Thoughts along the way.
I decided to aim for ~100 degree coverage, and not worry about floor and ceiling reflections for now.
Big round waveguides - need to be large, and while rear radiation can be suppressed, it's hard to find a way to reach down to the Schroeder frequency range (or even within an octave of it). From old experience MEH is beyond my ability/skill/patience. I need something simple to make. Coverage of ~100 degrees is not common in any available waveguide, and making one is not an option at the moment.
Medium/large co-ax unit - too deep, also too small to provide controlled radiation down low enough: e.g. 12" unit lacks directionality up to ~1kHz, even though it could play down to 100 Hz. I'd use a medium size unit with an added ring of smaller drivers to increase the effective radiating area, but that's not an attractive solution.
I therefore concluded that another fractal-array was the best option, in particular because they are easy to make with minimal woodworking tools or ability.
The next step was to decide on a tweeter - needs a waveguide to obtain any semblance of controlled dispersion, how many bands to include in the array and how many drivers to include in each band.
As before, these speakers will stand on a 12" base unit sitting on the floor (no floor bounce). I don't need much bass extension from the new speakers and as I had a bunch of SDS-135 units left over, a ring of those would form the lower/lowest band of the array.
Choice of the array shape and other drivers to follow.
I decided to aim for ~100 degree coverage, and not worry about floor and ceiling reflections for now.
Big round waveguides - need to be large, and while rear radiation can be suppressed, it's hard to find a way to reach down to the Schroeder frequency range (or even within an octave of it). From old experience MEH is beyond my ability/skill/patience. I need something simple to make. Coverage of ~100 degrees is not common in any available waveguide, and making one is not an option at the moment.
Medium/large co-ax unit - too deep, also too small to provide controlled radiation down low enough: e.g. 12" unit lacks directionality up to ~1kHz, even though it could play down to 100 Hz. I'd use a medium size unit with an added ring of smaller drivers to increase the effective radiating area, but that's not an attractive solution.
I therefore concluded that another fractal-array was the best option, in particular because they are easy to make with minimal woodworking tools or ability.
The next step was to decide on a tweeter - needs a waveguide to obtain any semblance of controlled dispersion, how many bands to include in the array and how many drivers to include in each band.
As before, these speakers will stand on a 12" base unit sitting on the floor (no floor bounce). I don't need much bass extension from the new speakers and as I had a bunch of SDS-135 units left over, a ring of those would form the lower/lowest band of the array.
Choice of the array shape and other drivers to follow.
With BButterfield's general rules for making fractal arrays in mind, and following some measurements on test baffles, I chose the following solution.
The choice was biased to use parts I had. The hardest choice was the tweeter and its waveguide. Somethin I've played with on and off for well over a decade. I decided to use SEAS 27TFFC in Visaton WG-148. I am aware this is not a universally-liked choice but (perhaps due to limits of my hearing being around 14kHz), I like them.
To reduce the extreme sensitivity in the 2-3 kHz range, the tweeters have 4.4 uF in series, which gives a relatively flat response. A little digital correction is needed around and about 10 kHz. These capacitors are the only passive cross-over components and the only driver correction used in the speakers.
I decided to have two mid-range bands, and that I would risk using only 4 (rather than 6) units per band (series, parallel wired). The crossover frequencies are - broadly speaking - about 1 kHz, and a little over 2kHz.
Given the frequency range I decided to use Faital Pro 3FE25-4 for the upper-mids. The square is angled somewhat in the style of the DNA speakers (not exactly the same angles). Tests showed that these do not require any equalisation beyond the crossover, with relatively smooth responses extending far beyond the 1-2kHz band. These share a 2l sub-enclosure that also acts as a brace from baffle to the back of the enclosure.
During setup measurements, I became extremely impressed by these little units due to innocuous frequency response and wide bandwidth. I should have used these from the start, but chose slightly smaller units in the earlier arrays.
The range from somewhat under 200Hz to 1 kHz is covered by Peerless by Tymphany SDS-135. These share a ~15l sealed volume (0.5x0.8x0.05m minus bracing etc.) . They hand over to 12" woofers on the floor.
These were spaced to provide a wide effective radiating area.
Image(s) to follow.
The choice was biased to use parts I had. The hardest choice was the tweeter and its waveguide. Somethin I've played with on and off for well over a decade. I decided to use SEAS 27TFFC in Visaton WG-148. I am aware this is not a universally-liked choice but (perhaps due to limits of my hearing being around 14kHz), I like them.
To reduce the extreme sensitivity in the 2-3 kHz range, the tweeters have 4.4 uF in series, which gives a relatively flat response. A little digital correction is needed around and about 10 kHz. These capacitors are the only passive cross-over components and the only driver correction used in the speakers.
I decided to have two mid-range bands, and that I would risk using only 4 (rather than 6) units per band (series, parallel wired). The crossover frequencies are - broadly speaking - about 1 kHz, and a little over 2kHz.
Given the frequency range I decided to use Faital Pro 3FE25-4 for the upper-mids. The square is angled somewhat in the style of the DNA speakers (not exactly the same angles). Tests showed that these do not require any equalisation beyond the crossover, with relatively smooth responses extending far beyond the 1-2kHz band. These share a 2l sub-enclosure that also acts as a brace from baffle to the back of the enclosure.
During setup measurements, I became extremely impressed by these little units due to innocuous frequency response and wide bandwidth. I should have used these from the start, but chose slightly smaller units in the earlier arrays.
The range from somewhat under 200Hz to 1 kHz is covered by Peerless by Tymphany SDS-135. These share a ~15l sealed volume (0.5x0.8x0.05m minus bracing etc.) . They hand over to 12" woofers on the floor.
These were spaced to provide a wide effective radiating area.
Image(s) to follow.
Image of new toy design
I built the above described speakers in only a few hours each. They ended up 608mm x 900mm (sitting ~400mm above floor) x 90mm deep.
Quick measurements of the first suggested it was worth completing both. Thus far I've used crude crossover filters and plan to improve these at some point (though that's more as a matter of principle than due to dissatisfaction with the sound, which I find surprisingly pleasant in the first month of listening).
Image of (nearly) the full test baffle for initial tests.
https://www.diyaudio.com/forums/attachment.php?attachmentid=869929&d=1598181177
Here is a sample, early Holm Impulse measurement: in room, ~2m distance, ~30 degrees (that's more or less in front of one normal listening position). The default "complex average" in Holm is a crude, heavily-averaged response measurement I used for initial checks of the balance after adjusting the crossovers with higher resolution - there's no equalisation between <200 Hz and ~10 kHz. The measurement is misleading below ~200 Hz, and not the final crossover between the new speakers (red) and a couple of different states of the woofers & sub-woofer array (blue & green) which had not yet been matched to the new speakers.
https://www.diyaudio.com/forums/attachment.php?attachmentid=869930&d=1598181177
As explained earlier, I'm not going to put in typically 1/12th octave averaged measurments used to set the crossovers. It requires a whole ensemble of plots and too much explanation to separate out interference effects from the speaker response. I don't plan to descibe the crossovers, certainly not in their current state.
This is a fun approach for someone who wants to play with fractal arrays & to develop the assocated filters.
Ken
I built the above described speakers in only a few hours each. They ended up 608mm x 900mm (sitting ~400mm above floor) x 90mm deep.
Quick measurements of the first suggested it was worth completing both. Thus far I've used crude crossover filters and plan to improve these at some point (though that's more as a matter of principle than due to dissatisfaction with the sound, which I find surprisingly pleasant in the first month of listening).
Image of (nearly) the full test baffle for initial tests.
https://www.diyaudio.com/forums/attachment.php?attachmentid=869929&d=1598181177
Here is a sample, early Holm Impulse measurement: in room, ~2m distance, ~30 degrees (that's more or less in front of one normal listening position). The default "complex average" in Holm is a crude, heavily-averaged response measurement I used for initial checks of the balance after adjusting the crossovers with higher resolution - there's no equalisation between <200 Hz and ~10 kHz. The measurement is misleading below ~200 Hz, and not the final crossover between the new speakers (red) and a couple of different states of the woofers & sub-woofer array (blue & green) which had not yet been matched to the new speakers.
https://www.diyaudio.com/forums/attachment.php?attachmentid=869930&d=1598181177
As explained earlier, I'm not going to put in typically 1/12th octave averaged measurments used to set the crossovers. It requires a whole ensemble of plots and too much explanation to separate out interference effects from the speaker response. I don't plan to descibe the crossovers, certainly not in their current state.
This is a fun approach for someone who wants to play with fractal arrays & to develop the assocated filters.
Ken
Attachments
Thanks. When I measured the 3FE25s on a test baffle, I was amazed at just how well-behaved they are. I'm not surprised that the '22s are just as good. One of the reasons I updated the thread was to provide another hit if someone searches for those little Faital Pro drives. They are a different stratum of quality from the wee Peerless mids up-thread.
Amusing: I'd put the 3FE25s and tweeter in the baffle, connected to a 4 pole speakon tweeter wired to one channel, one mid to the other, and made a measurement at 60cm on the tweeter axis. The response was so extended (IIRC ~14kHz) and so flat - because it was slightly off axis from the 3FE25, that I thought I'd messed up and was measuring the tweeter with its waveguide and capactor. Had me puzzled for a couple of moments.
These were by far the highest fun-to-effort ratio speakers ever, and the most immediately satisfying. The crossovers for the above measurement took about 10 minutes to set up. A huge contrast to failed attempts at an MEH a decade ago.
I'm now having the pleasure of listening to as much music as time allows. Patricia Rozario is singing Eternity's Sunrise at the moment.
Ken
Amusing: I'd put the 3FE25s and tweeter in the baffle, connected to a 4 pole speakon tweeter wired to one channel, one mid to the other, and made a measurement at 60cm on the tweeter axis. The response was so extended (IIRC ~14kHz) and so flat - because it was slightly off axis from the 3FE25, that I thought I'd messed up and was measuring the tweeter with its waveguide and capactor. Had me puzzled for a couple of moments.
These were by far the highest fun-to-effort ratio speakers ever, and the most immediately satisfying. The crossovers for the above measurement took about 10 minutes to set up. A huge contrast to failed attempts at an MEH a decade ago.
I'm now having the pleasure of listening to as much music as time allows. Patricia Rozario is singing Eternity's Sunrise at the moment.
Ken
Ken, thanks for your posts,
and links to other's work in this area, like BButterfield and FoLLgoTT.
Please continue to give updates.
You've inspired me to start another project, yay!
(Have been in a bit of a slump, having successfully concluded a pair of fairly large, but lightweight, synergies.)
So your type project comes at a great time to try something new
What do you call these things, btw? Psuedo-coaxials, fractal-arrays ???
Gonna use a Wavecore TW030WA14 for tweeter.
Surrounded by five 3FE25's tightpacked around the tweeter. (Per your enthusiasm over them.
Surrounded by five Faital 10FE200's tightpacked around the 3FE25's.
Hopefully, with plenty of latitude to play with xover frequencies for best patterning.
I really want to try minimize all center-to-center distances, as I'm convinced that's probably the biggest factor in the super sound of synergies.
But I don't know if that's advisable, as it appears you guys are widening spacing some for better pattern control ??
Should give about a 30" 'circular' diameter, which I'm thinking may give around 100 deg H&V pattern control down to around 250Hz.
Which is the second part of my largish synergies i'd like to try to duplicate, extended low freq control.
At any rate, I figure it's easy enough to try another baffle, less tightpacked...
I've no idea what to do for overall baffle size...any recommendations appreciated. And pls do point out any train wrecks I'm not seeing....
I already had some 10FE200's on hand, so total driver cost for one trial speaker is under $130 for a project that sounds exciting, double yay!!
You sold me
and links to other's work in this area, like BButterfield and FoLLgoTT.
Please continue to give updates.
You've inspired me to start another project, yay!
(Have been in a bit of a slump, having successfully concluded a pair of fairly large, but lightweight, synergies.)
So your type project comes at a great time to try something new
What do you call these things, btw? Psuedo-coaxials, fractal-arrays ???
Gonna use a Wavecore TW030WA14 for tweeter.
Surrounded by five 3FE25's tightpacked around the tweeter. (Per your enthusiasm over them.
Surrounded by five Faital 10FE200's tightpacked around the 3FE25's.
Hopefully, with plenty of latitude to play with xover frequencies for best patterning.
I really want to try minimize all center-to-center distances, as I'm convinced that's probably the biggest factor in the super sound of synergies.
But I don't know if that's advisable, as it appears you guys are widening spacing some for better pattern control ??
Should give about a 30" 'circular' diameter, which I'm thinking may give around 100 deg H&V pattern control down to around 250Hz.
Which is the second part of my largish synergies i'd like to try to duplicate, extended low freq control.
At any rate, I figure it's easy enough to try another baffle, less tightpacked...
I've no idea what to do for overall baffle size...any recommendations appreciated. And pls do point out any train wrecks I'm not seeing....
I already had some 10FE200's on hand, so total driver cost for one trial speaker is under $130 for a project that sounds exciting, double yay!!
You sold me
Mark, wow!
knowing that you've tackled much harder projects, I would suggest the way to proceed is look at BButterfield's rules of thumb (or the Horbach-Keele idea), and decide on your other requirements, in particular the depth of the speakers which essentially fixes the volume for each 10FE200, see below.
The 30"wide panel is going to be well-filled with radiating surface.
I'd bet that's going to be great for percussion, in room.
I don't know anything about the tweeter, so won't comment.
The gaps I left would have been smaller with five or six drives per ring, filling in gaps can't do harm. If I'd used bigger drives on the same diameters of ring, the fill-factor would have been better, it's not really about spreading out the ring: I wanted a factor of ~2 in the ratio of the two diameters, which is not a large ratio for fractal arrays.
Eliminating the floor-bounce notch is important to me, so I'm happy that the four SDS-135s roughly match the sensitivity of my woofers - they are driven from the same type of amplifier.
Depending how shallow you want the speakers to be, finding volume for the 10FE200s might be challenging. As you've already got the units, perhaps this does not matter much, though you're going to have to deal with quite a high Q peak. Hmm. That's going to make balancing out the low-end response a big challenge (not in principle as the right inverse filter solves the problem, but because separating out room modes and high-Q speaker response to deduce the required inverse filter is not at all easy).
On the other hand, if the speakers get deeper, a rear-wall reflection starts to appear. Needs a careful compromise, I think. I would not know how to make this work.
The larger panel is also going to need more attention to bracing. I used a timber frame to create the thickness of the box, but that might not work well for the volime needed for 10FE200s, so it's going to need to be a heavy construction.
I'm intrigued by your choice of five-drive rings. I'd stuck with 4 or 6 to give the array, wired series-parallel, either 1, or 2/3 or 3/2 of the drive impedance. With five I'd put 16 ohm drives in parallel, but I don't believe 10FE200 is available in 16 ohms.
For a 30" baffle, I'd probably go for a ring of six 6FE200 (Vas 6l, 10x less than the 10FE200, these are more mid-biased and you'd get the same sensitivity with Q~1). That's as big as I'd be happy to go in the low cost Faital Pro line (I've just looked at the selection chart, and the 8FE200 is a step too far in my view, you can see that the 200s share the same motor).
If you are aiming for something really loud, it won't be possible to keep it shallow (depending on the LF cutoff, of course, but I'm assuming 100-200 Hz).
As you note, it's not too hard to make a test baffle.
Have a lot of fun!
Ken
knowing that you've tackled much harder projects, I would suggest the way to proceed is look at BButterfield's rules of thumb (or the Horbach-Keele idea), and decide on your other requirements, in particular the depth of the speakers which essentially fixes the volume for each 10FE200, see below.
The 30"wide panel is going to be well-filled with radiating surface.
I'd bet that's going to be great for percussion, in room.
I don't know anything about the tweeter, so won't comment.
The gaps I left would have been smaller with five or six drives per ring, filling in gaps can't do harm. If I'd used bigger drives on the same diameters of ring, the fill-factor would have been better, it's not really about spreading out the ring: I wanted a factor of ~2 in the ratio of the two diameters, which is not a large ratio for fractal arrays.
Eliminating the floor-bounce notch is important to me, so I'm happy that the four SDS-135s roughly match the sensitivity of my woofers - they are driven from the same type of amplifier.
Depending how shallow you want the speakers to be, finding volume for the 10FE200s might be challenging. As you've already got the units, perhaps this does not matter much, though you're going to have to deal with quite a high Q peak. Hmm. That's going to make balancing out the low-end response a big challenge (not in principle as the right inverse filter solves the problem, but because separating out room modes and high-Q speaker response to deduce the required inverse filter is not at all easy).
On the other hand, if the speakers get deeper, a rear-wall reflection starts to appear. Needs a careful compromise, I think. I would not know how to make this work.
The larger panel is also going to need more attention to bracing. I used a timber frame to create the thickness of the box, but that might not work well for the volime needed for 10FE200s, so it's going to need to be a heavy construction.
I'm intrigued by your choice of five-drive rings. I'd stuck with 4 or 6 to give the array, wired series-parallel, either 1, or 2/3 or 3/2 of the drive impedance. With five I'd put 16 ohm drives in parallel, but I don't believe 10FE200 is available in 16 ohms.
For a 30" baffle, I'd probably go for a ring of six 6FE200 (Vas 6l, 10x less than the 10FE200, these are more mid-biased and you'd get the same sensitivity with Q~1). That's as big as I'd be happy to go in the low cost Faital Pro line (I've just looked at the selection chart, and the 8FE200 is a step too far in my view, you can see that the 200s share the same motor).
If you are aiming for something really loud, it won't be possible to keep it shallow (depending on the LF cutoff, of course, but I'm assuming 100-200 Hz).
As you note, it's not too hard to make a test baffle.
Have a lot of fun!
Ken
Last edited:
Thank you very much for all that info Ken !
The reason for 5 drivers is because 5 of the 3fe25s fit tightly around the Wavecore tweeter, with is 76mm dia.
Not much wasted space, and on paper, a piece of capped 12" sonotube fits around their 11" diameter profile to seal them all off.
My obsession with tight center-to-center spacing drove the Wavecore decision. Everything i really want to try, like maybe the Ciare CT-440 is more like 4+ inches, and more $$. Got the TW030WA14 for $30.
I've realized that the sonotube pushes the 10fe200s out about 1" from the array's center point, so whole array ends up being 32" in dia.
But again, all on paper...best laid plans of mice and men, eh?
If i didn't already have the 10fe200's, i would be trying something smaller than them. .... like any of your choices....
My 10"s are 4 ohms, so it works ok for 5 drivers i think, after i buy an 8 ohm to go with them.
On both the 3" and 10", i'll parallel [two 4 ohms in series with two 4 ohms in series with one 8 ohm] or the equivalent of three 8 ohm loads in parallel for a 2.7 ohm load. Using proaudio amps, so should be no problem.
If I take the internal cabinet up to around a 37-38" square equivalent, it should yield about 20L per inch of depth, net of the sonotube volume takeup.
I'm thinking the 10fe200's will be happy with 40L each, and i can keep cabinet depth to 10".
If so, I hope to cross to the 3" below the rear wall resonance, maybe 600Hz or so. Planning on a 100Hz low xover to sub (same as on every build)
Will be using FIR dsp tuning, so should be easy enough if the acoustic design is close to reasonable.
The wavecore is clearly the SPL limiter...it's only good to about 105 dB.
It will be welcome indeed if this project sounds so good to warrant a better tweeter
The reason for 5 drivers is because 5 of the 3fe25s fit tightly around the Wavecore tweeter, with is 76mm dia.
Not much wasted space, and on paper, a piece of capped 12" sonotube fits around their 11" diameter profile to seal them all off.
My obsession with tight center-to-center spacing drove the Wavecore decision. Everything i really want to try, like maybe the Ciare CT-440 is more like 4+ inches, and more $$. Got the TW030WA14 for $30.
I've realized that the sonotube pushes the 10fe200s out about 1" from the array's center point, so whole array ends up being 32" in dia.
But again, all on paper...best laid plans of mice and men, eh?
If i didn't already have the 10fe200's, i would be trying something smaller than them. .... like any of your choices....
My 10"s are 4 ohms, so it works ok for 5 drivers i think, after i buy an 8 ohm to go with them.
On both the 3" and 10", i'll parallel [two 4 ohms in series with two 4 ohms in series with one 8 ohm] or the equivalent of three 8 ohm loads in parallel for a 2.7 ohm load. Using proaudio amps, so should be no problem.
If I take the internal cabinet up to around a 37-38" square equivalent, it should yield about 20L per inch of depth, net of the sonotube volume takeup.
I'm thinking the 10fe200's will be happy with 40L each, and i can keep cabinet depth to 10".
If so, I hope to cross to the 3" below the rear wall resonance, maybe 600Hz or so. Planning on a 100Hz low xover to sub (same as on every build)
Will be using FIR dsp tuning, so should be easy enough if the acoustic design is close to reasonable.
The wavecore is clearly the SPL limiter...it's only good to about 105 dB.
It will be welcome indeed if this project sounds so good to warrant a better tweeter
Mark,
bigger and bigger (and possibly better), 38" square will be something to behold.
I see the sense of making a tight ring - saves worrying about the spacing, let the design take care of itself.
~10" deep is probably OK given how little of the sound will be going round the back (mainly below ~120 Hz), resulting Q~1.1 is not bad for the 10FE200s, and they will work down to 70Hz if you choose (I'd also cross higher though, like the 100 Hz you suggest).
In this case 10FE200s should indeed work quite well.
For me, the biggest challenge would be to make a stiff construction that's not too heavy to move. I look forward to seeing your results.
Ken
bigger and bigger (and possibly better), 38" square will be something to behold.
I see the sense of making a tight ring - saves worrying about the spacing, let the design take care of itself.
~10" deep is probably OK given how little of the sound will be going round the back (mainly below ~120 Hz), resulting Q~1.1 is not bad for the 10FE200s, and they will work down to 70Hz if you choose (I'd also cross higher though, like the 100 Hz you suggest).
In this case 10FE200s should indeed work quite well.
For me, the biggest challenge would be to make a stiff construction that's not too heavy to move. I look forward to seeing your results.
Ken
Last edited:
Ken, I agree with your biggest challenge....it's mine too.
38" square breaks my personal 75lb weight limit for a main speaker.
Right now, it looks like an internal 30"x30"x12" is as big as I can go and stay under weight.
That will give 30L net for each 10fe200, less than i wanted.
But the 10" section has far more SPL than the 3" or tweeter, so I can EQ the lower end response hump of the 10's down and be OK i think.
Speaking of SPL, I've switched the tweeter to a SG Acoutics unit SB Acoustics SB26STWGC-4 Fabric Dome Tweeter with Wave Guide
Three reasons really; more SPL, I figured out how to fit it and all the 3fe-25's into a 12" sonotube, and it has some directivity control beginning at 2kHz that appears will be needed to blend in the overall pattern control.
But just guessing about the pattern control, cause i'm not much on sims.
i gotta picture stuff in my mellon, then measure and find out why my pictures don't work Lol
Oh, I really don't mean to be pirating your thread with my project...not fair/cool.
I'll start my own once it gets going, but would like to come here for technical advice and experience.
That OK?
38" square breaks my personal 75lb weight limit for a main speaker.
Right now, it looks like an internal 30"x30"x12" is as big as I can go and stay under weight.
That will give 30L net for each 10fe200, less than i wanted.
But the 10" section has far more SPL than the 3" or tweeter, so I can EQ the lower end response hump of the 10's down and be OK i think.
Speaking of SPL, I've switched the tweeter to a SG Acoutics unit SB Acoustics SB26STWGC-4 Fabric Dome Tweeter with Wave Guide
Three reasons really; more SPL, I figured out how to fit it and all the 3fe-25's into a 12" sonotube, and it has some directivity control beginning at 2kHz that appears will be needed to blend in the overall pattern control.
But just guessing about the pattern control, cause i'm not much on sims.
i gotta picture stuff in my mellon, then measure and find out why my pictures don't work Lol
Oh, I really don't mean to be pirating your thread with my project...not fair/cool.
I'll start my own once it gets going, but would like to come here for technical advice and experience.
That OK?
Hmm ~1/2" 4 x 8 ft BB ply is ~48#, making a pretty big, OK braced speaker. Finish stiffening/mass loading it by placing enough removable [decorative] weight
FWIW, this is still overkill for all but high power/Xmax 'subs' as my norm for 'customer' builds were all 1/4" except for the baffle since many had to be moved by children, shorter and/or older adults: How Much Does Plywood Weigh? - Inch Calculator
GM
Good stuff GM,
I'm a fan of lightweight too....
I like to use thinnest panels possible for the displacement being handled, brace well while avoiding overdone, then find & fix resonances......and say enough.
And grin when I can move the box around
Here's the spec sheet i use for BB weight...it matches what I get. https://www.wolstenholme.com/pdf/Baltic Birch Plywood Specifications.pdf
With regard to my box under contemplation...
drivers weigh 33#, box without bracing 36#.
Bottom end displacement is about equal to a 15" with 10mm xmax.
Gonna be a real squeeze to keep it under 75#, i think....
I'm a fan of lightweight too....
I like to use thinnest panels possible for the displacement being handled, brace well while avoiding overdone, then find & fix resonances......and say enough.
And grin when I can move the box around
Here's the spec sheet i use for BB weight...it matches what I get. https://www.wolstenholme.com/pdf/Baltic Birch Plywood Specifications.pdf
With regard to my box under contemplation...
drivers weigh 33#, box without bracing 36#.
Bottom end displacement is about equal to a 15" with 10mm xmax.
Gonna be a real squeeze to keep it under 75#, i think....
following up your points and your question
Mark,
I missed your post earlier in the week when GM's reply fell on the next page. I'll look out for your design & build thread, meantime feel free to continue thoughts and questions here. It helps me to clarify my thoughts and explanations, as below.
Of course, you are right that in principle that the high-Q of the 10s can be perfectly compensated by an inverse filter (which you are capable of implementing, I've seen). I would still look for units with lower Vas. Minimizing the rear-wall reflection means a lot to me, hence I'd use the lower Vas to make thinner speakers. It took me a long time to settle on the final shape balancing several of these points.
By far the trickiest aspect of making a fractal array is handling the crossover to the tweeter. Indeed, perhaps that's the only substantial challenge, for people who can make suitable filters. Rings of mid units can be made to have similar directivity over a 2-3 octaves, using suitable filters, but it's hard to make the tweeter match, as that must be done acoustically. With a small-WG tweeter, I suspect you are heading for trouble integrating the tweeter and getting voice to sound natural.
In my AMT version, with the focus on floor & ceiling reflections, I matched the vertical directivity, approximately, and let horizontal go wide. In the later version, with the dome tweeter, I considered the tweeter in its WG first.
In-between times, I looked at dozens of tweeter+WG combinations to fine one that might be more sensitive than a standard dome, including CDs on various horns/WGs. 3D printing a shallow WG with recesses for the inner ring of midranges is worth exploring.
It would be great if we could figure out how to a smaller ring, and push problems up higher. The SB tweeter looks to be a good match to an array from ~5 kHz, but unless the faceplate can be machined away, it's hard to get mids close enough to reach as high as I believe necessary. The WG148 provides closer to constant directivity to lower frequency (~2kHz) which makes the crossover easier. Perhaps a starting point is to choose the effective diameter of the WG (i.e. ignoring the roll-over surround) ~1/2 the inner ring diameter. The SB tweeter WG looks like ~60mm diameter, or smaller and a 120 or even 150mm ring is hard to achieve.
I found that only a little simulation is really needed for the array itself. The mids act like point sources, so each ring of array behaves much like an emitting surface matching the diameter of the ring (centers). I often play around with "Edge" to aid what's going on in my "mellon", but the usual rules of thumb for directivity of (say) a 15" drive can be used, or the equations for a rigid piston of the chosen diameter. In Edge, the baffle can be set to either its true size or the size of the rear wall, as required.
Ken
Mark,
I missed your post earlier in the week when GM's reply fell on the next page. I'll look out for your design & build thread, meantime feel free to continue thoughts and questions here. It helps me to clarify my thoughts and explanations, as below.
Of course, you are right that in principle that the high-Q of the 10s can be perfectly compensated by an inverse filter (which you are capable of implementing, I've seen). I would still look for units with lower Vas. Minimizing the rear-wall reflection means a lot to me, hence I'd use the lower Vas to make thinner speakers. It took me a long time to settle on the final shape balancing several of these points.
By far the trickiest aspect of making a fractal array is handling the crossover to the tweeter. Indeed, perhaps that's the only substantial challenge, for people who can make suitable filters. Rings of mid units can be made to have similar directivity over a 2-3 octaves, using suitable filters, but it's hard to make the tweeter match, as that must be done acoustically. With a small-WG tweeter, I suspect you are heading for trouble integrating the tweeter and getting voice to sound natural.
In my AMT version, with the focus on floor & ceiling reflections, I matched the vertical directivity, approximately, and let horizontal go wide. In the later version, with the dome tweeter, I considered the tweeter in its WG first.
In-between times, I looked at dozens of tweeter+WG combinations to fine one that might be more sensitive than a standard dome, including CDs on various horns/WGs. 3D printing a shallow WG with recesses for the inner ring of midranges is worth exploring.
It would be great if we could figure out how to a smaller ring, and push problems up higher. The SB tweeter looks to be a good match to an array from ~5 kHz, but unless the faceplate can be machined away, it's hard to get mids close enough to reach as high as I believe necessary. The WG148 provides closer to constant directivity to lower frequency (~2kHz) which makes the crossover easier. Perhaps a starting point is to choose the effective diameter of the WG (i.e. ignoring the roll-over surround) ~1/2 the inner ring diameter. The SB tweeter WG looks like ~60mm diameter, or smaller and a 120 or even 150mm ring is hard to achieve.
I found that only a little simulation is really needed for the array itself. The mids act like point sources, so each ring of array behaves much like an emitting surface matching the diameter of the ring (centers). I often play around with "Edge" to aid what's going on in my "mellon", but the usual rules of thumb for directivity of (say) a 15" drive can be used, or the equations for a rigid piston of the chosen diameter. In Edge, the baffle can be set to either its true size or the size of the rear wall, as required.
Ken
Danley fan here...
Strictly speaking, multiple entry horns predate Danley, although some of us think he perfected the idea. I have no building skills myself (or soldering, hence my handle here.) I was lucky enough to get a used pair of Yorkville Unity U15 which is sort of a poor cousin to the Danley Synergy. I made some simple mods, run them active, and they are easily the best speakers I've ever owned, and I've owned quite a few (nothing over about $2,000 a pair of value tho.)
It's probably too late for you, but allow me to pontificate upon the virtues of Danley's speakers.
Some comments on Unity/Synergy: While the Geddes speakers are very highly rated here, they cannot evade the problem of non-coherent summing of drivers ( = comb filtering and inconsistent off-axis response.) In contrast, the Danley speaker addresses all these, being a virtual point source from (in Yorkville case) about 300 Hz up. That is very tough for any speaker to do. This isn't to say that other technologies won't yield excellent results, but they won't get you the coherence like Danley brings to the table, er, front of living room
Finally, HOM is controversial. I will give Geddes benefit of the doubt. Nothing stops you (except patent law) from putting your own foam or similar treatment into a horn. This is something easily experimented with, and benefits non-MEH as well (like Geddes's).
Strictly speaking, multiple entry horns predate Danley, although some of us think he perfected the idea. I have no building skills myself (or soldering, hence my handle here.) I was lucky enough to get a used pair of Yorkville Unity U15 which is sort of a poor cousin to the Danley Synergy. I made some simple mods, run them active, and they are easily the best speakers I've ever owned, and I've owned quite a few (nothing over about $2,000 a pair of value tho.)
It's probably too late for you, but allow me to pontificate upon the virtues of Danley's speakers.
Some comments on Unity/Synergy: While the Geddes speakers are very highly rated here, they cannot evade the problem of non-coherent summing of drivers ( = comb filtering and inconsistent off-axis response.) In contrast, the Danley speaker addresses all these, being a virtual point source from (in Yorkville case) about 300 Hz up. That is very tough for any speaker to do. This isn't to say that other technologies won't yield excellent results, but they won't get you the coherence like Danley brings to the table, er, front of living room
Finally, HOM is controversial. I will give Geddes benefit of the doubt. Nothing stops you (except patent law) from putting your own foam or similar treatment into a horn. This is something easily experimented with, and benefits non-MEH as well (like Geddes's).
Last edited:
Thank you, but we're exploring a different path here, one less frequently followed.
I don't see MEH as a good solution for my situation (a particular small room). That's in contrast with PA applications where the advantages are manifold. I'm not going to explain my reasoning here, so far off topic, after all there are thousands of posts on MEH and I've nothing to add. I believe I've read every thread. There are even a couple of good designs. In any case the prototypes I made years ago convinced me that I don't have the skills to make an MEH that would suit my needs.
I considered buying a pair of the U15s which were available used - I have no regret that I let them pass by. Geddes speakers served me well for almost a decade too. I traced their interactions with my room in some detail.
Consider wide and shallow speakers as being a design requirement in this thread.
Ken
I don't see MEH as a good solution for my situation (a particular small room). That's in contrast with PA applications where the advantages are manifold. I'm not going to explain my reasoning here, so far off topic, after all there are thousands of posts on MEH and I've nothing to add. I believe I've read every thread. There are even a couple of good designs. In any case the prototypes I made years ago convinced me that I don't have the skills to make an MEH that would suit my needs.
I considered buying a pair of the U15s which were available used - I have no regret that I let them pass by. Geddes speakers served me well for almost a decade too. I traced their interactions with my room in some detail.
Consider wide and shallow speakers as being a design requirement in this thread.
Ken
Hi Ken,
Many thanks for your reply a few days back. Very valuable info.
I have meant to say thanks for days.
Been studying all the posts I can find on these things, while working on a first proto attempt.
Just finished the box below which is 29x29x11" interior. Same drivers as dicussed. 68 1/2lbs, yea !
Like i said earlier, don't mean to pirate your build thread...just wanted to show I'm not blowing smoke Lol
Should be able to start measurements tomorrow.....
Speaking of synergies...perhaps you've seen this Danley patent ...
US8284976B2 - Sound reproduction with improved performance characteristics
- Google Patents
The depiction of the ring array, and how Tom likens it to a synergy, has been another of the motivating factors for trying this.
There are some comments about the ring at the very end of the patent, that tie with the patent snip below..
I believe the straight-sided conical horn (by far most often horn used), is a bit of a synergy's Achilles heel.
I'm hoping the ring array gives synergy like acoustic combining, and maybe smoother pattern control.
But that hope is, as you pointed out,..... if the dang rings can get close enough, and ring circumferences can match wavelengths, etc...
Synergy issues deja vu
Many thanks for your reply a few days back. Very valuable info.
I have meant to say thanks for days.
Been studying all the posts I can find on these things, while working on a first proto attempt.
Just finished the box below which is 29x29x11" interior. Same drivers as dicussed. 68 1/2lbs, yea !
Like i said earlier, don't mean to pirate your build thread...just wanted to show I'm not blowing smoke Lol
Should be able to start measurements tomorrow.....
Speaking of synergies...perhaps you've seen this Danley patent ...
US8284976B2 - Sound reproduction with improved performance characteristics
- Google Patents
The depiction of the ring array, and how Tom likens it to a synergy, has been another of the motivating factors for trying this.
There are some comments about the ring at the very end of the patent, that tie with the patent snip below..
I believe the straight-sided conical horn (by far most often horn used), is a bit of a synergy's Achilles heel.
I'm hoping the ring array gives synergy like acoustic combining, and maybe smoother pattern control.
But that hope is, as you pointed out,..... if the dang rings can get close enough, and ring circumferences can match wavelengths, etc...
Synergy issues deja vu
Attachments
Nice
Wow - looks fantastic. Have fun with the measurements. Remember you need to allow enough distance for the individual wavefronts to blend (I'm sure you know that). Of course that makes it harder as a longer window is needed...
Nope - I think that's new to me.
Ken
ps. the ring of 10FE200s does look particularly impressive
Wow - looks fantastic. Have fun with the measurements. Remember you need to allow enough distance for the individual wavefronts to blend (I'm sure you know that). Of course that makes it harder as a longer window is needed...
Nope - I think that's new to me.
Ken
ps. the ring of 10FE200s does look particularly impressive
Last edited:
Thx for the compliments Ken
i wish they were deserved...measurements were horrible !
First thing I learned was why all the fuss about diffraction....
The SG tweeter with wave guide was soooo bad, this morning I made a standard test baffle to see what's up. Until i recessed it to flush mount with the baffle, it still totally sucked.
Once recessed, it turned beautiful. (I'm a CD/horn guy...diffraction is so much less of an issue with them, i had no experience here...)
So it's obvious I need to make a new baffle for whole project....
I'll recess them all...
With regard to the 3" ring, the center of the 3fe25's is about 3 3/4" from tweeter center.
The 5 driver ring consistently gives two closely spaced sharp impulse peaks, about .11ms apart. If i kill response above 6kHz, the two peaks merge into one broader peak. I need to investigate this...but am going to get new baffle done first.
On both the 3" ring and 10" ring, I have 4 drivers on one amp channel, and a single driver on another channel to be able to compare ring vs single.
The 3" single is fine, ring goes pattern whacky above 1-2kHz...my fear is that the spacing is simply to great for use above there...hopefully new baffle changes things.
Tens worked fine, holding pattern nicely though about 400Hz.
Anyway, hope you don't mind all this in your thread. My build simply isn't worth a new thread till it's at least halfway successful .
And since you already get what I'm finding out....i figure if nothing else we all get to keep learning a little from what you've found out about these type arrays.
i wish they were deserved...measurements were horrible !
First thing I learned was why all the fuss about diffraction....
The SG tweeter with wave guide was soooo bad, this morning I made a standard test baffle to see what's up. Until i recessed it to flush mount with the baffle, it still totally sucked.
Once recessed, it turned beautiful. (I'm a CD/horn guy...diffraction is so much less of an issue with them, i had no experience here...)
So it's obvious I need to make a new baffle for whole project....
I'll recess them all...
With regard to the 3" ring, the center of the 3fe25's is about 3 3/4" from tweeter center.
The 5 driver ring consistently gives two closely spaced sharp impulse peaks, about .11ms apart. If i kill response above 6kHz, the two peaks merge into one broader peak. I need to investigate this...but am going to get new baffle done first.
On both the 3" ring and 10" ring, I have 4 drivers on one amp channel, and a single driver on another channel to be able to compare ring vs single.
The 3" single is fine, ring goes pattern whacky above 1-2kHz...my fear is that the spacing is simply to great for use above there...hopefully new baffle changes things.
Tens worked fine, holding pattern nicely though about 400Hz.
Anyway, hope you don't mind all this in your thread. My build simply isn't worth a new thread till it's at least halfway successful .
And since you already get what I'm finding out....i figure if nothing else we all get to keep learning a little from what you've found out about these type arrays.