Omnidirectional Speaker Project, any interest, help?

[Martyn]

You might need to stiffen up the strips supporting the tweeter - I'd guess they're going to buzz a little.

[Martyn]

I agree with you about the horn, but a horn is attempting to control the dispersion of a more-or-less unidirectional wave. Here we have a hemispherical (at least) source with an obstruction partially blocking the path of the wave. I think it'll take someone a lot smarter than I am to model what's going on. On the other hand, there are lots of people who fit that description...

[dlneubec]

Ok, here are some more photo's that show the details of the modular construction idea (see below).

The first photo shows the top baffle alone. As you might image, change this piece and you can accommodate a different size speaker or an array of speakers. It will be held in place with socket head bolts and hurricane nuts that are embedded in the backer piece below. Adjust the length of the Aluminum sleeves and you can change the height of the deflector ball, relative to the woofer, tweeter or both, or change to a completely different type of deflector solution. So, there's lots of potential for adjustment and experimentation here. Also, I've cut 13º chamfers with the table saw to help with potential diffraction. I might still either round over the outside edges a little more or cut another short 45º chamfer here.

The second photo shows the stacked box with the top baffle, baffle backer piece and the sealed top box sitting next to it. Swap in an unsealed, brace-type box and move the sealed section one down and you now have about a 19.2L sealed box rather than a 9.6L. What is not installed yet, are 45º angle blocks in the corners of the upper sealed box. They will be glued in and hurricane nuts installed, so the baffer backer can be held in place with bolts, no need for glue. I envision adding a 1/8"x1/2" closed cell speaker-type gasket around the top of the box to provide for a good seal. Also, I can reduce the volume of this box to less than 9.6L by simply installing some filler pieces, etc.

In the next photo, you can see the open shelf brace-type boxes stacked to make the remaining bottom portion of the cabinet. Change any of these to a solid bottomed section and you change the volume in the cabinet. Right now it is will be a 38.2L ported cabinet below a 9.6L sealed cabinet. The ports have not been added yet, nor has the downfiring RS225 been installed. I envision each box to be held in place as described above with bolts and hurricane nuts with speaker gasket added at each level to maintain a good seal without gluing anything except the bottom piece of each brace-type box.

Also note, if you wanted to add a front firing driver, you would simply have to make a section with an appropriately modified brace for the driver to fit within, so mixing omnidirectional with other options is doable as well.

The next photo shows how the shelf bracing is done with rabbit joints all around. In addition, I recessed the self brace 1/8" in from the exterior for a couple reasons. I thought it would look interesting in a kind of "form follows function" way, but also so if you wanted to be able to take the speaker apart and rebuild it after final finishing, you would need a break in the finish to allow it to be taken apart. This recessed area could allow that to happen. Also, if you wanted to finish it out and not take it apart, you could install a contrasting piece of hardwood trimmed to fit this 1/4" slot for a really interesting look, say cherry on the main box and maple inserts.

The last photo show the box reassembled with two shelf-brace sections left out, so now you have a 9.6L sealed box over a 19.2L box that is could sit on stands or on a sub. Hopefully this gives tou an idea how this concept can be very flexible.

The box is a tad wide at 13" sqaure, IMO, but I made it this size so that you could fit in as much as a 12" driver and also so you could do an array of drivers with a tweeter in the middle, as discussed previously, with about a maximum size around the RS150 for an array. The interesting thing about the array concept is that you might be able to add a powered 12" sub on the bottom, crossing to an array at 100hz to 300hz range, and have a pretty high SPL solution. I don't have any idea yet if the array concept will work in an omnidirection format, but if so, I would imagine it would take much bigger deflector, more like the ones on the Duevel's which in this case would be around 13" wide.

Ok, now I'm exhausted from all this typing and too lazy to reread this, so I hope there aren't too many typos, etc.

I'm open to anyone input and ideas here.

[dlneubec]

Wow.. you work fast! I do like the modular idea....

I was going to mention that for rough surfaced sytrofoam (though this looks like the finer stuff) you would want it very smooth for the tweeter... looks like you were already thinking of that for the woofer, and the tweeter is wood. Higher frequencies = shorter wavelengths = need smoother surfaces.

My other guess is that a larger diameter for the tweeter might work better... it will be interesting to see how the effects change with different size/shapes of reflectors.. Hmmm.. I am basically set up for measurment, need to clear some working space in the garage and get at it... (for other projects as well!)

Overall though, just diving in and listening to various things can tell you a lot! Have at it and let us know...

I see the styrafoam is an easy short term solution. I'm thinking wood, maybe in the for of a wood bowl, with the base sanded or ground off would be a possibility. Someone posted a link to a set of wood bowls that target sells above.

I used the same size ball for the tweeter as described in the Morrison text, assuming they had done some experimenting, but I also picked up 2" hardwood balls and an egg shaped ball for experimentation from the local Hobby Lobby.

I'm diving in, but it would be great to get someone involved who is setup to do measurements, etc. Attached is a pdf of the plans I built this from.

[dlneubec]

You might need to stiffen up the strips supporting the tweeter - I'd guess they're going to buzz a little.

I'd say you are right. The aluminum is only 1/16" thick. I was wanting to get 1/8" stock, but didn't find it. You could also add additional braces around the drive to stiffen it up. It is actually seems stiffer than I expected it to be.

[RonS]

Spent a little while this afternoon starting my test mules. This is strictly so that I can perform some measurements. If I go ahead with this, the final design will be a rectanglular box similar to the Morrisson design. I'm not sure what I'm going to do for the semi-sphere, I want to see how it measures without (Linkwitz doesn't use anything for his Pluto). I'll also hold some different shaped objects above the woofer and measure those (I'll hold them in place with a light stand and some grip arms, I'm a photographer). The woofer is a Vifa MG22 carbon fibre because I had an extra pair kicking around, and they go nicely with the sonotube Overall height is 36" to the top of the MDF lip. I'm also going to experiment with tweeter position, either pointing up with some sort of diffuser or maybe directed forward, or some sort of angle. Measurements will dictate the direction I go.

[dlneubec]

I did a little shopping after work. At my local True Value hardware store, I found a 7-5/8" plastic half-dome, intended as a light fixture. All it had was a 1/4" hole though the top. I taped over the hole and puddled some epoxy in it and it is drying now. It think, adheared to a wood disk on it's inside edge, it will work very good for the RS225 that I will have downfiring. However, it is only about 1/8" thick or less, so I need to find a good way to fill it so it will not vibrate. I've had two ideas. One is be to fill it with sand and then glue in the wood base to hold the sand in. The other is be to glue in the wood base, with a 1/4" hole or so in it and then use that foam filler stuff that expands to fill space and fill the half dome full.

Anybody have any other ideas?

I also found a full plastic globe also for a light fixture, but it is a little big at 6" diameter, assuming Morrison has it right. They say they use a 4.75" dome for a 6.5" driver (probably has about 4.75"-5" actual cone diameter like the RS180). Still, the plastic globe would be easy to cut and could be treated just like the half dome above. They were lest than $10 each.

At Target, I found a small, fairly round bottom bowl. It turned out to be painted plastic. I tried to sand it down on my belt sander to a more perfectly round shape, but eventually wore a hole through in one small spot. Again, I covered it with tape and puddled a bunch of epoxy in the bottom, which is now drying. It is about 4-7/8", which looked to fit perfectly against the wood mdf circle I cut and mounted the tweeter in. Again, I need a way to fill it to make it more solid, but otherwise I think it may work well. It cost $5.99 I think.

I wonder if you looked online a bit if you could find similar plasic half-domes for light fixtures. I think I'll take a look tonight.

[Dennis H]

I still think a cone shape or an elliptical solid (egg shape) make more sense than a sphere. The sphere has a big relatively flat spot on the bottom that will reflect the sound right back at the driver.

[JonP]

I see the styrafoam is an easy short term solution. I'm thinking wood, maybe in the for of a wood bowl, with the base sanded or ground off would be a possibility. Someone posted a link to a set of wood bowls that target sells above.

I used the same size ball for the tweeter as described in the Morrison text, assuming they had done some experimenting, but I also picked up 2" hardwood balls and an egg shaped ball for experimentation from the local Hobby Lobby.

I was looking around stores with my eyes "tuned" for speaker enclosure possibilites a while back.. good places are import stores (Pier 1, etc) container stores, and of course the hobby and crafts places. Interesting what you can turn up for a particular use, that was intended for something completly different... I'll have to put on my "acoustic reflector" colored glasses and look around...

I would think a reflector wouldn't need large mass, or even to be very strong, just able to be mountable and have a fairly (or get one put on) smooth surface, like you did in painting the styrofoam. More durable would be better, for the final product of course. And it should be massy enough to not vibrate easily.

The wood ball supplier mentioned earlier had the 3" (larger?) for moderately pricey, but half balls, and even balls with a flat spot on them were much cheaper. Guess they're easier to machine that way... but for us, they are better than a pure smooth ball. Your thinking that Morrison would have worked on it is a good point, but I keep looking at that and thinking a larger ball on the tweeter in proportion to the dome size would give you a more circular pattern. Maybe that isn't as neccicary, or dosen't sound as good. Or, maybe they just chose a partially flattened response shape, and worked with it.

For ultra dramatics, and more money, you can get mirror balls (viewing balls?) for gardens, out of stainless steel, and 6", 8" 10", 12" or thereabouts. Not very cheap though.

Ah, another thing, I like the rebuildable angle... Sounds like you can take your embedded T-nut idea and go a bit farther, say have blocks in each corner that would hold a nut, and holes thru the tops, then as you stack, you could tie each section together with 4 bolts, and the last ones going in under the bottom where the heads would be out of sight.

Too much fun... Hmmm... still have that section of 8" sonotube hanging around the garage....

[dlneubec]

I still think a cone shape or an elliptical solid (egg shape) make more sense than a sphere. The sphere has a big relatively flat spot on the bottom that will reflect the sound right back at the driver.

I tend to agree with you. I'm kind of thinking that the Duevel approach with the concave cone is probably the best. I'm exploring getting one made, so it is definitely in my plans to experiment with it.

In the meantime, I will try the dome shape and see how it works. It seems to be much more readily available in larger sizes.

BTW, I do have an egg shape hardwood piece that I can try with the tweeter, but nothing that big for the woofer.

[dlneubec]

I would think a reflector wouldn't need large mass, or even to be very strong, just able to be mountable and have a fairly (or get one put on) smooth surface, like you did in painting the styrofoam. More durable would be better, for the final product of course. And it should be massy enough to not vibrate easily.

Your thinking that Morrison would have worked on it is a good point, but I keep looking at that and thinking a larger ball on the tweeter in proportion to the dome size would give you a more circular pattern. Maybe that isn't as neccicary, or dosen't sound as good. Or, maybe they just chose a partially flattened response shape, and worked with it.

Ah, another thing, I like the rebuildable angle... Sounds like you can take your embedded T-nut idea and go a bit farther, say have blocks in each corner that would hold a nut, and holes thru the tops, then as you stack, you could tie each section together with 4 bolts, and the last ones going in under the bottom where the heads would be out of sight.

Too much fun... Hmmm... still have that section of 8" sonotube hanging around the garage....

For more mass, then maybe the option of filling the plastic domes I found with sand is the right way to go, though it might be a bit messy initially :B I suppose one could pack a bunch of plumbers rope caulk under that palstic dome to deadan vibration as well.

Right now I have 1.5" and 2" hardwood balls with a slight flat side and also an egg shape one. I'll be trying them all to see what differences I can hear.

What you describe about having blocks in each corner and 4 bolts is just what I was intending, though I was probably not clear about that in my description. We are thinking the same here.

Lets keep the ideas coming!

[mmoeller]

I tend to agree with you. I'm kind of thinking that the Duevel approach with the concave cone is probably the best. I'm exploring getting one made, so it is definitely in my plans to experiment with it.

In the meantime, I will try the dome shape and see how it works. It seems to be much more readily available in larger sizes.

BTW, I do have an egg shape hardwood piece that I can try with the tweeter, but nothing that big for the woofer.

If I were to have one of these made I would go with someone who could make a 1 off with a CNC type machine. Just for the accuary of the elliptical curvature. I'm sure there a alot of guys out there than could turn that sort of thing, but I think the amount of work that would go into turning would cost a ton more than a computer controlled process. Just a matter of time involved.
There are any number of plastics that can be machined into any shape you want. They might not have the weight your looking for though. Very interesting problem.

[dlneubec]

If I were to have one of these made I would go with someone who could make a 1 off with a CNC type machine. Just for the accuary of the elliptical curvature. I'm sure there a alot of guys out there than could turn that sort of thing, but I think the amount of work that would go into turning would cost a ton more than a computer controlled process. Just a matter of time involved.
There are any number of plastics that can be machined into any shape you want. They might not have the weight your looking for though. Very interesting problem.

I'm having my son-in-law look into using the Indiana University facilites available to him through their art department. They have a $150,000 CNC machine that can make just about anything, the way he describes it. All I had to give him was a cross-section drawing with at least one width and height dimension. It will be scanned in and adjusted to scale and the software will create a model from that cross-section, as I understand it. I should know in the next few days it that is going to work out. I just checked at the local Menards and they are now carrying 4'x8'x3/4" Baltic Birch for $59/sheet, so I might have to pick up some of that for the deflectors if this works out.

The following are the deflector shapes I have to work with at the moment. The plastic globe is about 6" dia. and would be cut in half. Either it or the 4-7/8" painted plastic bowl would be used for the upfiring RS180 and to hold the RS28. The large plastic dome is about 8" dia. and could be used for the downfiring RS225. The wood shapes are 1.5" and 1.75" balls and a 1.75" egg shape and for use with the RS28.

I also found one source for 5" and larger hardwood balls on the internet.
Premier Wood Products
I also sent an email to a company that makes commercial grade rubber balls up to 6" to see what they could do.
RubberMill
Another one might be platic domes and here is a source for them.
Global Plastic Services

I have not had much luck with wooden cones except one is available from Visaton, but I don't know the size and it would be about $100 American, without shipping!
Visaton

[JonP]

If I were to have one of these made I would go with someone who could make a 1 off with a CNC type machine. Just for the accuary of the elliptical curvature. I'm sure there a alot of guys out there than could turn that sort of thing, but I think the amount of work that would go into turning would cost a ton more than a computer controlled process. Just a matter of time involved.
There are any number of plastics that can be machined into any shape you want. They might not have the weight your looking for though. Very interesting problem.

I'm not sure, havent talked to my wood artist buddy with the lathe yet, but I think you can make templates and work from them. That could be very repeatable... Anyone "lathe aware" enough out there to know if this is the case?

There are several methods of machining, 3d printing, etc, commercially available. I researched them a bit when the original waveguide projects hit.
You can do anything, but the costs started at around $400 a piece and went from there... so that kind of did it for me. The ME guys at work didn't have any better commercial options than that.

If hand lathing to a template is doable, then that probably would be the cheapest, assuming finding a small guy who won't charge a load... And, I like the look of wood, especially the laminated BB look.

Ah, another note... Dan, if you have a Rockler store nearby, check to see if they have the special deal BB. Not sure if it's a regional thing, my local store gets in stacks of offcut pieces of 5/8" that are about 23x24"... it's 11 ply, most of it looks pretty nice. And, it's $4 a sheet... Probably can get 4 pieces of lamination per sheet, even for the largest lenses :T

[dlneubec]

I'm not sure, havent talked to my wood artist buddy with the lathe yet, but I think you can make templates and work from them. That could be very repeatable... Anyone "lathe aware" enough out there to know if this is the case?

There are several methods of machining, 3d printing, etc, commercially available. I researched them a bit when the original waveguide projects hit.
You can do anything, but the costs started at around $400 a piece and went from there... so that kind of did it for me. The ME guys at work didn't have any better commercial options than that.

If hand lathing to a template is doable, then that probably would be the cheapest, assuming finding a small guy who won't charge a load... And, I like the look of wood, especially the laminated BB look.

Ah, another note... Dan, if you have a Rockler store nearby, check to see if they have the special deal BB. Not sure if it's a regional thing, my local store gets in stacks of offcut pieces of 5/8" that are about 23x24"... it's 11 ply, most of it looks pretty nice. And, it's $4 a sheet... Probably can get 4 pieces of lamination per sheet, even for the largest lenses :T

I'm not a lathe guy, but I know 30 years ago you could use a template with a lathe, at least to create long thin peies, like table legs, so I'm sure you can now. Bowls are turned on lathe, so I assume you could template a bowl too. This is similar shape, in that it is secured on only one side of the lathe, like a bowl when it is turned. If you glued up the base of wood and provided a template, it doesn't seem like it should be that unreasonable to have one made. I like the idea of using BB for the deflectors and even if I end up with the ball-shaped deflectors, I may have some made using BB for the aethetic value.

Unfortunately, there is no Rockler at least within an hours drive that I know of. Those 23x24 sheets sounds like they would be great for this and many small speaker project. If you run it down to cost per board foot, that comes out at a cost of about $1.60/b.f. and the 13 ply full sheet is about $2.45/b.f. so that is a pretty decent savings, but not enough to drive too far for.

[dlneubec]

On another note, I'd like to hear what others think about wood materials in the modular build concept I'm using, from a resonance and acoustic standpoint.

What I'm thinking, is that since there are shelf braces or solid partitions every 4.5", that there is little need to be concerned about resonances in the cabinet. The largest unsupported side peice is 4"x11.5" and the largest unsupported span is 11.5"x11.5" for the baffle, and the baffle has a 3/4" backer piece of mdf. So, do you think you could 1) use smaller width material for the sides, say 1/2" baltic birch or mdf instead of 3/4". Alternatively, could you use solid hardwood, either 1/2" or 3/4" for the side panels? I assume that the bracing and baffle backers would at least need to stay 3/4" mdf or BB.

Acoustically, would the most important thing remaining to do be to block sound reflection off the cabinet from getting back to the driver, so would some acoutic treatment on the wall opposite the driver be the most beneficial, assuming you were modelling for minimum fill?

Anyone have any thoughts?

[sliceofhogan]

For the tweeter, couldn't you use a golfball? The dimple design might help dispersing the sound waves.

[Jed]

For some fancy egg shapes to reflect sound, find a potter to throw some forms for you. Leave the clay natural and use some sort of drywall compound layered over the surface. Sand to a smooth finish.

[dlneubec]

For the tweeter, couldn't you use a golfball? The dimple design might help dispersing the sound waves.

They are around the right size, about 1.62" as I recall. The dimples are designed to create spin and lift, but I have no idea if that effect is helpful for soundwaves or not.

[knifeinthesink]

I think someone already mention clay for a diffuser, and It would have to be damped, but that would probably be easy. I bought a $25 potters wheel for my niece last christmas at costco.

theres also modelling clay.

[dlneubec]

I think someone already mention clay for a diffuser, and It would have to be damped, but that would probably be easy. I bought a $25 potters wheel for my niece last christmas at costco.

theres also modelling clay.

I can see where clay would be easy to mold into odd spun shapes and could work well as you and Jed suggest. My biggest concerns would be, how accurate a profile could you get and how repeateable would the results be? Also how would you connect them to the wood structure and keep them removeable?

I guess if you were to glaze them and fire them, they could have a very hard smooth surface.

[dlneubec]

I heard (second hand) from my son-in-law that he will be able to make me some Duevel-like wood deflectors using IU's high end CNC system. I have to talk to him for sure about what he needs, but I assume he will need glued up bases. I will probably pick up a 4'x8' sheet of Baltic Birch to make them from.

I don't know how long it will take to make them, so I'm proceeding with the ball-shaped deflector for the time being. If all goes well, I should have one made up to audition sometime this weekend.

[Patrick Bateman]

I have a sneaking suspicion that it's not very critical.

Sorry to disagree, but the shape and the size WILL make a difference.

I listened to the speakers at CES, and I am familiar with their sound.

Here's a few things to consider.

First and foremost, you better use a high efficiency woofer. Nothing else will do. Because the design is *somewhat* omnidirectional, efficiency and/or power is essential.

Second, I said that the design is "somewhat" omnidirectional. That's because the flying saucer looking thing in actually a waveguide. I have posted extensively online about waveguides, in case anyone is interested. I've been working with horns and acoustic waveguides for fifteen years. Basically that flying saucer thing is a horn, but ROTATED around an axis. I'll bet that you could model it to some extent using hornresp or Martin King's worksheets. Just looking at the design, I can tell you that the sound will 'beam' a bit because the profile is pretty much exponential, as opposed to conical or oblate spheroidal, which won't beam as much.

Having said all that, I think it's a great project!! I'd love to see how this turns out. Omnis can sound very good. The ones at CES are using a compression driver, so I'll bet they have a lot more headroom than the less expensive Omnis, like the Decware, B&O and Mission (?).

Why not try out a BMS or B&C compression driver?

[JonW]

Dan-
This is a very cool project and I'm following along with interest. I quickly jumped in to suggest that it might sound too diffuse. But now I'm really curious to see how it turns out. I like your approach, too. Have fun. :T
-Jon

[Patrick Bateman]

Well, I got a good amount done on building and testing my modular prototype idea (see the attached pics).

I think the modular concept is going to be pretty nice. I'll report more on that later, with pics. I'm still working on the concept, but I think it may have the potential to allow me to build up various systems sort of like Lincoln logs, from TM's, to TMM's, to TMW's, with some part sealed, some ported, etc. and still disassemble them and change to another concept.

What is shown in the pics is my first concept, which is an RS180S-8 on top, firing up, into a 5" ball-shape (like the Morrison speaker). The RS180 is in a sealed cabinet at about 9L. The ball shape houses and RS128AS firing up into a 1.5" hardwood ball. The 5" ball is made from a styrafoam ball cut in half and then mounted to a MDF round. The MDF is the support for the whole thing. I painted several coats of truck bed liner spray on the styrafoam ball to close the cells. I see this as a temporary solution to allow me to test this concept. I'm anticipating trying to find another source for the 5" ball, maybe a wooden bowl, with the lip on the bottom sanded off.

The bottom 4/5's houses a RS225, firing down. It is in a 39L enclosure that will be ported, probably at the 4 corners, firing down also, but I didn't have the right ports.

I'm planning on trying a 2.5way crossover with this setup, an existing crossover made to use with two RS-180's and the Rs28, but instead the .5 will be the RS225. I got the idea that this might work from this thread, where Dennis H suggests the two might be interchangeable.
RS180 drop in for RS225 and Vice-Versa?
We'll see how it sounds anyway.

I'm working on getting a deflector made like the ones that Duevel uses and plan to experiment with this concept, as compared to the ball concept as well. I have not decided whether to use a ball or cone-type deflector for the RS225, but will probably start with a ball, since that is what I have on hand.

A couple of thoughts:

1st, note that the original speakers use a VERY small distance from the apex of the horn to the tweeter. This is to reduce reflections. As the distance gets smaller, it will clean up your response. The further you are from the apex, the lower in frequency you'll see reflections. So keep that gap small.

2nd, why didn't you copy the layout of the original speakers? The original layout is clever, because the output from the edge of each horn meets at the edge of that wood flying saucer contraption.

I'm really thinking that the original would be cloned quite easily using a $100 compression driver paired with the best prosound woofer you can find. The woofer will be critical, as it must run all the way up to where the compression driver tapers off.

[Patrick Bateman]

I keep looking at that and thinking a larger ball on the tweeter in proportion to the dome size would give you a more circular pattern. Maybe that isn't as neccicary, or dosen't sound as good. Or, maybe they just chose a partially flattened response shape, and worked with it.

The diameter of the ball will dictate the response of the driver, as well as the cutoff.

Here's some math, just off the top of my head.

Let's say you want to cross that tweeter over at 2khz. That's a wavelength of seven inches. If it were me, I would start with a mouth size of 3.5". So what I would do is buy a 7" ball, place it UNDER the tweeter, about 1/4" to 1/2" from the apex of the dome. The mirror side of the ball will serve the same function on the woofer as it does the tweeter.

Obviously the ball will act as a horn for nearly the entire output of the tweeter; but it will have a minimal effect on the woofer. But that's fine, as the woofer will go omni-directional at point that's equivalent to the diameter of the cone.

To sum it all up, you could probably make a nice design using a 7" stryofoam ball, a 8" woofer (with a 7" cone) and a crossover of 2khz.

How does that sound?

:: PB ::

[dlneubec]

A couple of thoughts:

1st, note that the original speakers use a VERY small distance from the apex of the horn to the tweeter. This is to reduce reflections. As the distance gets smaller, it will clean up your response. The further you are from the apex, the lower in frequency you'll see reflections. So keep that gap small.

2nd, why didn't you copy the layout of the original speakers? The original layout is clever, because the output from the edge of each horn meets at the edge of that wood flying saucer contraption.

I'm really thinking that the original would be cloned quite easily using a $100 compression driver paired with the best prosound woofer you can find. The woofer will be critical, as it must run all the way up to where the compression driver tapers off.

Hi Patrick,

I believe you may be correct, regarding the deflector spacing. I noticed the same thing. It would seem that the tweeter deflector should be especially close, since these are the smallest wavelengths. If you look at the Morrison example posted previously, you can see that the ball/dome deflector is very close to the tweeter. At least with a tweeter, you don't have to worry about cone excursion. I think my 1.5" ball, as I have it configured now, is probably not close enough, but I can fix that with a little aluminum tube trimming. I would think that just short of touching the RS28 mesh sheild would be what to shoot for, maybe 1/16" away. What do you think?

With mids midwooders and subwoofers, I assume that the deflector should be a little farther away. Maybe this can be determined based on the minimum wavelength for each driver. Obviously there has to at least be room for cone excursion. Don't you think there also a concern about nearly plugging the air and sound flow from a cone if you get too close, at least with the ball/dome concept?

There were actually two concepts that sort of fired this project. One was the Duevel, which uses the concave cone idea, and the one which I think will work best. However, the Morrison concept is also being considered for a couple reasons and it uses the ball/dome deflector concept. This is what I could do short term to test the overall concept and is maybe the most doable for the average DIYer. I do plan to make and test a Duevel based concept, except I do not plan to attempt the deflector/cone combination that they use in the Bella Luyna and Jupiter. On the surface, there just seems to be too much going on with the deflector/horn/compression driver combination to be a doable DIY project. My approach is to do something a bit simpler, more like the Duevel Venus, as far as the deflector goes.

The various options and concpets that could be done with this is the reason I came up with the modular box concept, so I could try one thing, dissasemble the box and build a completely different concpet.

[dlneubec]

The diameter of the ball will dictate the response of the driver, as well as the cutoff.

Here's some math, just off the top of my head.

Let's say you want to cross that tweeter over at 2khz. That's a wavelength of seven inches. If it were me, I would start with a mouth size of 3.5". So what I would do is buy a 7" ball, place it UNDER the tweeter, about 1/4" to 1/2" from the apex of the dome. The mirror side of the ball will serve the same function on the woofer as it does the tweeter.

Obviously the ball will act as a horn for nearly the entire output of the tweeter; but it will have a minimal effect on the woofer. But that's fine, as the woofer will go omni-directional at point that's equivalent to the diameter of the cone.

To sum it all up, you could probably make a nice design using a 7" stryofoam ball, a 8" woofer (with a 7" cone) and a crossover of 2khz.

How does that sound?

:: PB ::

Let me see if I undestand what you suggest. Fire the woofer up into a 7" ball and suspend the tweeter above the 7" ball firing down toward the same ball. Is that correct? That is certainly more like the Duevel approach and makes some sense.

It sounds doable, but how do you account for what the other makers of omni's are doing. Morrison, Visaton, and Morage are doing almost exactly what I am and use a much smaller dome for the tweeter deflector, except the Visaton uses a cone shape and the other two use a dome shape.
Morrison
Visaton
Mirage

Somewhere on the Morrison site, they mention using a 4.75" dia. dome deflector for the 6.5" woofer and a 1.5" dia. dome deflector for the tweeter on one of their models (maybe the model 17?). These would seem to be sensible places to start testing, don't you think?

[dlneubec]

Let me see if I undestand what you suggest. Fire the woofer up into a 7" ball and suspend the tweeter above the 7" ball firing down toward the same ball. Is that correct? That is certainly more like the Duevel approach and makes some sense.

It sounds doable, but how do you account for what the other makers of omni's are doing. Morrison, Visaton, and Morage are doing almost exactly what I am and use a much smaller dome for the tweeter deflector, except the Visaton uses a cone shape and the other two use a dome shape.
Morrison
Visaton
Mirage

Somewhere on the Morrison site, they mention using a 4.75" dia. dome deflector for the 6.5" woofer and a 1.5" dia. dome deflector for the tweeter on one of their models (maybe the model 17?). These would seem to be sensible places to start testing, don't you think?

I just remembered another example. Duevel themselves use the ball-shape deflector on their lowest end system, the "Planets". Each ball is about the size of the driver. The tweeter ball is much smaller than the woofer, and interestingly enough, they are not very close to the drivers at all, though the tweeter looks like it has a bit of horn effect to its throat area.
Duevel Planets

[mmoeller]

I just remembered another example. Duevel themselves use the ball-shape deflector on their lowest end system, the "Planets". Each ball is about the size of the driver. The tweeter ball is much smaller than the woofer, and interestingly enough, they are not very close to the drivers at all, though the tweeter looks like it has a bit of horn effect to its throat area.
Duevel Planets

I'm guessing again that the balls are located at a critical 1/4 wavelength or 3/4 wavelength. The difference could be a few inches at the frequencies were talking about.

Patrick spoke of a 2Khz bandwidth for a 7" sphere. I see the correlation of a 7" wavelength being around 2k. It seems to me that the distance to the ball is the crucial element. Not that the size of the ball is inconsequential, but it may have more to do with the amount of air, directed by the cone, that gets "redirected"

[dlneubec]

I'm guessing again that the balls are located at a critical 1/4 wavelength or 3/4 wavelength. The difference could be a few inches at the frequencies were talking about.

Patrick spoke of a 2Khz bandwidth for a 7" sphere. I see the correlation of a 7" wavelength being around 2k. It seems to me that the distance to the ball is the crucial element. Not that the size of the ball is inconsequential, but it may have more to do with the amount of air, directed by the cone, that gets "redirected"

Are you thinking that this suggests that the ball must be within 7" to be effective at 2k rather than needing a 7" diameter?

If this is the case, it might be a bit easier to apply that standard to the what minimum distances to apply for various drivers. Perhaps after that, concerns about the amount of air being re-directed shoul be applied?

[joecarrow]

These commercial designs are all very interesting, but Duevel's website has me wondering if you really want to look to them for the science of the approach.

Specifically, the workings of their speaker feet have me baffled:

http://www.cd-konzert.de/Produkte/Eaccessoires.htm

Unless you have some good FEA available to you, I really think that copying the existing designs and listening to see which you like best in your room is the best approach. Having said that, I think your test mules are dead-on as a way of figuring out what really works.

[dlneubec]

These commercial designs are all very interesting, but Duevel's website has me wondering if you really want to look to them for the science of the approach.

Specifically, the workings of their speaker feet have me baffled:

http://www.cd-konzert.de/Produkte/Eaccessoires.htm

Unless you have some good FEA available to you, I really think that copying the existing designs and listening to see which you like best in your room is the best approach. Having said that, I think your test mules are dead-on as a way of figuring out what really works.

Those are some high tech feet! My guess is that they have a rubber shock absorber between the steel pieces and the speaker. Then the ball bearing between the one disk and the next allows it to level to whatever floor situation is sits on....... Well, I tried anyway!

I hope this is the right approach, but heck, it's only a few hours, some mdf and stimulating thinking. At least the modular approach makes it adaptable for reuse without starting over, at least that's the plan.

[Patrick Bateman]

How do I explain the fact that some vendors are using smaller horns? Well that's simple!

Smaller horn = the low frequency cut off is higher.

It's as simple as that.

For example, in the "planet" speakers, the balls will have a almost negligible effect. For example, the 1" chromed ball will act as a horn, spun around a axis, whose size is the radius.

So let's do the math.

((speed of sound / length of radius) * 1/2 wavelength)=
((13800in / sec / 0.5inch) * 1/2 )=
13800hz

So basically that puny little chrome ball will do some very mild changes to the response, from 13800hz on up. I'll bet it's effect is *barely* audible.

But on the bigger models, it's effect will be pronounced, as I heard at CES.

[knifeinthesink]

I think this has been mentioned already, but Linkwitz' Pluto has no deflector, and he says that tested beside the orion it was sometimes difficult to tell which was playing.

If the results are that good without a deflector, then maybe its not neccessary.

What are the reasons for assuming from the outset that there must be a deflector.

[Dennis H]

I think this has been mentioned already, but Linkwitz' Pluto has no deflector,

Yeah, he's crossing at 1K where the 5" driver is omnidirectional so there's no need for a diffusor. And the 2" tweeter is omni up to 5K or so. So any diffusors would only be working for the top 2 octaves or so.

[JonP]

I'm guessing again that the balls are located at a critical 1/4 wavelength or 3/4 wavelength. The difference could be a few inches at the frequencies were talking about.

Patrick spoke of a 2Khz bandwidth for a 7" sphere. I see the correlation of a 7" wavelength being around 2k. It seems to me that the distance to the ball is the crucial element. Not that the size of the ball is inconsequential, but it may have more to do with the amount of air, directed by the cone, that gets "redirected"

Hi guys... I've been trying to visualize this by imagining the wave travel and the angles involved.

A certain diameter sphere nearby as a reflector will produce an angle that will get steeper or shallower with both the diameter of the sphere, and the distance from the radiating source. The same size sphere will lower the radiation angle as you get closer, and at the same distance a larger sphere will also lower the angle.

Also, at different frequencies how "big" the same diameter will be will vary, due to wavelength proportions as Patrick mentioned. So we have at least 3 variables at work.... But, as has been mentioned, sound waves are pretty "mushy" compared to light waves, and looking too closely at things in an optics sense is overdoing it. It's not going to be super focused and influenced by the lens unless it's very big at a particular wavelength..
It will be interesting to see what's the greater influence.. I'd also guess that close to the dome would be better, and less wierdness at the shortest wavelengths/highest freqencies.

Anyway... on a building items search note, I went to a local Michael's (craft store) during lunch today, found some various sized wooden balls... they have a small flat on them, as well as a hole that could hold a screw or bolt. From under 1" to 2 1/4" I got the latter and a 1 1/2". They were for "doll heads" but looked like they'd have a nice grain if sanded smooth and finished.

Michael's also had some styrofoam.. balls to 8", but the coarse, open cell stuff that you'ld have to finish. Actually they had shrink wrap around them, the resulting surface would probably work moderately well for preliminary tests, but it dosen't look nice. Wonder if plaster and mud would stick and sand out OK? I also found a small cell foam half sphere, 6", that was for floral arrainging.. not the super fine sealed cell stuff but very small pores. Very easy to damage, though.

Another nice thing they had, Xmas ornaments, clear blown glass, 3" and 4" clear spheres, pretty cheap. (4" were 4 for $5) Christmas is coming, eventually there will be lots of spherical objects out there..

I'll have to kludge up a support to put the wooden balls over a 27TDFC or RS28, and then get my measuring system into a workable condition...

[dlneubec]

Yeah, he's crossing at 1K where the 5" driver is omnidirectional so there's no need for a diffusor. And the 2" tweeter is omni up to 5K or so. So any diffusors would only be working for the top 2 octaves or so.

While not directed at the Pluto specifically, I have read somewhere that omnidirectionals that have upfiring drivers without a deflector (or diffuser, which is the correct term?) sounded thin in the midrange, possibly because there was less sound headed directly at the listener and much was redirected off the ceiling.

Perhaps this could also be a problem with valuted ceilings. It would seem that the Duevel approach is designed to ensure that the sound is radiated in more of a 360º cylinder from the speaker, with less heading toward the ceiling. Intuitively, that seems like a more logical approach to me, . I'm not certain if the ball/cone shaped diffusers will do as good a job at redirecting the sound into that type of cylinder pattern, but listening to both should tell me a lot or at least which I prefer.

[dlneubec]

Hi guys... I've been trying to visualize this by imagining the wave travel and the angles involved.

A certain diameter sphere nearby as a reflector will produce an angle that will get steeper or shallower with both the diameter of the sphere, and the distance from the radiating source. The same size sphere will lower the radiation angle as you get closer, and at the same distance a larger sphere will also lower the angle.

Also, at different frequencies how "big" the same diameter will be will vary, due to wavelength proportions as Patrick mentioned. So we have at least 3 variables at work.... But, as has been mentioned, sound waves are pretty "mushy" compared to light waves, and looking too closely at things in an optics sense is overdoing it. It's not going to be super focused and influenced by the lens unless it's very big at a particular wavelength..
It will be interesting to see what's the greater influence.. I'd also guess that close to the dome would be better, and less wierdness at the shortest wavelengths/highest freqencies.

Anyway... on a building items search note, I went to a local Michael's (craft store) during lunch today, found some various sized wooden balls... they have a small flat on them, as well as a hole that could hold a screw or bolt. From under 1" to 2 1/4" I got the latter and a 1 1/2". They were for "doll heads" but looked like they'd have a nice grain if sanded smooth and finished.

Michael's also had some styrofoam.. balls to 8", but the coarse, open cell stuff that you'ld have to finish. Actually they had shrink wrap around them, the resulting surface would probably work moderately well for preliminary tests, but it dosen't look nice. Wonder if plaster and mud would stick and sand out OK? I also found a small cell foam half sphere, 6", that was for floral arrainging.. not the super fine sealed cell stuff but very small pores. Very easy to damage, though.

Another nice thing they had, Xmas ornaments, clear blown glass, 3" and 4" clear spheres, pretty cheap. (4" were 4 for $5) Christmas is coming, eventually there will be lots of spherical objects out there..

I'll have to kludge up a support to put the wooden balls over a 27TDFC or RS28, and then get my measuring system into a workable condition...

The question seems to be is it a reflector/deflector/diffuser or a horn? I can see where it begins to effectively be a horn as it gets closer to the driver. The sides of the horn from by the bourndaries of the baffle and the diffuser, and it is a 360º horn. In that case, I can see where a larger diameter and surface would be more effective. I wonder if there would even be some db gain at some point?

On the other hand, must it act as a horn to deflect the sound in a 360º cylinder from the driver, or can the sound simply be redirected?

As far a testing, take a look at how I attached the ball to the tweeter. That is more of a finished look, in that the bolt that holds up the cross piece (that holds the ball) is covered by aluminum tubing. The tubing acts as a spacer and also looks better aesthetically. However, for testing, I will remove the aluminum spacer and use a longer bolt. I can use 3 hex nuts on the bolt to allow for esy adjustment up and down. That way I can change both the height and the ball diameter quite easily. See the attached pdf for how I intend to do this for the testing stage.

[dlneubec]

How do I explain the fact that some vendors are using smaller horns? Well that's simple!

Smaller horn = the low frequency cut off is higher.

It's as simple as that.

For example, in the "planet" speakers, the balls will have a almost negligible effect. For example, the 1" chromed ball will act as a horn, spun around a axis, whose size is the radius.

So let's do the math.

((speed of sound / length of radius) * 1/2 wavelength)=
((13800in / sec / 0.5inch) * 1/2 )=
13800hz

So basically that puny little chrome ball will do some very mild changes to the response, from 13800hz on up. I'll bet it's effect is *barely* audible.

But on the bigger models, it's effect will be pronounced, as I heard at CES.

I'm trying to understand the potential horn effect here. I think I get it. Take a look at the attached pdf. It shows how the "horn" changes from small diameter ball, further from driver; to larger diameter ball, closer to the driver; to a Duevel-like deflector. The horn shape is shown in bold black.

Does this look correct?

If so, I wonder if it also follows that the sound wave will be shaped in the direction of the horn. The Duevel-type heading more toward the listener and the ball type being more diffuse?

[mmoeller]

How do I explain the fact that some vendors are using smaller horns? Well that's simple!

Smaller horn = the low frequency cut off is higher.

It's as simple as that.

For example, in the "planet" speakers, the balls will have a almost negligible effect. For example, the 1" chromed ball will act as a horn, spun around a axis, whose size is the radius.

So let's do the math.

((speed of sound / length of radius) * 1/2 wavelength)=
((13800in / sec / 0.5inch) * 1/2 )=
13800hz

So basically that puny little chrome ball will do some very mild changes to the response, from 13800hz on up. I'll bet it's effect is *barely* audible.

But on the bigger models, it's effect will be pronounced, as I heard at CES.

I concede the fact that the size of the balls radius will have an effect. However, then how does the distance from the cone to the ball effect the "horn"? The closer it is the the smaller the entrance to the horn, and vice versa.

Check out this website. Has a great deal of information on Horn speakers. Including formulas for horn sizes and shapes. With a few assumptions you might be able to use some of these to determine size and distance of the ball.

http://melhuish.org/audio/horn.html

[knifeinthesink]

Dan,

Perhaps when you get measuring capabilities you could report the effects of going without the diffuser.

Id think that it would be possible to correct for thinness with the XO, just like we have to correct for dipole effect or baffle step in other designs. Since your tweeter can go very low, it would be interesting to experiment.

The issues around the deflector seem to be the current focus of the thread though and its cool that theres alot of discussion around a subject that is not so commonly discussed.

[Patrick Bateman]

The question seems to be is it a reflector/deflector/diffuser or a horn? I can see where it begins to effectively be a horn as it gets closer to the driver. The sides of the horn from by the bourndaries of the baffle and the diffuser, and it is a 360º horn. In that case, I can see where a larger diameter and surface would be more effective. I wonder if there would even be some db gain at some point?

On the other hand, must it act as a horn to deflect the sound in a 360º cylinder from the driver, or can the sound simply be redirected?

Is it a reflector or a deflector or a diffuser or a horn? I can't grasp that question. When I look at the duvel images, it is clearly a exponential horn that has been rotated around a 360 degree axis. (The axis is a vertical line that goes through the voice coils of each driver, from the top of the enclosure to the bottom. Imagine a vertical line perpendicular with the listener.)

If the solid wood "deflector" in the Duvel speakers was a cone, then the horn would be conical. The horn is formed by the negative space around the wooden "deflector."

And I can unequivocally state that there WILL be gain at some point; in fact I presented the math to calculate that point. If you can, see if you can get Martin King into this discussion; he is waaaaaaay beyond me when it comes to doing the math.

:: PB ::

[Patrick Bateman]

Dan,

Perhaps when you get measuring capabilities you could report the effects of going without the diffuser.

Id think that it would be possible to correct for thinness with the XO, just like we have to correct for dipole effect or baffle step in other designs. Since your tweeter can go very low, it would be interesting to experiment.

The issues around the deflector seem to be the current focus of the thread though and its cool that theres alot of discussion around a subject that is not so commonly discussed.

Duvel's solution is more elegant than tweaking the crossover. Tweaking the crossover may improve the sound at one point in the room, but Duvel's solution will cope with the power response. And Duvel's solution requires less power.

:: PB ::

[dlneubec]

Dan,

Perhaps when you get measuring capabilities you could report the effects of going without the diffuser.

Id think that it would be possible to correct for thinness with the XO, just like we have to correct for dipole effect or baffle step in other designs. Since your tweeter can go very low, it would be interesting to experiment.

The issues around the deflector seem to be the current focus of the thread though and its cool that theres alot of discussion around a subject that is not so commonly discussed.

Of course, you are right, until I get measruement capability I can only move so far forward, unless someone else who has that capability and is interested enough to build a test mule vlounteers. It's something I knew I would need to do eventually.

At this point, I know absolutely nothing about the measuring or the equipment, so to do so I will need lots of help or guidance. I do have a laptop. It is an HP Pavialion ZD8000 with a P4 3.0ghz processor, 2gb ram, running Windows XP Pro SP2. The audio is probably built into the MB, and is listed as a Conexant AC-97.

The $64,000 question is what do I need to buy and how do I set it up, from a software and hardware standpoint and also from a physical measurement standpoint. Also, what will it cost? Anybody care to advise a beginner on this?