Hello,
i would like to build a constant directivity waveguide of oblate spheroid form for beeing used with a 1 inch compression driver. The ultimate goal is to construct my own unity clone one day.
My problem is, that a compression driver has a round exit while i doubt i can build a round horn in unity size. So there has to be a transition between round and rectangular.
My questions, all in the attempt to get the best sound quality, are:
- What is the best geometry of the transition?
- How long should the transition be?
- Does someone know which exit angle the BMS 4552ND has? I want to match the angle which the horn has at the throat with it.
Thanks for reading so far, i would be happy about some help 🙂
i would like to build a constant directivity waveguide of oblate spheroid form for beeing used with a 1 inch compression driver. The ultimate goal is to construct my own unity clone one day.
My problem is, that a compression driver has a round exit while i doubt i can build a round horn in unity size. So there has to be a transition between round and rectangular.
My questions, all in the attempt to get the best sound quality, are:
- What is the best geometry of the transition?
- How long should the transition be?
- Does someone know which exit angle the BMS 4552ND has? I want to match the angle which the horn has at the throat with it.
Thanks for reading so far, i would be happy about some help 🙂
Hi,
the exit angle from the throat is determined by the expansion rate of the horn.
If you have the correct expansion rate (maybe doubling the area every 100mm along the horn), then that will be independent of the angle machined into the compression driver exit throat.
I would go from round to octagonal then to square or rectangular.
The segments making the octagonal shape can be tapered from true eight equal sides diminishing to zero sized segments as you reach square.
the exit angle from the throat is determined by the expansion rate of the horn.
If you have the correct expansion rate (maybe doubling the area every 100mm along the horn), then that will be independent of the angle machined into the compression driver exit throat.
I would go from round to octagonal then to square or rectangular.
The segments making the octagonal shape can be tapered from true eight equal sides diminishing to zero sized segments as you reach square.
SunRa said:
I found a excel sheet in an other thread about waveguides. Earl Geddes posted there too, i think it was based on his equations. I think it was called "geddes on waveguides" or somethink like that.
AndrewT said:
According to Earl Geddes, in a OS waveguide one can match the waveguide throat expansion angle to the one of the compression driver. That was also stated in the above mentioned thread. This should reduce the HOM generation of the waveguide. Then the angle rises to the real waveguide angle, which is kept constant for the bigger part of the waveguide.
I think you described a conical waveguide, which is rather identical with a OS waveguide except for the throat design. In a conical waveguide, angle matching is not possible since the expansion is constant all the way to the throat. With that, one gets a discontinuity at the throat, which generates HOMs.
Concerning the throat shape, i think in similar lines as you. When one morphs a square and a circle of equal area into another, the inbetween shape should be somethink like an octagon.
FrankWW said:
yes, i should do that.
Thanks so far for the answers. If someone has experience with building such waveguides, i would love to hear about it.
I have been through the same thoughts and concluded that I would use some kind of filling/putting or whatever you may call it ... actually I might use a lathe to make the throath in hard wood doing the transition to the squared flare, and let filling do the rest ... mouth edges will be rounded and with felt underlined ... could be tricky to get smooth though ... a full scale drawing will help a lot
Why not turn the throat until the flare becomes stright sided and then make a 12 section conical. It's make the transition much easier.
Also if you used the came throat section, you could use a router and make the conical section fron laminated sheets of MDF. Yourkville have a successful round Unity.
Also if you used the came throat section, you could use a router and make the conical section fron laminated sheets of MDF. Yourkville have a successful round Unity.
Brett said:
I just send BMS a mail, but since its weekend, it will take some days until they respond.
Tinitus & Brett, your ideas got me thinking. Maybe a round horn would be the way to go.
The throat could be done as you suggest, by drilling it out of wood.
For the the straight part of the horn, one could make a wooden rectangular case for the horn, then fill in the round shape with a base of polyurethane foam and a layer of epoxy or something like that. Terminating the case with a round wood cutout that resembles the horn mouth and throat areas on both sides would help doing the foam / epoxy part.
The throat could be done as you suggest, by drilling it out of wood.
For the the straight part of the horn, one could make a wooden rectangular case for the horn, then fill in the round shape with a base of polyurethane foam and a layer of epoxy or something like that. Terminating the case with a round wood cutout that resembles the horn mouth and throat areas on both sides would help doing the foam / epoxy part.
If you did the throat section, the conical could also be cut out of something like very thin ply ot artist's thick cardboard to test. Get it close, then make sawdust.
In case I wasn't clear before, a 12 (or 8 or whayever) like this except even easier as it's conical.
http://www.steinmusic.de/horn/artikelee.html
In case I wasn't clear before, a 12 (or 8 or whayever) like this except even easier as it's conical.
http://www.steinmusic.de/horn/artikelee.html
Sure, its a nice vintage looking horn, but maybe a bit outdated ... havent new research proved that a conical squared horn works better, when done properly with the right throath transistion and rounded mouth edges(with foam/felt) ?
Conical, maybe. Square? Where? I was only giving a suggested construction technique.tinitus said:
Exactly. With a turned oblate throat section and a 12 sided 'round' conical the transition from throat to conical would be easier to smooth. Once you decide on the design, the 12 pieces are all identical so easier to make.
Small problem is that a "round" shape takes up a bit more space than a quadrangular 😀 this vintage horn would surely be more fun to build, but I believe that transition to the throath part will actually be more difficult, but nothing that cant be solved
MaVO
If one starts with a source that is to small to produce any directivity on its own (less than about 1/3 wl across), it radiates into space as a spherical wave.
If one takes a straight sided conic horn and places that point source at the apex, it radiates a segment of a spherical wave.
There is no opportunity to produce or radiate higher order modes in this condition but when ever a curved wall horn changes wall angle, there is the chance of producing higher modes. The OS horn does this change in angle as slowly as possible to minimize / eliminate the effect.
In the real world, one finds that 20KHz has a wavelength of about 5/8 inch, already smaller than the one inch throat on compression drivers.
Also, all drivers are not equal, larger 1.4 inch and even more so 2 inch drivers have an internal geometry which produces a converging wavefront at its summation point, what one wants for a horn is a diverging wavefront.
The BMS driver mentioned (which I use in some of our products at work), has a different internal structure which does not involve multiple paths which sum in a converging wave but has one summation path ending at the center at about 3/8 inch dia.
From there forward inside the driver is a conical expansion which produces a diverging wavefront.
When coupling into a more rapid conical horn (then the inside of the driver), the idea is o do it quickly before the dimension get much larger.
In practice, if you make a square horn with a one inch driver, using a one inch pipe to spread the filler in the corner works well and forms a round shape.
You mention the Unity horn, in working on it and its later offshoot the Synergy horn, a strange thing had occurred to me.
This was that there is a form of coloration, which I have not heard described before.
As the all in one horns got better and actually summed as one source, each improvement usually resulted in a reduction in the ability to tell exactly where and how far away the speaker was. While this was interesting as you could walk up and stick your head in the speaker and still not hear any separation of the image into separate ranges, the cool part was when in stereo.
Reducing the speakers’ location signature so much had a dramatic effect in improving the stereo image. The driver to driver self interference that is present in multiway speakers is what produces the Q’s that your brain uses to tell where the source is independent of the signal.
While it has taken some 7 years or so, the current Synergy horns like an SH-50 can reproduce a square wave over a decade in range, covering all three drive / frequency ranges and radiates as if it were one source with constant directivity.
Best,
Tom Danley
If one starts with a source that is to small to produce any directivity on its own (less than about 1/3 wl across), it radiates into space as a spherical wave.
If one takes a straight sided conic horn and places that point source at the apex, it radiates a segment of a spherical wave.
There is no opportunity to produce or radiate higher order modes in this condition but when ever a curved wall horn changes wall angle, there is the chance of producing higher modes. The OS horn does this change in angle as slowly as possible to minimize / eliminate the effect.
In the real world, one finds that 20KHz has a wavelength of about 5/8 inch, already smaller than the one inch throat on compression drivers.
Also, all drivers are not equal, larger 1.4 inch and even more so 2 inch drivers have an internal geometry which produces a converging wavefront at its summation point, what one wants for a horn is a diverging wavefront.
The BMS driver mentioned (which I use in some of our products at work), has a different internal structure which does not involve multiple paths which sum in a converging wave but has one summation path ending at the center at about 3/8 inch dia.
From there forward inside the driver is a conical expansion which produces a diverging wavefront.
When coupling into a more rapid conical horn (then the inside of the driver), the idea is o do it quickly before the dimension get much larger.
In practice, if you make a square horn with a one inch driver, using a one inch pipe to spread the filler in the corner works well and forms a round shape.
You mention the Unity horn, in working on it and its later offshoot the Synergy horn, a strange thing had occurred to me.
This was that there is a form of coloration, which I have not heard described before.
As the all in one horns got better and actually summed as one source, each improvement usually resulted in a reduction in the ability to tell exactly where and how far away the speaker was. While this was interesting as you could walk up and stick your head in the speaker and still not hear any separation of the image into separate ranges, the cool part was when in stereo.
Reducing the speakers’ location signature so much had a dramatic effect in improving the stereo image. The driver to driver self interference that is present in multiway speakers is what produces the Q’s that your brain uses to tell where the source is independent of the signal.
While it has taken some 7 years or so, the current Synergy horns like an SH-50 can reproduce a square wave over a decade in range, covering all three drive / frequency ranges and radiates as if it were one source with constant directivity.
Best,
Tom Danley
Hi Tom, well I guess I understand about half of it, maybe due to forein language
But what I really dont understand about the unity synergy(tapped) horn ... the one with a center tweeter and 4 mids placed around it ... normally we would demand a smooth surface and transition in the horn, does that mean that the mids can only be CD drivers or is it possible to use smaller hifi fullrange drivers without upsetting the horn ?
Is there any simple answer to why its possible to mount the mids in the middle of the horn, and even woofers, when the theory say that the transition(like that word) of the throath and mouth is important to the function of the horn
http://www.danleysoundlabs.com/pdf/danley_tapped.pdf
But what I really dont understand about the unity synergy(tapped) horn ... the one with a center tweeter and 4 mids placed around it ... normally we would demand a smooth surface and transition in the horn, does that mean that the mids can only be CD drivers or is it possible to use smaller hifi fullrange drivers without upsetting the horn ?
Is there any simple answer to why its possible to mount the mids in the middle of the horn, and even woofers, when the theory say that the transition(like that word) of the throath and mouth is important to the function of the horn
http://www.danleysoundlabs.com/pdf/danley_tapped.pdf
Hi
The Synergy horn is the further extension of the Unity horn where one horn body is driven from multiple points. The Tapped horn which allows one to make a smaller bass horn is an offshoot of working on the Unity / Synergy horns which places both faces of the driver in the horn passageway.
“What the sound wants” depends on acoustic size, at a very small dimension (compared to wavelength) sound can travel flawlessly through a tiny capillary tube coiled up in knots, at a much larger dimension sound will reflect off a surface. In other words, it travels like a pressure or a wave depending on acoustic dimension.
In the horns above, one doesn’t want a discontinuity in the horn path, you don’t want anything which causes the sound to have to bend or change directions once it has reached a large acoustic size.
The sound from the side-mounted drivers enters the horn where its expansion rate is suitable for that frequency range (keeping in mind that in part a horn is a high pass filter based on the rate of expansion).
The small holes or in some units slots that the sound enters into the horn by serve two purposes.
The size of the hole is made as small as possible to minimize the horn discontinuity and the air in the holes combined with the trapped air below the cone form a low pass acoustic filter.
Sound the driver produces above that filter is significantly attenuated or block from loading the horn. That filter is placed just above the electrical crossover low pass and its function is then to greatly reduce the harmonic distortion by blocking those upper harmonic frequencies from entering the horn and being radiated.
“Is there any simple answer to why its possible to mount the mids in the middle of the horn, and even woofers, when the theory say that the transition(like that word) of the throath and mouth is important to the function of the horn”
Theory describes what they measured at the time.
One can drive a horn anywhere along its length and it behaves similarly to the end drive case with a couple exceptions.
First, as one goes up in frequency, there will be a point where the closed end of the horn is ¼ wl away. At this point, there is a reflected signal back from the throat which causes a deep cancellation notch.
As one increases the frequency further, one finds the output is progressively self canceled by the reflected signal. This combined with the acoustic low pass filter sharply limits the hf content of the side mounted drivers.
Below that point is where the drivers are actually used with that first notch being above the low pass corner.
Second, as one moves to a mid horn position, the driver parameters which are ideal for that frequency range shift from those of the end mounted case.
So far as full range drivers, keep in mind that with the acoustic filter (holes and trapped volume), the driver must retain piston mode operation, more complex modes do not make it out the holes and into the horn path.
So far as the number of mid drivers, symmetry is your friend and on a round horn like the Yorkville U-15 and U-215, only 3 mid drivers are used on a round horn and it has slots instead of holes.
The key is that where each range taps in, the dimensions across the horn must be about a quarter wl or less at the highest frequency each range produces.
This way, one has insured that the conical horn is driven by a point source with no directivity, allowing the horn to be the pattern control and not driver-to-driver interference or spacing. With the hf driver farthest to the rear and lf drivers most forward, one can also eliminate the phase shift that normal crossovers produce, ending with something like a linear phase FIR filter in DSP although its made with passive parts and physical spacing.
Hope that helps,
Tom Danley
There is a CLF viewer and data file for the SH-50 (and others) that allows you to examine the radiation pattern and response at any angle etc if interested.
The Synergy horn is the further extension of the Unity horn where one horn body is driven from multiple points. The Tapped horn which allows one to make a smaller bass horn is an offshoot of working on the Unity / Synergy horns which places both faces of the driver in the horn passageway.
“What the sound wants” depends on acoustic size, at a very small dimension (compared to wavelength) sound can travel flawlessly through a tiny capillary tube coiled up in knots, at a much larger dimension sound will reflect off a surface. In other words, it travels like a pressure or a wave depending on acoustic dimension.
In the horns above, one doesn’t want a discontinuity in the horn path, you don’t want anything which causes the sound to have to bend or change directions once it has reached a large acoustic size.
The sound from the side-mounted drivers enters the horn where its expansion rate is suitable for that frequency range (keeping in mind that in part a horn is a high pass filter based on the rate of expansion).
The small holes or in some units slots that the sound enters into the horn by serve two purposes.
The size of the hole is made as small as possible to minimize the horn discontinuity and the air in the holes combined with the trapped air below the cone form a low pass acoustic filter.
Sound the driver produces above that filter is significantly attenuated or block from loading the horn. That filter is placed just above the electrical crossover low pass and its function is then to greatly reduce the harmonic distortion by blocking those upper harmonic frequencies from entering the horn and being radiated.
“Is there any simple answer to why its possible to mount the mids in the middle of the horn, and even woofers, when the theory say that the transition(like that word) of the throath and mouth is important to the function of the horn”
Theory describes what they measured at the time.
One can drive a horn anywhere along its length and it behaves similarly to the end drive case with a couple exceptions.
First, as one goes up in frequency, there will be a point where the closed end of the horn is ¼ wl away. At this point, there is a reflected signal back from the throat which causes a deep cancellation notch.
As one increases the frequency further, one finds the output is progressively self canceled by the reflected signal. This combined with the acoustic low pass filter sharply limits the hf content of the side mounted drivers.
Below that point is where the drivers are actually used with that first notch being above the low pass corner.
Second, as one moves to a mid horn position, the driver parameters which are ideal for that frequency range shift from those of the end mounted case.
So far as full range drivers, keep in mind that with the acoustic filter (holes and trapped volume), the driver must retain piston mode operation, more complex modes do not make it out the holes and into the horn path.
So far as the number of mid drivers, symmetry is your friend and on a round horn like the Yorkville U-15 and U-215, only 3 mid drivers are used on a round horn and it has slots instead of holes.
The key is that where each range taps in, the dimensions across the horn must be about a quarter wl or less at the highest frequency each range produces.
This way, one has insured that the conical horn is driven by a point source with no directivity, allowing the horn to be the pattern control and not driver-to-driver interference or spacing. With the hf driver farthest to the rear and lf drivers most forward, one can also eliminate the phase shift that normal crossovers produce, ending with something like a linear phase FIR filter in DSP although its made with passive parts and physical spacing.
Hope that helps,
Tom Danley
There is a CLF viewer and data file for the SH-50 (and others) that allows you to examine the radiation pattern and response at any angle etc if interested.
- Status
- Not open for further replies.