Seas H1209 + Excel W18 + SS HDS in 90 degree 3.5cm deep waveguide + separate Beta 15A in U-frame.
3rd model: same as above but tweeter section replaced with SS 10F on 12cm wide baffle extention + 19mm Vifa ring radiator in small waveguide.
3.5th model: pure experiment – combo of naked Beta 15A and SS10F
4th model: Excel W18 on rectangle baffle 28x35cm, SS10F naked (no baffle), Beta 15A in U-frame. Mundorf 17AMT planned to be added as tweeter but put on hold due to high cost of the latter. Additional SEAS H1209 in TL cabinet (from other project) was used to support low bass below 50Hz with obvious benefit.
5th ..?
Observations/conclusion are as follows:
1) impossible to get decent low bass on reasonably sizes baffle even with 4 of 15" woofers in total. BTW even speaker of 45 x 120cm frontal size is not reasonable for most people/rooms. Attempts to equalize down were resulting in easily exceeding Xmax of 4mm but still not enough low end. Coupling to the floor (taking off spikes and standing base on the mat) helped quite a bit, effect got stronger if panel put sideways, but U/H frame (no spikes) once tried proven to be notably better.
2) Two Beta 15A per side sounded significantly worse than one (apart from low end extension or headroom at loudness levels beyond sane). Tried parallel and series connection – without much difference. Going back to 1 woofer a side like removes a vail from midbass.
2) 45cm baffle is too wide for W18EX. Moving it to ~30cm showed immediate improvement in imaging and "naturalness" of midrange. Despite obvious higher dipole roll-off difference was not as pronounced as expected, namely there wasn’t much below 130-150Hz in either case.
3) Seas H1209 seems to be bad candidate for OB - on 45cm wide baffle it didn’t do much below 80Hz and apparently lacked dynamics. Although above 150Hz was "cleaner" than Beta 15A (meaning you hear more details as it sounds neat and accurate, but same thing reproduced with 15 inch whilst a bit rough its way more "live" and exciting).
4) SS HDS in waveguide had quite good directivity down to ~3kHz but annoying null @ 10kHz.
5) arrangement with monopole tweeter benefited from placing 4cm open-cell foam mat around W18 (guess making transition dipole to monopole smoother?))
6) SS10F taken as dipole fullrange on minimal baffle sounds gorgeous on voice recordings, properly detaching the sound from speaker creating illusion you listening to real singer live. Could not achieve same effect with rear firing tweeter of small waveguide. There was always a hint of the sound coming clearly out of the tweeter, can sound clearer and more focused but always somehow artificial.
7) Beta 15A in 40cm deep U-frame sounds better without or with very little stuffing despite cavity resonances showing up around 230Hz. Medium to heavy stuffing causes loss of that OB ease and clarity in the bass and starting to sound strained instead also exciting notably more room resonances, although measured frequency response looks better.
8) naked Beta 15A and SS10F crossed LR4 @ 700Hz sound surprisingly not bad, esp on certain accosting recordings like brass orchestra or large symphonic. Voices don’t sound right though. Bass below 80-100Hz is absent so another beta in H/U-frame or sealed sub to be added, although do not see a point in digging deeper into this experiment.
9) actually "3rd model described above was fully satisfactory sonically, however was not that great to look at and taking too much space. At the time was replaced by pyramid shape TL monopole speakers described here in other tread . . just to try smth different as well as aesthetical reasons
10) after living few months with monopoles decided to make A/B comparison and give dipoles another try. Wow-factor of "you-are-there" spaciousness and natural (albeit diffuse) sound is really something not forgotten too easy. As such decision take to reconstruct dipole system but tilting it a bit towards cardioid response:
combination of monopole (sealed/TL) + dipole bass, acoustic absorber (i.e. fiber+thick foam mat) around backside lower mid and thin absorber around upper mid with waveguided monopole upper treble should in theory bring together best of both concepts and allow placement not too far into the room
Observations collected from these OB prototypes as follows:
1st model: rectangle flat baffle 45x120cm, per side 2 x Beta 15A + Excel W18EX + SS HDS front firing per side and seas TDF rear firing.
2nd model: trapezoid shape flat baffle 20cm at top widening to 40 cm at base housing Seas H1209 + Excel W18 + SS HDS in 90 degree 3.5cm deep waveguide + separate Beta 15A in U-frame.
3rd model: same as above but tweeter section replaced with SS 10F on 12cm wide baffle extention + 19mm Vifa ring radiator in small waveguide.
3.5th model: pure experiment – combo of naked Beta 15A and SS10F
4th model: Excel W18 on rectangle baffle 28x35cm, SS10F naked (no baffle), Beta 15A in U-frame. Mundorf 17AMT planned to be added as tweeter but put on hold due to high cost of the latter. Additional SEAS H1209 in TL cabinet (from other project) was used to support low bass below 50Hz with obvious benefit.
5th ..?
Observations/conclusion are as follows:
1) impossible to get decent low bass on reasonably sizes baffle even with 4 of 15" woofers in total. BTW even speaker of 45 x 120cm frontal size is not reasonable for most people/rooms. Attempts to equalize down were resulting in easily exceeding Xmax of 4mm but still not enough low end. Coupling to the floor (taking off spikes and standing base on the mat) helped quite a bit, effect got stronger if panel put sideways, but U/H frame (no spikes) once tried proven to be notably better.
2) Two Beta 15A per side sounded significantly worse than one (apart from low end extension or headroom at loudness levels beyond sane). Tried parallel and series connection – without much difference. Going back to 1 woofer a side like removes a vail from midbass.
2) 45cm baffle is too wide for W18EX. Moving it to ~30cm showed immediate improvement in imaging and "naturalness" of midrange. Despite obvious higher dipole roll-off difference was not as pronounced as expected, namely there wasn’t much below 130-150Hz in either case.
3) Seas H1209 seems to be bad candidate for OB - on 45cm wide baffle it didn’t do much below 80Hz and apparently lacked dynamics. Although above 150Hz was "cleaner" than Beta 15A (meaning you hear more details as it sounds neat and accurate, but same thing reproduced with 15 inch whilst a bit rough its way more "live" and exciting).
4) SS HDS in waveguide had quite good directivity down to ~3kHz but annoying null @ 10kHz.
5) arrangement with monopole tweeter benefited from placing 4cm open-cell foam mat around W18 (guess making transition dipole to monopole smoother?))
6) SS10F taken as dipole fullrange on minimal baffle sounds gorgeous on voice recordings, properly detaching the sound from speaker creating illusion you listening to real singer live. Could not achieve same effect with rear firing tweeter of small waveguide. There was always a hint of the sound coming clearly out of the tweeter, can sound clearer and more focused but always somehow artificial.
7) Beta 15A in 40cm deep U-frame sounds better without or with very little stuffing despite cavity resonances showing up around 230Hz. Medium to heavy stuffing causes loss of that OB ease and clarity in the bass and starting to sound strained instead also exciting notably more room resonances, although measured frequency response looks better.
8) naked Beta 15A and SS10F crossed LR4 @ 700Hz sound surprisingly not bad, esp on certain accosting recordings like brass orchestra or large symphonic. Voices don’t sound right though. Bass below 80-100Hz is absent so another beta in H/U-frame or sealed sub to be added, although do not see a point in digging deeper into this experiment.
9) actually "3rd model described above was fully satisfactory sonically, however was not that great to look at and taking too much space. At the time was replaced by pyramid shape TL monopole speakers described here in other tread . . just to try smth different as well as aesthetical reasons
10) after living few months with monopoles decided to make A/B comparison and give dipoles another try. Wow-factor of "you-are-there" spaciousness and natural (albeit diffuse) sound is really something not forgotten too easy. As such decision take to reconstruct dipole system but tilting it a bit towards cardioid response:
combination of monopole (sealed/TL) + dipole bass, acoustic absorber (i.e. fiber+thick foam mat) around backside lower mid and thin absorber around upper mid with waveguided monopole upper treble should in theory bring together best of both concepts and allow placement not too far into the room
photos of the prototypes..
1st one (JAMO R909 -alike)
1st one (JAMO R909 -alike)
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Thanks Stoom!
>>> 2) Two Beta 15A per side sounded significantly worse than one (apart from low end extension or headroom at loudness levels beyond sane). Tried parallel and series connection – without much difference. Going back to 1 woofer a side like removes a vail from mid bass.
I always wondered about this. Thanks for your comment. I also wonder how much better a U or H frame would sound compared to straight up on a baffle. Others may have different opinions but I appreciate yours.
The speakers look great! Have you tried pulling them a bit more out from the rear wall?
Godzilla
>>> 2) Two Beta 15A per side sounded significantly worse than one (apart from low end extension or headroom at loudness levels beyond sane). Tried parallel and series connection – without much difference. Going back to 1 woofer a side like removes a vail from mid bass.
I always wondered about this. Thanks for your comment. I also wonder how much better a U or H frame would sound compared to straight up on a baffle. Others may have different opinions but I appreciate yours.
The speakers look great! Have you tried pulling them a bit more out from the rear wall?
Godzilla
rear wall was some 3meters behind, theses are just open ikea bookshelves practically transparent to sound
question abt U/H/N vs flat baffle - generally low end sounds better upper midbass/low midrange worse. but considering Betas should be ideally crossed @ around 150-180 Hz think ideal solution would be 35cm H frame placing acoustic center in the middle which allows part cancellation of cavity resonances, which will also be sufficiently above crossover point
question abt U/H/N vs flat baffle - generally low end sounds better upper midbass/low midrange worse. but considering Betas should be ideally crossed @ around 150-180 Hz think ideal solution would be 35cm H frame placing acoustic center in the middle which allows part cancellation of cavity resonances, which will also be sufficiently above crossover point
Hi,
With two drivers, have you try this design :
H frame, Alpha15A mounting back to back Xo 175 hz 12db, in this case front wave and rear wave are the same. Distance to rear wall 50 cm..
Mesurement :
Red is predict, black is mesurement at 3,5 m (listening point).
I build several OB, U and H with one or two drivers, and I agree one driver works better than two in "classical" design.
To understand this, you must mesure front and rear wave, and compare.
Generally rear wave show a very big hump at lowmid.
Phil.
With two drivers, have you try this design :
An externally hosted image should be here but it was not working when we last tested it.
H frame, Alpha15A mounting back to back Xo 175 hz 12db, in this case front wave and rear wave are the same. Distance to rear wall 50 cm..
Mesurement :
An externally hosted image should be here but it was not working when we last tested it.
Red is predict, black is mesurement at 3,5 m (listening point).
I build several OB, U and H with one or two drivers, and I agree one driver works better than two in "classical" design.
To understand this, you must mesure front and rear wave, and compare.
Generally rear wave show a very big hump at lowmid.
Phil.
With OB's it's always about pulling them out into the room.
My single H-frames are almost too heavy for that.
That's why I asked about casters.
the stand :
Phil.
An externally hosted image should be here but it was not working when we last tested it.
Phil.
by "leakage" meant the dipole from to back cancellation, an inch is enough for sound to slip thru regardless of the wavelength. end there is difference between having say 50cm cancellation path on 4 sides of driver OR 50cm on 3 sides and practically infinite on one.
Philfr - maybe your H-frame is sufficiently deep to support low bass as it is but in my experiments I found than box/frame bottom placed directly on the floor adds some weight to low end. Rubber or smth needed to avoid rattling noise as neither the floor or box bottom surfaces perfectly flat. Understand your use of rubber mat was different, i.e. to decouple box from the base of the stand firmly placed on the floor with spikes..
Philfr - maybe your H-frame is sufficiently deep to support low bass as it is but in my experiments I found than box/frame bottom placed directly on the floor adds some weight to low end. Rubber or smth needed to avoid rattling noise as neither the floor or box bottom surfaces perfectly flat. Understand your use of rubber mat was different, i.e. to decouple box from the base of the stand firmly placed on the floor with spikes..