So this is an idea of mine. I've never built a multi-way before so I'm needing feedback before going on with the design.
I've delved deep into the great works of Gainphile, Gedlee, Harman, and Linkwitz. It's pretty obvious to me that the most important characteristic a speaker can have is constant directivity; especially from ~700hz - 7khz according to Gedlee.
Gainphile and Linkwitz focused on DSP controlled dipoles. Gedlee focuses on waveguides with foam plugs to control HOM's. Harman uses some pretty crazy waveguides.
I was thinking that, conceptually, there is actually a much easier way. Speaker driver beaming points are very well known and mathematically extremely simple to calculate. At and below ka=2, a driver will be omnidirectional. Tweeters are almost all 1". Thankfully this is perfect as it allows us to be omnidirectional all the way up to 8.5khz. We only really care to get up to 7khz, anything more is the cherry on top.
Now ideally in my mind, we would cover close to, or all of the 700hz-7khz bandwidth with a single driver to maintain coherency in this critical region. There are a few tweeters today that have low Fs and great FR linearity to Fs. Here is one of them.
SB Acoustics :: SB21RDC-C000-4
Now I guess all we have to do is crossover @ 700hz with an 8" driver with similar sensitivity, and call it a day? You in theory, would have omnidirectional constant directivity from the lowest frequencies of your 8" driver, to 8.5khz. All without DSP, bi-amping, horns/waveguides, or complicated OB solutions.
What do you guys think?
I've delved deep into the great works of Gainphile, Gedlee, Harman, and Linkwitz. It's pretty obvious to me that the most important characteristic a speaker can have is constant directivity; especially from ~700hz - 7khz according to Gedlee.
Gainphile and Linkwitz focused on DSP controlled dipoles. Gedlee focuses on waveguides with foam plugs to control HOM's. Harman uses some pretty crazy waveguides.
I was thinking that, conceptually, there is actually a much easier way. Speaker driver beaming points are very well known and mathematically extremely simple to calculate. At and below ka=2, a driver will be omnidirectional. Tweeters are almost all 1". Thankfully this is perfect as it allows us to be omnidirectional all the way up to 8.5khz. We only really care to get up to 7khz, anything more is the cherry on top.
Now ideally in my mind, we would cover close to, or all of the 700hz-7khz bandwidth with a single driver to maintain coherency in this critical region. There are a few tweeters today that have low Fs and great FR linearity to Fs. Here is one of them.
SB Acoustics :: SB21RDC-C000-4
Now I guess all we have to do is crossover @ 700hz with an 8" driver with similar sensitivity, and call it a day? You in theory, would have omnidirectional constant directivity from the lowest frequencies of your 8" driver, to 8.5khz. All without DSP, bi-amping, horns/waveguides, or complicated OB solutions.
What do you guys think?
Last edited:
Hello,
Theorically it could work but in practice it's going to be difficult:
the tweeter you linked have an FS of 760hz so you probably can't cut it at 700hz. It'll probably distort and have not enough protection to avoid destroying it.
Depending of filter order and type it is usually 2*fs for fc (sometimes lower 1.5*fs if steep slope is used).
Maybe it could be ok for protection using brickwall FIR but you'll probably have issues with sound character (increase of distortion near tweeter fs).
Don't forget you'll have diffraction loss and other cabinet related problems to take into account too.
But the idea could work using a Full Range in place of tweeter.
Or if you use some waveguide to put your tweeter on (there is a SEOS waveguide designed for that purpose) and compromise to push your fc higher, maybe using a smaller diameter mid and a woofer going three way.
Theorically it could work but in practice it's going to be difficult:
the tweeter you linked have an FS of 760hz so you probably can't cut it at 700hz. It'll probably distort and have not enough protection to avoid destroying it.
Depending of filter order and type it is usually 2*fs for fc (sometimes lower 1.5*fs if steep slope is used).
Maybe it could be ok for protection using brickwall FIR but you'll probably have issues with sound character (increase of distortion near tweeter fs).
Don't forget you'll have diffraction loss and other cabinet related problems to take into account too.
But the idea could work using a Full Range in place of tweeter.
Or if you use some waveguide to put your tweeter on (there is a SEOS waveguide designed for that purpose) and compromise to push your fc higher, maybe using a smaller diameter mid and a woofer going three way.
Last edited:
A constant wide directivity speaker sounds different from a constant narrow directivity speaker. The former is easier to obtain, the latter is not. I believe the latter is what is needed. Geddes has explained this very nicely in his papers.
Also, upward firing tweeter can probably give even smoother wide constant directivity.
To get more xmax to cross lower, use a 2.5in stiff aluminum cone full range like an SB65WBAC25-4. Put it behind a throat and waveguide to pinch it off at 1.4in dia throat. Then you get closer to constant directivity like shown here with Bushmeister's horn.
Thread here: http://www.diyaudio.com/forums/multi-way/285030-bookshelf-multi-way-point-source-horn.html
Thread here: http://www.diyaudio.com/forums/multi-way/285030-bookshelf-multi-way-point-source-horn.html
"The open baffle gives significantly wider dispersion at high frequencies than a box speaker with the same 8" driver on the same baffle, see Ref. 2, Design Models F. This is caused by the radiation from the rear of the baffle, having undergone enough phase shift as it comes around to the front, so that it adds to the total sound at off-axis angles. See Dipole Models. "
A dipole has constant directivity only up to certain frequency. Above that it will show funny response especially in polars. Not a simple way!
From same SL site linked By xrk
A cone and a dome will either change directivity as linearly as shown in post #1. That graph is based on flat and rigid circular membrane simulation.
Look at this link https://www.klippel.de/fileadmin/kl...ure/Papers/KLIPPEL_Sound_Radiation_Poster.pdf
From same SL site linked By xrk
A cone and a dome will either change directivity as linearly as shown in post #1. That graph is based on flat and rigid circular membrane simulation.
Look at this link https://www.klippel.de/fileadmin/kl...ure/Papers/KLIPPEL_Sound_Radiation_Poster.pdf
Just wanted to point out that ka = 2 does not mean omnidirectional. ka = 2 means it is theoretically possible to get omnidirectional dispersion given a perfect piston. In reality, ka = 1 is more accurate of when typical drivers will still maintain omnidirectionality, so about 4.2kHz for a 1'' tweeter. By 8.5kHz most tweeters are at least 5dB down at 90 degrees.
So this is an idea of mine. I've never built a multi-way before so I'm needing feedback before going on with the design.
I've delved deep into the great works of Gainphile, Gedlee, Harman, and Linkwitz. It's pretty obvious to me that the most important characteristic a speaker can have is constant directivity; especially from ~700hz - 7khz according to Gedlee.
Gainphile and Linkwitz focused on DSP controlled dipoles. Gedlee focuses on waveguides with foam plugs to control HOM's. Harman uses some pretty crazy waveguides.
I was thinking that, conceptually, there is actually a much easier way. Speaker driver beaming points are very well known and mathematically extremely simple to calculate. At and below ka=2, a driver will be omnidirectional. Tweeters are almost all 1". Thankfully this is perfect as it allows us to be omnidirectional all the way up to 8.5khz. We only really care to get up to 7khz, anything more is the cherry on top.
Now ideally in my mind, we would cover close to, or all of the 700hz-7khz bandwidth with a single driver to maintain coherency in this critical region. There are a few tweeters today that have low Fs and great FR linearity to Fs. Here is one of them.
SB Acoustics :: SB21RDC-C000-4
Now I guess all we have to do is crossover @ 700hz with an 8" driver with similar sensitivity, and call it a day? You in theory, would have omnidirectional constant directivity from the lowest frequencies of your 8" driver, to 8.5khz. All without DSP, bi-amping, horns/waveguides, or complicated OB solutions.
What do you guys think?
You might want to take a look at the Linkwitz Pluto. It's more or less designed along the lines that you are talking about, above (if I am getting the right sense of what you are saying). It's an up-firing 5 or 6" woofer and a front firing 1-inch full range driver on a stalk above the woofer. SL has data on it at his web site (see link below).
I have heard this loudspeaker and I liked its sound a lot, although I recall it being a bit SPL limited in a larger room it was in. I prefer the sound of the Pluto to the current LXmini with the 3" cone driver up top.
Link to Pluto specs and measurements at Linkwitzlab.com
I would caution you against crossing over a true 1" tweeter below 1.5kHz, even more so with a ring-radiator. If you want to cross down around 700Hz (an octave below that) you need a more robust driver. Even if the FR extends below 1kHz that does NOT mean the driver should be used there. Distortion of tweeters is often increasing steeply in this FR band, with decreasing frequency. A 1"-2" full range driver will have a suspension that is better equipped for 700Hz. Here is one you might consider:
Peerless PMT40N25AL01-04
Last edited:
And exactly which compression driver (not sure what you mean by large format, maybe 1.4" and above?) would you suggest?
I've never seen a strict definition, but seems to me that 1.4" throat and above is one criterion, the other being 3" or larger diaphragms.
The EV DH1A comes to mind.
The maximum acoustic output of a 1" diaphragm around 700Hz is very low.
If you listen at a whisper the IM distortion may not be horrible.
I once derived some approximate limits but those notes are packed. But even a 3KHz limit on a 1" dome is pushing it.
I have worked toward "Constant Directivity" since the 1970s.
> constant narrow directivity speaker... is not {easier}.
A constant narrow angle is not terribly hard with a LARGE horn. IMHO much easier than constant broad directivity.
90x90 degree 1KHz up can be done with about a 12x12x12" box. A 1" throat will stay 90 degrees well above your 7KHz goal; and there are tricks I used to go higher.
I once built a very tight 35x30 900Hz-up horn to throw a small patch of speech across a very large room. (Booth to stage talkback) Face was about 2 foot across, horn was over 3 feet long. At 30 degrees, throat size was not a limit for high freq directivity.
I have pencilled a horn 60 deg at 250Hz and 30 deg above 500Hz, 4 foot mouth 8 feet long. A 4" cone would easily hold the pattern to 7KHz. At the face it would sound funny. From 40 feet away it should sound stunning. The real goal is to throw sound 4,000 feet to annoy some neighbors. (But they are not worth 4 sheets plywood and a fat 15".)
Neither of these would be at ALL pleasant living-room speakers!!
"Broad" pattern in a single driver conflicts with power output requirements.
The "conventional" multi-way systems do different things but tend to be reasonable compromises on ALL requirements.
The game-breaker would be many drivers in array. You can staple cardboard into a dodecahedron, put a 4" cone in each face, and get 360 degrees every way. It doesn't suck bad. It sucks some in interference between drivers. (And doing it in wood would be a woodworker's nightmare.)
Remember that most packaged music has been mixed/mastered on rising-directivity monitor speakers. The "Constant Directivity" goal has merit, but more in the abstract than real-world.
If you listen at a whisper the IM distortion may not be horrible.
I once derived some approximate limits but those notes are packed. But even a 3KHz limit on a 1" dome is pushing it.
I have worked toward "Constant Directivity" since the 1970s.
> constant narrow directivity speaker... is not {easier}.
A constant narrow angle is not terribly hard with a LARGE horn. IMHO much easier than constant broad directivity.
90x90 degree 1KHz up can be done with about a 12x12x12" box. A 1" throat will stay 90 degrees well above your 7KHz goal; and there are tricks I used to go higher.
I once built a very tight 35x30 900Hz-up horn to throw a small patch of speech across a very large room. (Booth to stage talkback) Face was about 2 foot across, horn was over 3 feet long. At 30 degrees, throat size was not a limit for high freq directivity.
I have pencilled a horn 60 deg at 250Hz and 30 deg above 500Hz, 4 foot mouth 8 feet long. A 4" cone would easily hold the pattern to 7KHz. At the face it would sound funny. From 40 feet away it should sound stunning. The real goal is to throw sound 4,000 feet to annoy some neighbors. (But they are not worth 4 sheets plywood and a fat 15".)
Neither of these would be at ALL pleasant living-room speakers!!
"Broad" pattern in a single driver conflicts with power output requirements.
The "conventional" multi-way systems do different things but tend to be reasonable compromises on ALL requirements.
The game-breaker would be many drivers in array. You can staple cardboard into a dodecahedron, put a 4" cone in each face, and get 360 degrees every way. It doesn't suck bad. It sucks some in interference between drivers. (And doing it in wood would be a woodworker's nightmare.)
Remember that most packaged music has been mixed/mastered on rising-directivity monitor speakers. The "Constant Directivity" goal has merit, but more in the abstract than real-world.
From all your recommendations, I have decided it's a better path to go with the geddes approach. The SEOS 15 & 18 are constant directivity waveguides based on Geddes' oblate spheroid design, which reduces HOM's. The SEOS 15 and 18 both allow for a very low crossover point too! The SEOS 15 waveguides are very affordable, and the Denovo DNA-360 also looks excellent for an affordable price!
Now what I'd really like to do is use a woofer that can get down to 30hz open baffle, and has high sensitivity. The Acoustic Elegance LO15 looks perfect! 97.5dB/2.83V, 33.8Hz Fs, and Qts of .82.
The crossover will be the difficult part, lol. What do you guys think?
Links to everything for anyone interested in doing the same or similar:
SEOS Constant Directivity Waveguides: The SEOS? Project DIY Sound Group
Denovo Compression Driver: Denovo Audio DNA-360 compression driver DIY Sound Group
Acoustic Elegance 15" Open Baffle Sub/Woofer: Acoustic Elegance LO15 woofer for Open Baffle Applications
Now what I'd really like to do is use a woofer that can get down to 30hz open baffle, and has high sensitivity. The Acoustic Elegance LO15 looks perfect! 97.5dB/2.83V, 33.8Hz Fs, and Qts of .82.
The crossover will be the difficult part, lol. What do you guys think?
Links to everything for anyone interested in doing the same or similar:
SEOS Constant Directivity Waveguides: The SEOS? Project DIY Sound Group
Denovo Compression Driver: Denovo Audio DNA-360 compression driver DIY Sound Group
Acoustic Elegance 15" Open Baffle Sub/Woofer: Acoustic Elegance LO15 woofer for Open Baffle Applications
Last edited:
From all your recommendations, I have decided it's a better path to go with the geddes approach. The SEOS 15 & 18 are constant directivity waveguides based on Geddes' oblate spheroid design, which reduces HOM's. The SEOS 15 and 18 both allow for a very low crossover point too! The SEOS 15 waveguides are very affordable. 
The B&C DE-250 Geddes uses is ruler flat with a very nice bandwidth.
Now what I'd really like to do is use a woofer that can get down to 30hz open baffle, and has high sensitivity. The Acoustic Elegance LO15 looks good but the FR I found online says otherwise. The SB Acoustics 15" SB42FHCL75-6 is ruler flat, but a bit lower sensitivity. An MTM can be done if I need more volume..
The crossover will be the difficult part lol. What do you guys think?
Links to everything for anyone interested in doing the same or similar:
SEOS Constant Directivity Waveguides: The SEOS? Project DIY Sound Group
B&C DE 250:
B&C DE250-8 1" Polyimide Horn Driver 8 Ohm 2/3-Bolt
SB Acoustics SB42FHCL75-6: 15" SB42FHCL75-6 :: SB Acoustics
The B&C DE-250 Geddes uses is ruler flat with a very nice bandwidth.Now what I'd really like to do is use a woofer that can get down to 30hz open baffle, and has high sensitivity. The Acoustic Elegance LO15 looks good but the FR I found online says otherwise. The SB Acoustics 15" SB42FHCL75-6 is ruler flat, but a bit lower sensitivity. An MTM can be done if I need more volume..
The crossover will be the difficult part lol. What do you guys think?
Links to everything for anyone interested in doing the same or similar:
SEOS Constant Directivity Waveguides: The SEOS? Project DIY Sound Group
B&C DE 250:
B&C DE250-8 1" Polyimide Horn Driver 8 Ohm 2/3-Bolt
SB Acoustics SB42FHCL75-6: 15" SB42FHCL75-6 :: SB Acoustics
Last edited:
Yes but given the diameter of woofers and characteristics of De250 or clone (denovo) you will have to compromise on vertical directivity of whole system.
Is it a problem? I do not think it's a no-no and this kind of configuration have more to offer than this issue. But it's my taste other could think or feel differently.
Many mtm ( some being state of the art) do compromise on centre to centre distance between drivers, even one of my reference with that kind of design: RM Monitor
In this particular RM range CTC is 2/3 wavelength at fc (approx).
Better to keep this CTC as small as possible, 1/4 wl being ideal (but need low fc to be implemented in real world especially with large driver):
Biro Technology
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.