Omnidirectional Speaker Project, any interest, help?

[dlneubec]

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?

Hey Patric, I somehow missed this post of yours. I was wondering what your background was, since you seemed very familiar with horns, etc. In this case, it's kind of like firing a driver into the backside of one of those waveguides like Zaph used in his Waveguide TMM, isn't it?
Waveguide TMM

In terms of a high efficieny main driver, what do you think of the small driver array concept? Something like four 5"-6" drivers, like the RS150-8, in a series 2x2 parallel? That gets to about 92.9db 1w/1m.
RS150-8
Add a high efficiency, small flange tweeter centered into the array, perhaps something like the Vifa XT25SC50 (listed at 92.5db)?
Vifa XT25SC50

I guess you would probably have to have a 10"-12" powered sub playing up to 100-200hz to get enough low end. Early in this process I conceptualized a couple of these (see the pdf below)

[JonP]

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.)

:: PB ::

It may not be clear that most of the thread discussion is around a Duvel Venus type of design, where the tweeter has a single reflecting shape. Dan thought that the 2 part more hornlike one with the compression driver would be overly difficult to DIY...

[JoshK]

Why not try a B&C woofer? They seem to make good bang for the buck drivers and are hi-eff. Their 6" or 8" woofers would work well for this I'd imagine. That is what I intend to try when I approach this project (for surrounds).

[Patrick Bateman]

Hey Patric, I somehow missed this post of yours. I was wondering what your background was, since you seemed very familiar with horns, etc. In this case, it's kind of like firing a driver into the backside of one of those waveguides like Zaph used in his Waveguide TMM, isn't it?
Waveguide TMM

In terms of a high efficieny main driver, what do you think of the small driver array concept? Something like four 5"-6" drivers, like the RS150-8, in a series 2x2 parallel? That gets to about 92.9db 1w/1m.
RS150-8
Add a high efficiency, small flange tweeter centered into the array, perhaps something like the Vifa XT25SC50 (listed at 92.5db)?
Vifa XT25SC50

I guess you would probably have to have a 10"-12" powered sub playing up to 100-200hz to get enough low end. Early in this process I conceptualized a couple of these (see the pdf below)

If I were you, I would avoid an array of woofers like the plague. Arrays are tricky enough WITHOUT waveguides. I'm building a unity clone right now, and honest to God, it's by far the HARDEST project I've EVER done. And I have a full set of measurement gear; I can measure response, impedance, waterfall, etc... If you want a good chance of success, I would stick with a single woofer and a single tweeter. I'd use some high efficiency kit. For my Unity I'm using a BMS 4540nd, which is $80 at assitanceaudio.com. A good woofer choice is trickier, because there aren't a lot of eights or tens that offer high efficiency and low bass. And FORGET about high efficiency sixes; they are for midrange ONLY. I'm using a B&C 8NDL (sp?) in my car, and I'm very happy with it. You can get those over at usspeakers.com iirc.

As for my background, I've been building horns for around 15 years now. I just learned most of it through trial, error, and a LOT of time spent with Martin King's worksheets from quarterwave.com.

If you want to see MY horn loaded array, it's here: my horn loaded array

[Dennis H]

And FORGET about high efficiency sixes; they are for midrange ONLY.

The original Unity is only good down to the lower midrange, e.g. 300 Hz or so, low enough that you can cross to a pro 15". If you try to go lower with the horn, the mouth size really gets huge.

[dlneubec]

Why not try a B&C woofer? They seem to make good bang for the buck drivers and are hi-eff. Their 6" or 8" woofers would work well for this I'd imagine. That is what I intend to try when I approach this project (for surrounds).

I've never used a pro driver. That's probably worth looking into, thanks.

[dlneubec]

I hope to get my modular test mule together this weekend using a few drivers I have lying around (RS180S-8, RS225S-8, RS28SA) using all dome shaped diffusers. The RS180 will be sealed and up firing into a 6" dia. dome shape that houses the RS28, which will be upfiring into a 1.5" hardwood ball. I also have a 4-7/8" dome to try as a diffuser the RS180 and a 1.75" round and 1.75" egg shaped hardwood diffusers that I can exchange with the 1.5" ball. The downfiring RS225 will fire into an 8" plastic dome. It will be ported in 39 liters, with 4 dowfiring ports, one at each corner (1-3/8" dia.), tuned to about 35hz.

I'm using plastic domes that were intended as light covers. They are smooth and hard, but only about 1/16" thick. I need to add some mass to them to try and attenuate any possible vibration. I'm looking for ideas as to what I might fill these with to accomplish that. So far, I've thought of some type of modelling clay, rope caulking, silicone caulking, silly putty, etc. It would need to be something that would not dry out and get hard, I would think. The 6" dome houses the RS28 body, or about 1" deep x 3" dia. of it, but the 8: one is entirely empty, so a farily large volume will be needed to completely fill it. I could possibly line it, for example with rope caulk, rather than fill it.

Anyone have any ideas or thoughts on filler material?

[joecarrow]

I started to think about what would be stiff, would stick to the sides, and wouldn't shrink after hardening (which could cause buzzing and rattling, I'm sure), and I came up with that yellow expanding foam you can get at the hardware store for sealing up cracks in your house.

I think the name I remember is "Great Stuff", and it comes in versions that cure extra hard and slightly pliable. I'd expect the pliable one to give more damping, but I expect that either one would work.

[dlneubec]

I started to think about what would be stiff, would stick to the sides, and wouldn't shrink after hardening (which could cause buzzing and rattling, I'm sure), and I came up with that yellow expanding foam you can get at the hardware store for sealing up cracks in your house.

I think the name I remember is "Great Stuff", and it comes in versions that cure extra hard and slightly pliable. I'd expect the pliable one to give more damping, but I expect that either one would work.

I had thought about "Great Stuff" but dismissed it becuase I did not know thay had a version that is slightly pliable. I used that when I had my house built almost 20 years ago to fill all the holes I could find to the outside or crawl space.

[dlneubec]

Well, I stopped by the hardware store at lunch. I found two kinds of "Great Stuff" but the only difference seemed to be how much they expand. Both indicated they dried to a sandable, etc. finish, no indication either remained pliable permanently.

However, I did find a tub of plumbers putty. It says it stays soft and plaiable, does not shrink, etc. and is very dense, heavy and should stick well to a smooth surface. A 3lb tub cost $4.99. I think this stuff will work quite well. I think I cab just line inside of the dome with maybe a 1/2" layer of it, rather than completely fill it up. We'll see.

Here's a thought, I wonder if you could make a deflector from "Great Stuff? Take a glass bowl or something with a dome interior shape, coat it with Pam or a similar spray on oil, fill it with Great Stuff, using the bowl as a form and then sand it smooth and spray paint it after it hardens. You would be able to cut it, route sections out as needed, etc. The only thing is, it would be very light.

[Jed]

Well, I stopped by the hardware store at lunch. I found two kinds of "Great Stuff" but the only difference seemed to be how much they expand. Both indicated they dried to a sandable, etc. finish, no indication either remained pliable permanently.

However, I did find a tub of plumbers putty. It says it stays soft and plaiable, does not shrink, etc. and is very dense, heavy and should stick well to a smooth surface. A 3lb tub cost $4.99. I think this stuff will work quite well. I think I cab just line inside of the dome with maybe a 1/2" layer of it, rather than completely fill it up. We'll see.

Here's a thought, I wonder if you could make a deflector from "Great Stuff? Take a glass bowl or something with a dome interior shape, coat it with Pam or a similar spray on oil, fill it with Great Stuff, using the bowl as a form and then sand it smooth and spray paint it after it hardens. You would be able to cut it, route sections out as needed, etc. The only thing is, it would be very light.

Why not use plaster instead of Great Stuff? Much more durable. I'd imagine Great stuff might act like a sound absorber at some frequencies due to its inherent lack of density.

Jed

[Paul W]

How about building a 360 degree horn tweeter/mid? The radius of the balls is fairly small so the shape they produce is too shallow to be much of a horn. Looking at section area vs distance from the throat, it seems a surprisingly slow flare from the dome to the outer edge would be appropriate...something shaped like a childs snow bowl/sled could be used as a mule or mold.

Assuming 1/4 wave to center of the horn, and 1/4 wave mouth height, anything below a few hundred Hz is impractically large. Something like the RS52 dome mid is probably the practical limit for a 360 horn. Ring radiators and the BMS annulars also come to mind, as does the Seas Excel removable phase plug.

Look out B&O

[Patrick Bateman]

I hope to get my modular test mule together this weekend using a few drivers I have lying around (RS180S-8, RS225S-8, RS28SA) using all dome shaped diffusers. The RS180 will be sealed and up firing into a 6" dia. dome shape that houses the RS28, which will be upfiring into a 1.5" hardwood ball. I also have a 4-7/8" dome to try as a diffuser the RS180 and a 1.75" round and 1.75" egg shaped hardwood diffusers that I can exchange with the 1.5" ball. The downfiring RS225 will fire into an 8" plastic dome. It will be ported in 39 liters, with 4 dowfiring ports, one at each corner (1-3/8" dia.), tuned to about 35hz.

I'm using plastic domes that were intended as light covers. They are smooth and hard, but only about 1/16" thick. I need to add some mass to them to try and attenuate any possible vibration. I'm looking for ideas as to what I might fill these with to accomplish that. So far, I've thought of some type of modelling clay, rope caulking, silicone caulking, silly putty, etc. It would need to be something that would not dry out and get hard, I would think. The 6" dome houses the RS28 body, or about 1" deep x 3" dia. of it, but the 8: one is entirely empty, so a farily large volume will be needed to completely fill it. I could possibly line it, for example with rope caulk, rather than fill it.

Anyone have any ideas or thoughts on filler material?

For a test mule, I wouldn't even bother with filler material. Unless the spheres are COMPLETELY flimsy, like made out of paper or something. If you're worried, shoot a little foam inside of the sphere; that will make it rigid. That's what I make my horns out of; foam and plastic (or epoxy actually.)

Also, don't even sweat weather the surface is smooth. The "smoothness" is dependent on the frequency. Here's more math to prove it. Let's say you use a styrofoam ball for the tweeter, and there's a series of indentations that are about 1/8" deep along the surface. That would cause frequency response problems at 110400hz, 55200hz, and 27600hz. But anywhere lower than that, it's not even an issue. At lower frequencies, the wavelengths are just too long. Actually that's one of the reasons that eggcrate foam is so useless for absorbing sound; it only works at a narrow band of frequencies that's dictated by the height of the bumps in the foam. It's not what the foam is MADE of that absorbs sound; it's the SHAPE that's important.

(speed of sound / .125") = (13800 inches per second / .125") = 110,400hz

[jacoviii]

hi guys sikped the last part of the post in a hurry

but imo the shape of the deflector shuold near match the shape of the cone
so on a old cone of the same kind make a mold by filling the cone with plaster or the green fibery bondo
it should be rock soild because you do not want any bad reverbs
maybe use a smaller or larger speaker from the same sieres /guop for the form

[dlneubec]

I got my modular test mule together this weekend using a RS180S-8, RS225S-8 and RS28SA. The RS180 will is sealed in about 8L cabinet up top firing upwards. The RS180 was planned to fine up into A 6" dia. dome shape the RS28, which I planned to be up-firing into a 1.5", 1.75” hardwood ball or an 1.75” egg shaped hardwood diffuser. The RS180 and RS28AS was from a prior project and well broken. The down-firing RS225 fires into an 8" plastic dome. It will be ported in 39 liters, with 4 down-firing ports, one at each corner (1-3/8" dia.), tuned to about 35hz.

The plastic domes that were lined with about ¼” or so of plumbers putty to make them heavier and reduce any potential vibration. I used about 1lb. in the 6” and 2lb’s in the 8”.

The crossover used was from a RS180S, RS28AS 2.5way MTM, designed by Roman Bednarek. Here is a link to the crossover.
RJB Audio RS180 MTM

I think it was Paul H. who suggested that the RS180 might be a drop in replacement for the RS225 designs on this board with a simple re-tuning of the bass and cabinet. I had the thought that if that were true, perhaps the RS225 could be dropped in place of an RS180, as a .5way woofer. So that’s what I did here. No change was made to the crossover.

I finished the construction and had it ready to try out by Saturday night.

So how did it sound? Well, in the initial setup not that good. I had the 6” ball probably within ½” of the top of the baffle and the 1-1/4” ball within maybe 1/32” of the RS128. I’m not good at describing how speakers sound, but these seem to be very bright, in both the mid-range and the treble regions, though the sound-stage was very big and the bass was deep and solid.

I then started to mess with different settings. I found that the sound improved as I moved the 6” deflector away from the RS180 and also as I moved the ball away from the tweeter. The Egg shape sounded better than the either ball, IMO. On a whim, I decided to take the 6” deflector/RS28 module and sit it toward the back edge of the RS180, facing forward and up at about a 15º angle. I used a little bit of plumbers putty I had left over to hold it at this position. All of a sudden, the sound was greatly improved. Everything warmed up, the sound seemed fuller and more realistic, yet the sound-stage remained large. I got an idea of how I could mount the tweeter/deflector housing using eye-bolts attached to the upright bolts to allow me to hold it 1” or so above the RS180 and adjust the front to back position of the RS180. You can see this in the photo’s attached. I will be going to the hardware store to pick up some longer eyebolts so I have a greater range of adjustment. I think I prefer the sound most with the RS180 set back a little of center (as far back as I can go now with these eyebolts).

In comparing this to my NaO Mini’s, I’m startled by just how good they sound, given the non-optimized crossover, trying an RS225 as a drop in for a RS180 .5way woofer and given the completely different orientation of the drivers, etc. The omni seems more open and definitely has a bigger sound-stage. My wife thought it sounded more real than the NaO Mini’s. I listened to quite a bit of music today, including some Eva Cassidy, Patty Griffin, Norah Jones, Chicago, Fleetwood Mack, etc. and felt the omni sounded damn good compared with the NaO Mini. I think it is good enough that I may be ready to jump in build a second one.

I think this thing has a lot of potential. I need to pick up some measuring equipment and skills so that I can understand what I am hearing and then see if I can find someone to help with a new crossover design or tweaking of this one. It should be even better when the RS225 gets broken in. I’m going to want to experiment with the height of the RS225 as compared to the 8” dome and also try a Duevel-like diffuser, perhaps for all the drivers.

I tried to show some of the construction in the photo’s. You can see the base, 8” dome and RS225 in one. You can see the RS225 from above with the 4 corner ports. You can see the next box stacked up and see the corner pieces that have hurricane nuts in them that allow the next box to be bolted to the one underneath. You can see the seal I put around each box layer to keep the whole cabinet sealed, yet able to be dismantled. In the RS225 cabinet, I lined the walls with 1” Styrofoam insulation as you can see. This is one very solid box.

[mmoeller]

Great work. Your making such fast progress. I'm totally impressed. I would be down at least a chunk of a finger.

I wouldn't give up on the horn concepts. It seems to me that standing the tweeter section and ball on end like that would really defeat the purpose. The omni-directional effect from the tweeter is completely gone, and the Mid no longer has a predictable horn above it. The mid will sound omni-directional for low frequencies, but the idea was to direct some of the higher frequencies it produces. I would tend to think that moving the deflectors a bit further away might change the brightness of the initial trial. There is some gain and frequency effects from the balls.

[dlneubec]

Great work. Your making such fast progress. I'm totally impressed. I would be down at least a chunk of a finger.

I wouldn't give up on the horn concepts. It seems to me that standing the tweeter section and ball on end like that would really defeat the purpose. The omni-directional effect from the tweeter is completely gone, and the Mid no longer has a predictable horn above it. The mid will sound omni-directional for low frequencies, but the idea was to direct some of the higher frequencies it produces. I would tend to think that moving the deflectors a bit further away might change the brightness of the initial trial. There is some gain and frequency effects from the balls.

I probably gave the impression I'm giving up on the horn/diffuser concepts. I'm not. I've simply adapted to something that is working for now. It's clear to me that their is some horn effect and this is causing a significant change in the frequency response, maybe even some SPL gain. I also suspect that the Duevel-like diffursers will be better than the dome shaped ones and I do plan to have some of those made to try out also.

It is also clear to me that to evaluate it further, I will most likely need to take measurements, something I have neither the equipment, skill, or knowlege to do at the moment. I'm just now starting to research what I need to do to be able to take the appropriate measurements. Until I can, about all I can do is tune by ear with experimental positions and my ear tells me that the horns will not work with the current crossover, but the vertically placed tweeter does. One other setup that was pretty good was with the 6" diffuser directly above the RS180, as high as I could get it and then with the eggshaped diifuser above and slightly behind the tweeter, so that the point of the egg was about even with the back edge of the tweeter dome. The sound was pretty good, but still a little bright and forward.

BTW, they seem very omnidirectional even with this arrangement. As others have said, the Pluto does not have a diffuser and interestingly enough, the driver to tweeter arrangement that Linkwitz uses is close to what I have set up now.

The driver's I'm using are handy in that they are well known and I had a crossover already designed for them (sort of). But long term I'm nost sure if they are the best solution. I'm concerned about how the sound is effected by having a .5way driver firing at the floor, since it is playing much higher than would probably be desireable in that positino. I'm wondering if a 3-way would be better, where the bottom firing driver handles more of the lows and is perhaps a 10" or 12", perhaps a sealed powered sub. I have the two powered, sealed subs I use with the NaO Mini's that have an active crossover designed to give them flat response to play with. I'm interested in making a new bottom baffle and trying one of those 12's bottom firing too. There are lots of experiments yet to be tried, which is why I wated to do the modular build.

If anyone can head me in the right direction as far as taking measurements, I'm all ears (actually, eye's I guess)!

[dlneubec]

It just occurred to me that it wouldn't be too tough for someone already with measurement capability to test the upfiring RS180 and RS28. I imagine there are a number of folks out there who have these drivers around. All you would have to build is the top selaed box, baffle and the dome diffusers. It would not be too time consuming or costly.

Any volunteers?

Well, I had to ask............... :rf

Hey, I could make one up, (since I will need one for the eventual pair) and ship it to someone who is interested. I don't think it would be too heavy, especially if the drivers did not have to be shipped as well.

[JonP]

It just occurred to me that it wouldn't be too tough for someone already with measurement capability to test the upfiring RS180 and RS28. I imagine there are a number of folks out there who have these drivers around. All you would have to build is the top selaed box, baffle and the dome diffusers. It would not be too time consuming or costly.

Well, wouldn't you know, I have the above drivers kind of sitting around... and even some RS225's!! :T Time to dust off a microphone...

I have been thinking about your results so far.. and yes, there will be very likely effects by the diffusors, acoustic centering changes, very different baffle respose, etc... that will change the FR of the drivers enough that they would HAVE to have crossover work done. It's surprising they sound that good just as is!

On the tweeter in ommni results, I'd guess that's why it sounds better vertical.. there likely are some adjustments to be done from the box based crossover, and it just happens to work better in the vertical position. It probably hasn't gotten a fair shake yet..

I could give you some info for testing gear and software... drop me a mail and let me know what you have already...

[dlneubec]

Well, wouldn't you know, I have the above drivers kind of sitting around... and even some RS225's!! :T Time to dust off a microphone...

I have been thinking about your results so far.. and yes, there will be very likely effects by the diffusors, acoustic centering changes, very different baffle respose, etc... that will change the FR of the drivers enough that they would HAVE to have crossover work done. It's surprising they sound that good just as is!

On the tweeter in ommni results, I'd guess that's why it sounds better vertical.. there likely are some adjustments to be done from the box based crossover, and it just happens to work better in the vertical position. It probably hasn't gotten a fair shake yet..

I could give you some info for testing gear and software... drop me a mail and let me know what you have already...

I agree. I still believe that the original concept will work and that the frequency response can be addressed with proper measurements and a new crossover design.

I also assumed that the tweeter sounded better vertically because that's how it was designed for in the crossover I'm using.

I played with the orientation and height of the diffusers a little more last night after getting some longer bolts. With the 6" diffuser about an inch higher, the midrange is not as bad. Using the egg-shaped diffuser on the tweeter, setting it about 1/2" above the tweeter and back, so that the point of the egg is close to the rear edge of dome, I get a decent sounding response, not as good a the vertical orientation, but better than with no diffuser at all, in which cas the highs are really attenuated. I have a cd with a chromatic scale on it and some warble tones. You can definitely hear some peaks across the scale. You can hear some peaks across the scale with the vertical orientation also, but they seem less severe.

I sent you an email regarding the measurement gear help, etc. Let me know if you do not get it. Thanks for the offer to help.

[garretwp]

Will you have any more photos of this project? I would really like to see how things are coming along!

- Garrett

[dlneubec]

Will you have any more photos of this project? I would really like to see how things are coming along!

- Garrett

Did you see these?

previously posted pics

I can post a few more, but they are of the construction process, since there is little new since last Sunday as far as finished items. I have played with some other tweeter/diffuser orientations that I could take some pics of and post, but have not done so yet.

Incidentally, I added a 2ohm resistor to attenuate the tweeter in the existing 2.5way crossover and they sound much smoother. I will have to try the various diffuser settings again to see how that attenuation effects the brighter sound I heard with the diffusers over the RS180 and tweeter.

I'm currently trying to pick up what is needed to do measurements, which will be a critical step in moving the design much further along. It is the only way to get some kind of handle on what the deflectors do. All I can do now is tune by ear.

Ok, here's a couple more, but they are nothing new.

[dlneubec]

My measurement gear should all be here today (I hope). :banana: :banana: :banana:

If I can figure out how to use it, maybe I'll get some measurements over the weekend to help us understand what effect various diffuser locations, heights, sizes, shapes, etc. have on frequency response.

In my test build, I ran speaker wires for the top firing RS180 and the RS28 through one of the 4 bottom ports to the crossover, which is outside so it can be messed with or changed out until it is finalized. However, I had to drill some holes though the bottom of the RS180 sealed cabinet to get the wires through for it and the RS28 and then another hole through the top baffle to get the wire from the sealed box to the RS28. I caulked those holes and used solderless disconnects on the drivers, so I could disconnect them when I wanted to dissassemble the cabinets.

I relized that I could accomplish this flexibility better with two sets of binding posts between the bottom ported section and the top sealed section, one for the RS180 wires and one for the RS28. Then I will add a 3rd set of binding posts on the back of the top seciton to the outside and attached wires from the RS28 to these posts. In the future, when the final crossover is dropped into the box, it will just be a matter of connecting the wires to it. I can also take the cabinet apart easily, by disconnecting the wires at the binding posts between the sealed and ported cabinet and no open holes will need to exist or be caulked. I hope to install these over the weekend and take some pics of how that is done.

It looks like a long weekend.

[dlneubec]

Well, I spent most of a day trying to get my measurement equipment tested and up and running. I figured out, with the help of John K (Music and Design) that the problem is that my audio card in my HP laptop only has two jacks, one for speaker out and one for mic and I need a separate line in. I ordered an M-Audio Transit this morning. I was also going to use the output from my audio card for testing, un-amplified, but will now pull in the old receiver I have in the garage to add amplification. Hopefully, with these changes and additions, I will be able to do some measuring next weekend.

I did take apart the modular boxes to add longer legs at the base. I increased the leg length by 1" to give the down-firing RS225 a little more breathing room from the diffuser. It did seem to sound a little more clean and open in the bass and mid bass, but that may be my imagination. :B

While I had it apart I epoxied in the threaded inserts that hold the boxes together, to keep them solid. I increased their size also, from 8x32 thread socket head bolts and inserts to 10x24 socket head bolts and inserts. Finally, I added the binding posts between the sealed and ported boxes for easy disassembly and then one out of the sealed box to the tweeter. See pics below.

I sure hope to get more done next weekend!

[Patrick Bateman]

Well, I spent most of a day trying to get my measurement equipment tested and up and running. I figured out, with the help of John K (Music and Design) that the problem is that my audio card in my HP laptop only has two jacks, one for speaker out and one for mic and I need a separate line in. I ordered an M-Audio Transit this morning. I was also going to use the output from my audio card for testing, unamplified, but will now pull in the old reciever I have in the garage to add amplification. Hopefully, with these changes and additions, I will be able to do some measuring next weekend.

I did take apart the modular boxes to add longer legs at the base. I increased the leg length by 1" to give the donwfinring RS225 a little more breathing room from the diffuser. It did seem to sound a little more clean and open in the bass and mid bass, but that may be my imagination. :B

While I had it apart I epoxied in the threaded inserts that hold the boxes together, to keep them solid. I increased their size also, from 8x32 thead socket head bolts and inserts to 10x24 socket head bolts and inserts. Finally, I added the binding posts between the sealed and ported boxes for easy disassembly and then one out of the sealed box to the tweeter. See pics below.

I sure hope to get more done next weekend!

What software are you using to measure?

If I were you, I would do gated measurements outside, from a minimum distance of 2 meters. If you do them any closer, you're going to get a funky response due integration issues.

:: PB ::

[dlneubec]

What software are you using to measure?

If I were you, I would do gated measurements outside, from a minimum distance of 2 meters. If you do them any closer, you're going to get a funky response due integration issues.

:: PB ::

I will be using Sound Easy.

[dlneubec]

I finally have my measurement hardware and software up and running (at least it seems so at this time). It has been a major struggle. At first I planned to use my laptop soundcard, but found out after buying a UB802 (mic preamp w/phantom PS), a Behringer ECM8000 mic, mic stand, mic cable, and various other connecting cables, resistors, etc. to make up a measurement jig, that my soundcard needed a "Line In" as well as a Line Out, and all I had was a Mic In and Line out.

So, I then buy the M-Audio Transit, an external USB Audio device, which seems to be regularly recommended for laptop users. I got the Transit in, and it was a nightmare trying to get it to install on Win XP Professional, though it is supposed to be fully compatible. I finally think I got it set up and get some really inconsistent results. It looks like there is a timing issue and so I mess with the Latency setting on the Transit until I get a pretty consistent result. Combined with this is the learning curve for using SoundEasy. It looks to be a very powerful program and came highly recommended from several I've worked with in the DIY community, but it is certainly not easy or intuitive to use.

At any rate, I know think I can get some impedance and Freq. Resp. readings, but still have a ton of studying to do to understand the best way to set up the measurement for such a different type of speaker. It is pretty straight forward for you typical box, but the omnidirectional is a whole different animal. I did take some spl readings on the entire system, using the crossover that I have. I'm not sure if I did it all correct, but it was very easy to see how changes to position or orientation of the up-firing woofer diffuser and especially the tweeter diffuser affected the response. For these, I measured at 1m at about 34" high, which is right between the upfiring woofer and tweeter. Attached are a couple graphs that I got with some different woofer and tweeter diffuser settings. I can't even tell you where I had it setup for these, I just tried to do a few screen copies to show how radically things changed with diffuser changes.

I was warned that there might be some phase problems with the existing crossover, due to switching in the RS225 for the RS180. I believe that is what the 5-6db dip at around 450hz is. That was consistent regardless of how I changed the diffuser settings, however, when I unhooked the RS225 and ran one with the upfiring RS180 only, the dip disappeared completely.

I think my next move should be to measure the response one driver at a time, without the 2.5way crossover, starting with the tweeter and working on the diffuser type, position, etc. in order to get the flattest or most easily controlled response I can, then do the same with the top firing woofer, and finally, the bottom firing one. With these settings in hand try and then see how it looks with the tweeter and the top firing woofer together and whether further adjustments can improve the overall response and then finally with all three to see how that looks. If I get anywhere with those steps, I guess it will be on to whatever the next step needs to be.

I’ll post more after I get some individual driver readings, perhaps posting results from multiple positions of the driver/diffuser combo.

[dlneubec]

I did a little testing on the tweeter last night and made these general observations. It tested with the tweeter front firing with no diffuser and up firing with no diffuser and up firing with diffusers, including a 1.5"dia ball, 1.75"dia ball and a 1.75"dia egg shape with the narrow point at the tweeter. I tested the diffusers at approximately 1/32", 1/4", 1/2" and 3/4" above the mesh screen on the RS28, centered on the dome. I measured at 1m out, 34" high, which is be right between the up-firing woofer and the tweeter, using a 3.95ms time window. The drivers were measured without the existing crossover.

The highest SPL and flattest response was from the tweeter in a front firing arrangement setting above the woofer. Height did not seem to matter much.
The Egg shaped diffuser was generally the lowest SPL and most ragged response at all heights. I believe it can probably be removed from further testing.
Generally, the SPL increased and the response flattened out as the diffuser got closer to the tweeter. Generally, those at 1/32" were the best for all sizes.
Generally, the larger the diffuser surface, closer to the tweeter seems to be better. The 1.5" ball was better than the Egg, which was narrower. The 1.75" ball was better than the 1.5" ball.
The Egg shaped diffuser, at 1/32" from the tweeter had an even greater high end roll off than no diffuser on the up firing tweeter at all.

What I have yet to try is offsetting the diffusers, say to the rear of tweeter center.

The next things I intend to try with the tweeter are larger diameter diffusers than 1.75". I also want to try cutting a flat spot the same size as the phase shield on the RS28 on the 1.75" ball and actually setting it so it touches flat spot on the top of the mesh screen on the RS28.

I hope to have the Duevel-like diffusers in the next couple weeks and want to test those as well.

Attached are a few graphs as examples. They were set to add 20db to the result.

The first is the front firing tweeter, set 4.75" high above the top firing woofer baffle and 6" back from the front edge of the box.
The second is the best of the rest, the 1.75"ball, centered and 1/32" above the tweeter mesh.
The next is the egg shaped diffuser 3/4" up.
The last is the 1.5" ball 3/4" up.
The last is the up-firing tweeter with not diffuser.

I'm a noob at this, so there could easily be something wrong in my setup. These are all plotted with phase also. Take a look at how the phase changes when the tweeter is front firing versus up-firing.

Any thoughts?

[dlneubec]

In case anyone's interested, here are a series of results for the 1.75" diffuser at 1/32", 1/4", 1/2" and 3/4" heights, FWIW.

Is anyone still interested or are you all sick of this thread? :Z

[Dennis H]

Thanks for keeping us posted Dan. Interesting stuff.

[edjosh23]

Dan,

I don't frequent this forum often, but I have been following this thread. I find your research quite interesting. I was wondering if a tweeter with a rising frequency responce would be better. Your responce curve of the tweeter facing forward was good, with a slight dip at around 4khz, which is not present when the tweeter is omnidirectional with a deflector. I decided i would look for a some tweeters with rising responce curves.... couldn't really find anything useful, but perhaps you may know where to look for that.

Josh

[Paul W]

Dan,
Still interested and following your progress. I'm curious to know what you find with a large radius very close to the dome.
Paul

[dlneubec]

Dan,

I don't frequent this forum often, but I have been following this thread. I find your research quite interesting. I was wondering if a tweeter with a rising frequency responce would be better. Your responce curve of the tweeter facing forward was good, with a slight dip at around 4khz, which is not present when the tweeter is omnidirectional with a deflector. I decided i would look for a some tweeters with rising responce curves.... couldn't really find anything useful, but perhaps you may know where to look for that.

Josh

I have learned a few things, as most newbies do. John K. of Music and Design has told me that the deep notch (notches?) in the diffuser plots are due to resonance between the dome and the diffuser, which is one of the reasons why the front firing tweeter is much flatter. Maybe that's why the larger dia. diffusers, placed closer to the dome seem better. Also, as you go up in frequency, the tweeter output get more directional. All the directional ouput would normally be pointed at the user and shows up in the front firing plot. On the diffuser plots, however, the diffuser is forcing that sound out 360º, so there has to be a roll off at the top end as the output gets more and more directional and the sound is forced to spread out. Perhaps a tweeter with a rising output would help some, but I have not researched that.

[mmoeller]

Very cool stuff. I'm interested to see measurements of off axis performance, between the front facing tweet an the diffusers w/ upward firing tweeter. One might assume that the objective of the diffusers would be to create a uniform response for the off axis performance. The front firing tweeter would no doubt have a better on axis responce, but the omni directional should have a more uniform off axis. The high frequency roll-off might be correctable, albeit beyond my knowledge as to how. I could guess the edjosh23 might have a good solution. Or you might be able to correct with the crossover.

The second set of graphs tells alot about the results. You can see the higher you set the diffuser the lower in frequency the anomalies show up. The 3/4" graph show the dip at about 9kHz. Wavelength at 9kHz is 1.5", hmm 1/2 wavelength is 3/4". Funny how math works out sometimes. In my other posts I spoke alot about 1/4 wavelengths, and 3/4 wavelengths. Seems that the experience of Mr. Bateman has show that the 1/2 wavelength was key.

It is also interesting that without the diffusers there is a much more pronounced diffraction effect. If I'm reading the graphs correctly.

Still paying attention.

[mmoeller]

Well looks like we were typing at the same time.

[dlneubec]

Dan,
Still interested and following your progress. I'm curious to know what you find with a large radius very close to the dome.
Paul

Ok, last night I tested with 2", 2.25" and 2.5" dia. diffusers. Since the 1.75" dia diffuser was better than the egg or the 1.5" dia diffuser, I assumed going even larger would offer continued improvment. In reality, there seemed to be very little difference, maybe some slight improvement. I tried these at the same heights as the smaller diffusers, up 1/32", 1/4", 1/2" and 3/4".

However, all along I've been trying the diffusers moved toward the rear of the dome also and thought they sounded better, though I had no measurement capability at that time. So I checked this out with the 2", 2.5" and 2.25" diffusers. I moved them back from center in approx. 1/8" increments. On all three, there was a point where the frequency flattened out from about 7or 8khz to 20khz. The 2.5" one was probably the best and it was as it was set back about 3/8". Attached is a plot of it. It looks to me like that position does a couple things. One, it decreases the effect of resonances between it and the tweeter, pushing the single notch up over 20khz. Two, it increase the spl, by forcing more of the sound forward and is probably radiating in less than 360º as a result. In the plot below, the 2.5" ball was basically touching the mesh grid over the tweeter. It seems much improved over those centered over the tweeter, but still does not match the front firing results.

I still wonder about the cone or convex cone shape as a diffuser. I suspect you will still have to have the natural roll off, but it may better address the resonance issues. I'm going to take a few days off from testing while I do some more reading about testing and also to do some wood glue-ups that will be used to make some Duevel-like convex cone shaped diffusers.

[mmoeller]

... it increase the spl, by forcing more of the sound forward and is probably radiating in less than 360º as a result.

Only one way to find out, measure off axis. :T

[dlneubec]

One other thing I did last night was to start trying testing the up firing RS180. I started with the 6"dia diffuser (holds the up firing tweeter and is directly above and centered on the RS180) at about as close as I could get it, which put the tweeter face up 4.25". I tested also at 4.5" and then at 5". There did not seem to be much difference between the settings. Attached is a plot of the diffuser at the up 5" position.

It looks ok to my eye up to about 1.8khz or so and then lots of irregularities appear. I don't know if these notches are due to resonances between the diffuser and the RS180 or if it is just the natural breakup modes of the RS180 Maybe someone with more experience looking at these and who know what the RS180 FR looks like could give us a better idea.

It does look like if you were to combine this plot with the plot from the tweeter with the rearward diffuser above, and could flatten the response out at about 90db, you would have a pretty flat response from around 300hz to about 7khz and then it drops a couple db out to about 18khz. The down-firing RS225 will bring up the low end also.

[dlneubec]

Only one way to find out, measure off axis. :T

No doubt you are correct. I definitely need to do some off axis comparisons.

After I take a few days off from measuring, (while I read and try to learn more about what I'm doing!) the first thing I will probably do is try some off axis measurments. I would envision comparing the front firing tweeter setup to one with a larger diffuser close and centered over the tweeter and to the one with the 2.5" diffuser set back from from center of the tweeter.

One other thing I tried early on, before I started taking measurements, was to try the tweeter faced on an upward angle somewhere between direct front firing and up firing, with no diffuser. I actually liked the sound pretty well, so it would be interesting to see how that measures both on and off axis.

[mmoeller]

Looks like you would be ok with the 4-5" range for the distance on the woofer. It's about a 5k-6.7k spread for the 1/2 wavelength. The second defined notch in your graph shows this. The first one must be something related to the woofer response. The graphs from partsexpress show that the FR starts to degrade right between 1k and 2k.

[dlneubec]

Looks like you would be ok with the 4-5" range for the distance on the woofer. It's about a 5k-6.7k spread for the 1/2 wavelength. The second defined notch in your graph shows this. The first one must be something related to the woofer response. The graphs from parts express show that the FR starts to degrade right between 1k and 2k.

Just to make it clear, the dimensions I'm giving is to the face of the tweeter baffle. The distance to the diffuser is considerably less. Attached is a couple photo's of the 5" high setup. This also shows the 2.5" dia. diffuser, set back of the center of the tweeter by about 3/8".

[Dennis H]

Dan, do you still have the stock plastic diffusor under the grill of the RS28? I wonder if it wouldn't be better to remove that. Seems like it might interfere with what you're trying to do.

[dlneubec]

HI Dennis,

Yes I do. I hadn't even thought about it. Is it hard to take the RS28 apart? I've never considered it before. It might be interesting to see what difference removing it would make.

[Dennis H]

Dave Ralph's page shows how to take the tweeter apart. It looks pretty easy.

http://www.speakerdesign.net/dayton_rs/tweeter/reference_series_tweeter.html

[Paul W]

Typical dome tweeter response narrows with increasing frequency so, with flat on-axis response, total radiated power is lower at HF.

Directing a relatively narrow HF cone into a broader doughnut shape means relatively low HF power must cover a broader area. Achieving flat frequency response with broad area coverage would require HF boost at the dome. HF boost could yield flat frequency response in a full 360 degree circle and would also provide a lower difference in power response across the tweeter frequency range.

Did I get that right or do I need more coffee?