The Contrappeso Project

[dlneubec]

Edit: Hi all. Many of you were aware of this project and some were not. The reason is that is was buried a few pages into the "RAAL Omni Discussion" thread. I have had a few requests to have it in a separate thread, so I asked a moderator, Kevin P to help me with it. He was great with it and this is the result.

My project ended up being called The Contrappeso's, which is Italian for counterbalance. Below you can read through the thread winding up to the final results. For those of you who followed the original thread, it is still available tp be posted in with regards to RAAL's omni design, which was the inspiration for this one. Below is where that thread was split.

Ok, I'm thinking of testing out this concept, using some inexpensive buyout drivers, or the like.

The midrange is set, from some Vifa buyouts, here:
Midrange
These are 16ohm drivers and 4 in parallel yield around a 92-93db sensitivity, but should be about 6db down from that on the listening axis.

The woofers are still undecided, but are currently two 12" sealed woofers in a 12" tube front and rear firing. These were PE buyouts from about 2 years ago. In a sealed 40 liters, a parallel pair will get down to around an F3 in the mid 30's, f10 around 21hz and will take more power than I'd ever give them and not get close to Xmax (8mm). The drawback is these would have a qtc of around 1.0 in that small volume, with a sensitivity of around 88db/2.83V. These may be changed out for two smaller woofers in the 6"-8" range and perhaps smaller bass bin tubes (diameter or length).

The tweeter will be the biggest challenge. This is probably not full range omni since I would the tube diameter will begin to make it directional to some degree as you go up in frequency. Of course RAAL used a side mounted ribbon, which is probably ideal and may be where I end up. I will start testing with 4" dia. schedule 40 pvc tubes for prototyping and various neo dome tweeters. The domes spherical output, as compared to a ribbons more horizontal cylinder output, could be a problem, but I think it is worth testing. With an 8” long tube, it would be about 1 wavelength to the top or bottom edge of the tubes at around 3.3 khz or so. Perhaps if I crossover above that, I will have better results. I have a couple Dayton ND16's I can try, a couple buyout 3/4" tweeters, a pair of NLA Vifa D26NC55's and if any of them look promising, I will probably pick up a couple Vifa OX20SC00-04's to try.

If the domes may prove problematic, I may be forced to try and find a narrow ribbon, which could inflate the budget a significant amount. If anyone has any suggestions for ribbons, I’d love to hear them. It has to fit inside the 4” tube, in between the 3-1/2” drivers and probably fire into a narrow slot, since the tube diameter is a restriction. I would assume that it’s face plate, at the very least, would have to me removed.

Below is a mock up of what it might look like with an aluminum finish on the tubes. Crossing the woofers low, around 200-250hz, placing them near the floor and pushing the woofers off to the side and well to the rear of the MT allows both the front and back firing woofers to be closer to at least 2 of the room boundaries while keeping the omni midrange and tweeter further from the wall, which omni's usually prefer. There should be no bsc needed in this design.

The build for this should be pretty easy since the tubes form the enclosures, but the design could be a serious challenge and may simply not work, especially the tweeter configuration. It’s a very contemporary look that many would not appreciate.

If this does work out, it may open the door to a project with more serious drivers.

What do you guys think?

[CraigJ]

[B][I]

Ok, I'm thinking of testing out this concept...
What do you guys think?

Dan,

I can't wait to hear what you build as I've been excited about this project since it was first mentioned. I had thought about using the Fountek FR88-EX after @onMarsh and Augerpro mentioned them and read that they tested well for John Krutke. For woofers, the plan was using a pair of RSS315HFs in a sonosub like enclosure. Based on Aleksandar's comment "it seemed natural to me to use two cans, stick them next to one another, and slam a ribbon in the middle, but...and there's a big BUT in there, fighting housewives is also a futile and doomed effort. Invariably, they are the nimber one ribbon destroyer" Not too worried about my wife's dusting, I had thought about using a partially covered (to get a 1/8" narrow throat) Neo3 PDRW between the cans, facing sideways or forwards, depending upon testing.

4" x .125" x 24" aluminum tubing can be purchased at https://www.speedymetals.com/pc-4629...-aluminum.aspx for about $14.47 plus shipping. Optional LED lights can be purchased at Home Depot.

CraigJ

[Saurav]

Wow, but, I'm confused Which is the front view and which is the side view? The midranges... there's 4 of them, to in each tube, firing up and down? Where is the tweeter? I saw the pictures of the RAAL design that looks like a floor-standing lamp, but even there I wasn't clear on exactly which way all the drivers were pointing.

[savage25xtreme]

the left view is the front view..... errrr ya.

The tweeter would be between the 2 midrange tubes. I think its a slick idea. Interested in hearing your thoughts about the dome tweeter between the tubes, a mini waveguide from a point source? in theory anyways?

[CraigJ]

From Aleksandar Radisavljevic's post: "a side firing ribbon, into a very narrow slot between the two cans was the solution. After some tweaking and shaping of the ribbon slot, it turned out to have a surprisingly good behaving response. The narrow slot was the key to dispersion and flat response. You see, the dispersion of of any horn improves tremendously with very narrow throats if you don't change the lenght. Also, the dispersion of a rounded edged horns is much better too. In this case, we have a horn that is shallow, has a very narrow throat ( <1/8") and the highs are nicely creeping along it's rounded sides all the way to the side where it meets the backwave of the same magnitude and phase, adding 6 dB, just where the response in highs starts to loose it, and voila! Omni!" Easy! :rofl:

Not sure where Dan has located his tweeter.

[Saurav]

OK, that's a little bit clearer. Seems like a B&G Neo3 could be made to work with something like that? Cut slits in the PVC pipe and then slide it onto either side of the tweeter. I tried something similar on a much smaller scale:



But I was trying to fix the dipole null on the Neo3, while this would be more of an attempt to put a dipole horn on the tweeter.

[dlneubec]

Craig,

I don't think what Rudolf illustrates is what RAAL has done. Rudolph's looks like a dipole approach to me, which it most definitely is not, as Aleksandar says in his post.

My understanding is that the ribbon resides in one can and fires into the side of the other can, about 1/8" away. There may be other "shaping" going on at the tweeter/can interface. It seems to me that only with a side firing driver would you get the same response front and rear, unless you went to two drivers in a bipole arrangement.

[dlneubec]

OK, that's a little bit clearer. Seems like a B&G Neo3 could be made to work with something like that? Cut slits in the PVC pipe and then slide it onto either side of the tweeter. I tried something similar on a much smaller scale:



But I was trying to fix the dipole null on the Neo3, while this would be more of an attempt to put a dipole horn on the tweeter.

This is clearly dipole, with a figure 8 dipole radiation pattern, not omnidirectional.

I don't think this is what RAAL is doing. What I have in mind is not dipole at all. I suspect mixing an omni pattern with a dipole would have serious phase cancellation issues, since the out of phase back wave of the dipole would cancel in phase rearward omni output from the mids. Not a good mix at all, I don't think. A bipole would probably work. Perhaps RAAL's approach, if I understand it, could be characterized as a marriage of omni and a single driver bipole?

[dlneubec]

I think the neo3 is going to be too side to work in this configuration. I believe it will need to be a very narrow slot ribbon. I would presume that some significant low end boost will occur as a result of the waveguide effect, so perhaps a small ribbon can be pushed to a lower crossover point as a result.

If you cut a slot in the side of a can wide enough to side fire a neo3, it will push the planar surface too far away from the "horn" surface on the opposite can and you will end up with a cavity in there. As Aleksandar said and as I found in testing upfiring tweeter fo some of my other omni projects, the key to a good response seems to be a very narrow space. I suspect it has something to do with keeping that distance shorter than the wavelengths it has to reflect.

I worked with another guy who wanted to put a dipole mid above an upfiring, omni woofer and their were serious out of phase cancellations. I believe he eventually put a shelf between the woofer and mid and covered the back of the mid with cloth and some stuffing, I think, to attenuate the back wave. I just don't think we can mix dipole and omni due to the out of phase nature of the dipole back wave.

Dan,

I can't wait to hear what you build as I've been excited about this project since it was first mentioned. I had thought about using the Fountek FR88-EX after JonMarsh and Augerpro mentioned them and read that they tested well for John Krutke. For woofers, the plan was using a pair of RSS315HFs in a sonosub like enclosure. Based on Aleksandar's comment "it seemed natural to me to use two cans, stick them next to one another, and slam a ribbon in the middle, but...and there's a big BUT in there, fighting housewives is also a futile and doomed effort. Invariably, they are the nimber one ribbon destroyer" Not too worried about my wife's dusting, I had thought about using a partially covered (to get a 1/8" narrow throat) Neo3 PDRW between the cans, facing sideways or forwards, depending upon testing.

4" x .125" x 24" aluminum tubing can be purchased at https://www.speedymetals.com/pc-4629...-aluminum.aspx for about $14.47 plus shipping. Optional LED lights can be purchased at Home Depot.

CraigJ

[CraigJ]

Craig,

I don't think what Rudolf illustrates is what RAAL has done. Rudolph's looks like a dipole approach to me, which it most definitely is not, as Aleksandar says in his post.

Dan,

I see what you mean and I'm sorry for the confusion. When I looked at Rudolf's illustration, it looked to me as though the tweeter ribbon was in the tube and the red was how the tubes were connected. Yes, as I wrote, not illustrated, "side firing ribbon" with a <1/8" throat.

What if instead of putting the tweeter inside the tube, you put a Neo3 (rear cup attached) with a diy 1/8" throat between the tubes facing towards the opposing speaker? Surely not up to your beautiful build standards, but this is a prototype.

I just saw that your main build thread is at PE.

[Saurav]

I suspect mixing an omni pattern with a dipole would have serious phase cancellation issues

Quite likely, yes.

If you cut a slot in the side of a can wide enough to side fire a neo3

I can envision a couple of ways around that. You could cut a narrower slot, and fill the space with something to keep reflections down (first picture). You could also cut a wider slot and add a 'chord' (2nd), or just move the tubes closer together (3rd). Neither of these might work from an aesthetics POV though.

[dlneubec]

Dan,

I see what you mean and I'm sorry for the confusion. When I looked at Rudolf's illustration, it looked to me as though the tweeter ribbon was in the tube and the red was how the tubes were connected. Yes, as I wrote, not illustrated, "side firing ribbon" with a <1/8" throat.

What if instead of putting the tweeter inside the tube, you put a Neo3 (rear cup attached) with a diy 1/8" throat between the tubes facing towards the opposing speaker? Surely not up to your beautiful build standards, but this is a prototype.

I just saw that your main build thread is at PE.

Ok, we are on the same page. I've still not decided if I want to jump into this project or the other one I posted over at PE, but the omni is more intriguing and most likely more challenging.

[dlneubec]

Quite likely, yes.
I can envision a couple of ways around that. You could cut a narrower slot, and fill the space with something to keep reflections down (first picture). You could also cut a wider slot and add a 'chord' (2nd), or just move the tubes closer together (3rd). Neither of these might work from an aesthetics POV though.

To get the horn to work worth a darn, I'm thinking you want to maintain the horn shape and flare up to as close to the narrowest portion of the horn as possible. But, I have little experience with horns, so one of your solutions might do fine.

[Saurav]

That's a good point.

To be honest, I was thinking about the 'Rudolf' picture when I said I could see how a Neo3 would work, and didn't think about the omni vs. dipole difference. I'm not sure a Neo3 is the best starting point for what you're trying to do.

[CraigJ]

Saurav,

Why does the Neo3 have to be on the inside? What if you recessed it's backside into one tube, built an 1/8" thick x <1/8" wide "mouth" on the front side of the Neo3, and mounted this between the tubes?

Aleksandar did say "Not to be bragging about it, but that is the first horn of it's kind ever made and it does have significant advantages over anything that you can stick a ribbon into." Gotta be something different than we've tried before.

Dan, I really like the stand you've designed.

[Saurav]

That might work. If I'm visualizing it right, that kinda sounds like the middle drawing in my earlier post, except the Neo3 isn't on the inside. I don't know how that tweeter will behave with a 'wall' behind it.

[Rudolf]

To be honest, I was thinking about the 'Rudolf' picture when I said I could see how a Neo3 would work, ...

Actually I have no idea how that ribbon is working between those cylinders. I only saw a photograph with something ribbon-like lurking from deep inside the converging walls. My picture was a proposal to ask John Marsh if that was the actual alignment.

[dsrviola]

Another Omni

Holistic Audio Arts:

http://www.stereomojo.com/RMAF%202010/RockyMountainAudioFest2010pt4.htm

and for a better picture of those illuminators, scroll past 3/4 of the way down:

http://www.positive-feedback.com/Issue51/rmaf10.htm

Nothing original here except for the DSP required before a DAC (at least for the smaller model.) More info. here:

http://store.virtueaudio.com/product-p/ha-h1speaker.htm

[JonP]

Wow... nice pair of Illuminator posteriors... :drool: Interesting how small the tweeter reflector is, I'd expect bigger.

Also interesting to see what Dueval is doing a few pics down. Looks like a compression tweeter firing thru the horn/lens assembly, and a larger (8"?) driver below.

[CraigJ]

Dan,

What do you think of placing a Fountek neoCD1.0 with the faceplate removed inside the tube (55mm width)? My mock up show about 1/4" distance between ribbon and inside of tube. May have to remove the mounting screw holes to reduce the width of the driver...

Craig

[dlneubec]

Craig,

Let me take a look at it. Do you know what it looks like with the face plate off? I've never seen a ribbon with the face plate off, so I don't know how they are constructed, typically. Is the ribbon element pretty much flush with the face of the main body, or does it set behind it or protrude out from it?

Obviously, the narrowor the body, the closer it can be positioned relative to the curved tube wall. Below is a top view of two 1/4" wall tubes, 4" I.D. with the body of a tweeter that is 2-1/8" wide. The two lines extending from the body of the tweeter represent the ribbon width. The distance to the inside of the wall at the edge of the ribbon is about 1/4" and the distance to the outside another 1/4". So, the body of the tweeter sits at least 1/2" behind the face of the cylinder. What is done in this space might be critical. I would think a narrower body would be much better, but there may be nothing available to DIYers that will work well. This is why ribbons might be tough, unless you the response will work with a deep narrow slot out to the face of the tube. You could certainly route a small roundover on the slot edges, but I don't know if that would help. On the other hand, a wavelength at 20khz is about .68" long, less than the depth of the slot when using the neoCD1.0 shown below, so maybe this would not be a problem.



Below is the same drawing with a Dayton ND20 shown opposite the ribbon tweeter. As you can see, do to smaller body, the small neo dome can be placed much closer to the face of the tube. This is why I think dome tweeters should to be explored. Something like the Vifa OX20SC00-04 might be a really good dome option, with its tiny circular footprint and no flange. All the Dayton ND's have a small waveguide/flange they are mounted in, though it might be possible to remove them. The ND16 is a little smaller yet, with it's 5/8" dome.



On the other hand, the ND series flange is such that it might allow the outside of the tweeter WG to be place so that it matches the top and bottom of the tweeter cutout in the tube and thus be a little proud of the tube as it curves back toward the side. This proud edge might then be sanded flush with the tube.

[dlneubec]

For the ribbon, one might actually butt the tubes together, roundover the slot, fill the void area with cotton or something and glue a triangular piece to side of the opposing tube to deflect the sound out the front and back. Something like this (similar to Saurav's sketch)

[dlneubec]

Here's another thing I've been thinking about. For the purposes of flexibility of room placment, I could make the MT head so that it swivels up to 90º. That way the bass bin could be placed to that its length is perpendicular to the front wall (i.e. when close to the side wall) or the MT head turned 90º so the bass bin tube runs parallel to the front wall (i.e. when not sidewalls are nearby). This would probably complicate getting the wiring into the MT section a bit, however.

[Saurav]

On the other hand, a wavelength at 20khz is about .68" long, less than the depth of the slot when using the neoCD1.0 shown below, so maybe this would not be a problem.

I'm not too confident about that. My only experience with something similar is when I had a screw-on CD tweeter and a bolt-on horn (or maybe it was the other way round), and I tried an adapter which didn't quite match up, so there was a small gap between the tweeter and the WG throat. I'd say it was smaller than 1/4", I could just feel it with my finger. That caused a very deep null somewhere below 20kHz, I'm thinking around 12-14kHz, but I'm going from memory here. Filling that space in with polyfill reduced the depth of the notch, which only helped to confirm that it was that gap that was causing the problem.

I think you'll have to fill it with something, and you may even need to fill it with something solid, so you essentially cover up the ribbon's radiating area except for the small gap / slit down the middle.

Knowing you, you'll try all of these variants before you build the final version It'll be interesting to see what you find. After my experiments with dipoles, I've been curious about trying omnis, but I don't trust my ability to measure them.

[dlneubec]

I agree with you. I suspect getting a DIY version for a ribbon is going to be a bi*ch unless you can find the perfect ribbon or have access to someone like RAAL who can design a tweeter specifically for this application.

[JonP]

Interesting set of design possibilities... I guess doing some experimenting is probably the only way to find out the "better track".

I imagine you have to have thought about it Dan, but what do you think of a short "line array" of dome tweeters, like 4-6, down the side of the cylinder? Or, extending the cylinders to go to 8 tweeters or so? If small and close enough spaced, I wonder if they could blend well, without it having to get unreasonably long...

[Saurav]

Heh, I was wondering about a line of tweeters too. Ideally you'd end up with a cylindrical wavefront to match the 'WG' (and the accompanying narrower vertical directivity), without introducing too much comb filtering that will affect the horizontal response.

[dlneubec]

No, the line of domes hadn't even occured to me. Great idea, definitely worth exploring!

I wish I had some experience with line arrays to know what the drawback of a small line of drivers would be. I guess I can model it in Edge. The length of the tube, however, probably has to be limited in order maintain CTC spacing and the MTM format.

As you said, experimenting with this probably the only way to know. I've been doing a bit of thinking about how I might be able to mount/implement both neo dome and ribbon tweeters and have some additional ideas. A line of domes will probably require a different mounting solution.

[CraigJ]

Dan,

I don't know how much of the build you'd like us to contribute, but I could easily run over and pick up a Fountek driver at Madisound on Monday and put it inside the 4" pvc I have. Heck, if you instruct me on how to measure an omni, I could also assist doing that as well.

Regarding domes, I think PE is introducing a new 1/2" dome "soon" to go with their CBT kit if that is a smaller option.

Happy Holidays,

Cj
p.s. @ madisound, I could also pick up eight NS3s, or eight little Founteks.......stop me

[dlneubec]

Thanks for the offer to help, Craig! I don't mind if others want to jump in and parallel on the effort for this one. If you want to, feel free to copy the look I have created, if you like it. I can send you a dimensioned cad file as well. We an post any results, construction photo's, etc. here or at PE.

One thought that occured to me regarding adapting ribbons is instead of removing the face plate completely from the tweeter, use it, but cut it down in width, just enough to leave the opening to the ribbon and any chamfer or roundover intact. Cut the rest of the flange away and you end up with a section that sits out in front of the tweeter body, but much narrower than the body in width. You'd use this protrusion as the side space filler and match the opening you create in the tube to match this. Perhaps adding a triangular filler piece to the opposing tube to act as a divider between the front and rear and to keep the space narrow and gradually increasing. Perhaps better yet, make your own custom narrow face plate to extend to the edge of the tube, something like this:



In my experience, I had great results measuring an omni just like any other speaker and designing for essentially flat on axis. Since the power response is so smooth and equal around the room. what you see on axis is what is heading off axis and the response should theoretically be identical at any poiint around the speaker. I don't think that will be true of this speaker, in that I suspect the top octave will roll off as you get 90º off axis due to the waveguide and shortening of wavelengths, so off axis measurements will be a must to understand what is going on. Of course, you should do off axis measuerments for any design, so there is nothing new there.

I can tell you that after hearing the CBT's at MWAF this year, I'm very interested in seeing what PE comes up with for a kit. I thought they sounded really great. Those tiny full range drivers might be perfect for this project as well. Let me know if you see anything on them.

For now, I have the ND16, ND20, ND28, D26nc55 domes and even the neo SB29 dimple dome I can play with. I'm in no hurry. This will project will take a lot of research and testing to get right, so might as well go slow.

Oh, while I'm thinking of it, the idea I had to adapt a dome to the tube is to use a smaller pvc tube as the mount. Pick a smaller pcv tube the correct ID to just fit around the dome and it's surround. Now drill a hole though the side of the big tube the size of the OD of the small pcv tube. Cut the pcv tube to a depth that allows the tweeter body to just clear the inside of the larger tube surface. Glue the small pvc tube into the side and sand off the extra tube that protrudes from the surface as it curves toward the side. Then provide a chamfer or roundover as needed on the remaining opening. This allows for a flat surface on the inside for the tweeter to mount to, allows you to vary the depth of the tweeter mount to fit the body diameter and allows you to contour the outside of the opening to fit the outside curve on the big tube's wall. Do you follow me and does that make sense? I haven't done a drawing of this idea yet, but will.

[dlneubec]

Here is what I was talking about in the last paragraph above, for mounting a dome tweeter by using a smaller diameter pvc tube glued into the side of the larger tube.

I suppose a stack of them could be done this way also. The small pvc tube wall, for 1-1/4" pvc is about .145" so that would add about .290" extra between the tweeter bodies, a little under 5/16" (.3125"). If the tweeters are 1-1/4" dia., like the ND16, for example, that would make the CTC spacing on the tweeters a little over 1.5", about 1 wavelength at 9khz.

[Saurav]

Dan, I thought you might find this interesting:

http://www.diyaudio.com/forums/multi-way/180301-lq-osd-acoustic-alternative.html

I can't say I've absorbed even a small fraction of what he's talking about. But it seems related to omni. I think.

[RAAL]

Hi, everyone!

dlneubec, I'm flattered!

We still haven't finished the design, but it will be very close to how you solved it. Single tilted pole or plate holding the top cans, no moving parts.

The main thing is to keep those cans as thin as physically possible to allow for good dispersion. We are cutting off the speaker flange to make it fit into 80 mm tube as I didn't want to go any bigger than that and I think 4" tube might be over the max. If possible, don't use plastic tubes.
Even more important in that case is to use a narrow ribbon tweeter and the solution with a ribbon and a little fork waveguide is very interesting.
As for the domes, I'm not so sure, but the trials is the mother of engineering, so give it a go. The problem there is that we see tweeter active area from 90 degrees. If any cancellation or comb filtering happens at 90 degrees, I wouldn't use it. If it's OK, try to take out the ferrofluid out of the gap before making trials. It ruins the detail and small tweeters are usually full of it.
In any case, you would want to have equal volume in both cans and little dome tweeters will fill that bill without having to worry about volume inbalance. A retail true ribbon solution is going to eat-up a big chunk of internal volume and that's the only good reason not to go the true ribbon way.

I guess that the only tweeter in retail, which is still affordable and has the smallest construction ( after it's waveguide is taken off), is this:

https://www.madisound.com/store/manuals/fostexdrivers/ft7rprev.pdf

It even has the phase plug that you can use, as well as it's waveguide after some cutting and trimming.

It will also allow for some attenuation after, as you don't want to run into integration problems.

Speaking of integration and crossover, you will see that above certain frequency, the response on the side will get dips and peaks due to time missalignment of the mids. this is why you need to keep the diameters small, if you want true omni. If it will be compact enough, you will end up having the first dip at 90 deg. around 3k, and you will make the crossover right there, to keep this effect at a minimum.

That's all the advice I can think of now. I'll check up on the progress when I can!
Enjoy the build!
Cheers,

[dlneubec]

Hi Aleksandar,

Thanks so much for the advice. I have found that I can get aluminum pipe for a fairly reasonable cost, especialy at thinner wall thicknesses. For example a 3.5" outside diameter tube (14 guage, 6063 aluminum) at 48" length and .083" thickness is only $22.00. The same is available in 4" diameter for about the same price. 14 guage is a little over 2mm thick, so I don't know if that is strong enough or not. From the same source I can get the same aluminum in 8" diameter, 14 guage, 48" long for about $80. That would be enough to do 2 bass bins, but for the drivers I've been looking at, that is really short on volume. I may have to go to somthing like 12" schedule 40 pvc pipe to keep the price reasonable.

I'm wondering if the midrange tube diameter is that critical. Yes, the smaller diameter tube allows you to get the side by side mids closer together, but that is assuming you locate the mids in the center of the tube. If you used a larger tube, you might be able to shift the driver off center and perhaps closer together than the smaller tube would allow, depending on the shape of the drivers flange. Of course, that means that the off axis response is not exactly the same around the tube, since the distance to the edges of the tube varies. However, that could have some positive results from a diffrations standpoint and one one side you have the adjacent tube affecting your driver anyway, so it may not be detrimental overall. The larger diameter would also point to a lower crossover point, which might be problematic. Measurements should tell the story.

Very good point about the volumes used up by the typical ribbon tweeter. I had not thought about that. The Fostex tweeter you linked is indeed very small compared to the other commercial ribbons I found. It does look like it has some excellent potential and may allow the use of its own curved waveguide, etc. Thanks for posting it!

I will probably start with a small neo dome and test it out. I purchased 4 of the Vifa OX20 tweeters and they appear to be perfect for this kind of use, so I will try them first. If they don't work, then it is probably because domes of any kind will not work. I will start with 4" pvc tubing initially just to see what I run into. If it appears to have potential with initial measurements, I will look into removing the ferrofluid from and testing the second pair.

I should be a fun and challenging project.

[dlneubec]

I'm trying to find a way to increase the likelihood that a simple neo dome tweeter will work in this design. The main concern is due to it's 360º radiation pattern and how it will radiate up toward the ceiling as well as out toward the listening position and back towards the front wall. However, as the radiation pattern goes from horizontal to vertical, the waveguide created by the tubes gets narrower until at 90º off the listening axis (up of down), the waveguide has become a slot that is perhaps only 1/8" or so wide.

Here's one way that this potential problem might be addressed. It would take a material that will not absorb too much sound, but is soft and pliable. I don't know if the typcal foam rubber you get from a for bedding will absorb much sound or not, but this is the kind of material I have in mind. Start with some that is 2" thick. Cut a circle out of it about 7" in diameter or so. Now cut that circle in half. Assemble the two cans together with these half circles placed between them on the top and bottom, centered above the tweeter. The bottom of the two curved parts would be adjacent to the tweeter. When pressed and held together, the cans would compress the center of the two pieces so that they are only 1/8" apart, however, the outside egdes of the foam would fit to the edges of the can, effectively making a vertical component to the waveguide. Below are some drawings that show a view of this from the top and front. I think from these, you can better follow what I'm trying to describe. You would end up with both a horizontal and vertical flare to the WG.

So, any thoughts? Anyone know if this kind of readily available foam would absorb too much sound?

[RAAL]

Off-centering the mids to keep them closer together is a good point, but if anything is even remotely loose or elastic, you will loose the advantage of vibration cancelling, which is another strong point of the original design.

I mount the midrange flanges (after machining) directly to the rim of the Al pipes. The pipe works as a very stiff connection rod between the opposing drivers and provides such vibration cancellation that only the diaphragms are left moving. If you do that, you'll get just the pure sound from your mids and no cabinetry vibration, whatsoever. If you make couple of shortcuts here and there, you'll not have the same thing like I do, and if you don't end up happy with this, it will reflect bad on my design. You know, people associate loudspeakers to one another by their design features and I'm afraid that if you make a poorly executed version, it will be attributed to my design, not your execution.
I think that this design is so simple and so good, that you should just followed and not change it in any way.
Using cheap Al pipes, and do some lathe machining is certainly better than making complicated wooden cabinetry.
Missing to make a rigid speaker-to-pipe connection will lose the vibration canceling, making it just somewhat better than wooden cabinetry, but not acoustically disappearing.
Sizing up the pipes will decrease the lateral HF distribution.
Using dome tweeter instead of the narrow slot ribbon will further still decrease lateral HF distribution, beyond calling it omni.

The whole point with this design was to make a ribbon play omnidirectionaly, have enough midrange surface area (4 x 29.5 sq.cm) of almost like 6.5" driver, but without cone breakups in the pass band and to make the cabinetry dissapear acousticaly.

The success of simple designs depends on their execution. Change one thing and you changed everything.

I'm not saying that you shouldn't try to change anything, just when you do, consider how it influences the rest and change the design accordingly.

For example, if you are to use a dome tweeter that you are to listen on a side, just mount it on a stick, 4" above the can, point it upwards, and use just one can, as you don't need the waveguide then. You would still listen it on a side, and it will be omni for sure.

[dlneubec]

I don't have the means to do special milling of aluminum, so I will have to rely on adhesives or other mechanical connections. I did find some thinner walled pvc coupling hubs that offer some interesting possibilities. They are 3-5/8" long and have a extruded ring or lip in the middle of them that extends about 3/16" into the tube. If I were to glue three of them together and cut the ends off of the two outside tubes, the extruded lip would act as a place to mount the baffle. I was thinking of using something like 1/4" aluminum plate, perhaps, as a baffle. It would be epoxied into the tube or mechanically connected (see below). I would think that would make for a pretty rigid connction between the drivers and in combination with the opposing driver arrangement, will have the vibration cancelling effect we both are looking for. One interesting thing about this 3-piece approach is that the center section might be prepared for the tweeter before the composite of three tubes are glued together. This could make it much easier to adapt for a ribbon tweeter, for example, during the prototyping stage.

One other possibility might be to use some small diameter threaded rod from driver mounting hole to the opposing driver mounting hole, across the tube, that could be bolted on each end to pull the drivers and structure together very rigidly. I'm not sure it the steel bolts would have any negative effect on the unshielded drivers, however.

An upfiring tweeter is not a solution I'm interested in exploring. I've done numerous omni designs over the past 5 years, and have tested that solution, amongst others, and I'm quite familiar with how that would turn out. I'm not set on a dome tweeter, but I want to start with one to see how the measurements look before I reject it. I probably will try that Fostex ribbon, before I decide which way to go.

If this design is not an exceptional performer, no one will hear it, as it will never go beyond a prototype stage. If it is an excellent performer after a prototype, then I'd be more willing to go to more expensive materials for the final build. I'm afraid I don't have the budget to start with the most expensive materials at the beginning of an experimental design.

[EdL]

Dan,

Just read through the thread.

Reticulated (open cell) foam, 30 ppi, as used in aquarium filters. It is available in a variety of thicknesses, can be shaped with an electric carving knife, bears a striking resemblance to the foam used by RAAL and Geddes for diffraction (HOM?) control.

If you prototype with the dome tweeters, how about having opposing pairs staggered (1/2 space), instead of directly opposite?

In your diagram above, I would be interested in making the tweeter backset adjustable with a screw, to allow searching for a sweet spot....might even do that for the cylinder spacing...just a brainstorm...hide the threads in the foam?

[Rick Craig]

I thought it would be good to update everyone on the CBT kit from PE. Fortunately I work in the same building where the design is happening and it's been enjoyable to see things progress. The prototype tweeters have been tested and production units should be in the works fairly soon; however, I don't know when the parts will be available for sale. I believe the tweeters will be available separately from the kit.

The cabinets will be in knockdown form and cut with a CNC (almost impossible to do these by hand with a router given the complexity). I'm not sure what the final pricing will be but the folks at PE think this will be a good selling kit. The Behringer DSP crossover is being used for the kit but of course you could use something like a DEQX or similar unit. The Dayton demo was from the original DSP curves and Marshall Kay has been experimenting with some alternate voicing since that time. The good thing is that you can tailor this to your room since you'll have the freedom to adjust the DSP settings.

[dlneubec]

Dan,

Just read through the thread.

Reticulated (open cell) foam, 30 ppi, as used in aquarium filters. It is available in a variety of thicknesses, can be shaped with an electric carving knife, bears a striking resemblance to the foam used by RAAL and Geddes for diffraction (HOM?) control.

If you prototype with the dome tweeters, how about having opposing pairs staggered (1/2 space), instead of directly opposite?

In your diagram above, I would be interested in making the tweeter backset adjustable with a screw, to allow searching for a sweet spot....might even do that for the cylinder spacing...just a brainstorm...hide the threads in the foam?

I remember reading about the retiuculating foam that folks were using for their Geddes clone speakers. Isn't that essentially trasnparent to sound waves? I would think we need something that will act more like a solid surface in this case, since it would act as the vertical sides to the waveguide. Perhaps it could be covered with something else, not sure about the possibilities there.

My thinking all along has been one tweeter in one of the two cans, not two, especialy across from one another, because I suspect the tubes will need to be as close as the drivers will allow. With domes, it will probably be the dome itself that will determine the minimum spacing. However, opposing, staggered pairs would allow for the inter-tube spacing to remain the same. Two tweeters would certainly increase sensiivity or headroom. I presume the two tweeters mounted in a vertical line would have the additional benefit of reducing vertical dispersion, which might be a positive here. Definitely worth experimenting with. I'll take a look in The Edge. Thanks for the suggestion.

I will certainly want to be able to adjust the spacing between the tubes at the prototype stage. As you suggest, it will be valuable to be able to determine the ideal spacing based on measurements. Yes, I was thinking of hiding the tube connectors in the foam, should that be the way it ends up going. One could also use aluminum tubing, wood dowels, acrylic tubing, etc. to cover/hide threads and act as spacers. Even a stack of washers could do the trick. I have not given much thought yet on connection methods, but I need to. If you have any further thoughts, feel free to post them.

I'm not seeing an easy way to make the tweeter back mount adjustable, since the tweeter is mounted in a tube with sealed drivers, unless a separate sealed chamber is made for the tweeter. Did you have anything in mind from a construction standpoint?

Here is a drawing with using the pvc hubs I mentioned above and with the Vifa OX20 tweeter in place.

[dlneubec]

Rick,

Thanks for the heads up on the CBT. This is one kit I might be interested in. I will be anxious to hear the pricing. The tweeters, if available, might work on this project as well.

[dlneubec]

I did some prototyping today and thought I'd pass some photo's along. As I mentioned above, I found some pvc coupling hubs at Lowe's and decided to use them for the prototype. I also did some thinking about how to make the tweeter front to back position adjustable and came up with a way to do that as well.

As will be made clearer in the photo's, I used a smaller diam. pvc tube (1-1/4"id) to create a tube the tweeter can slide in and out of. It turns out that the ID of the 1-1/4" pvc is a little bigger than 1-1/4". I decided to use 1-1/4" hardwood dowell to slide in an out of the tube to provide the back for the tweeter. It turns out it is just a little smaller than 1-1/4" od. I had some 1/2" width UHMW adhesive back tape that turned out the provide the perfect spacer for a snug fit between the dowell and the ID of the pvc tube. For prototyping, I'm going to simply try double sided masking tape to hold the tweeter against the dowell and I still have to come up with a way to get the wires into the tweeter.

Once the 1-1/4" pvc tube is inserted into the side of the pcv coupling hub, the outside of the protruding sections of the 1-1/4" tube were sanded flush with an upright belt sander. Then I used a chamfer bit on the router table to create a chamfer around the ID of the tube and then sanded the sharp edges to create a bit of a roundover. I will probably try the opposing, offset dome tweeter option previsously discussed as well, in which case the location of the 1-1/4" tubes will shift up and/or down on the opposing tubes, to create the offset.

Each larger woofer tube is made from 3 sections of the pcv coupling hub. The center one holds the tweeter and the top and bottom ones are cut down so that the internal flange on the hub can act as a backer for the baffle. For prototyping, I'm simply going to use 1/4" plywood as the baffle.

I think the photo's below will make this clear. The drivers are the Vifa (Tympany) OX20SC00-04 3/4" neo dome and the 16 ohm verion of the Vifa (Tympany)TC9FD-05-16, which is a 3-1/2" full range driver, that is on buyout at PE for $8 each. The larger woofers in the bas bins are still undecided and will depend on how things work out on the MT section. Below are Tympanies published FR for these two drivers. Actually the mid FR below is for the 8 ohm brother to the 16 ohm version. BTW, 4 of the 16 ohm drivers in parallel, shows a sensitivity of 93.5 @ 2.83v/m. As far as making the tube diameters smaller, that appears to be unlikely since the Vifa driver is 3-7/8" across, so it would take a considerably smaller driver to fit in a 3" dia tube and I already own 8 of these drivers.

Also of note, I would expect th sagging response of the Vifa ox20 to be completely filled in by the effect of the WG. Take a look at the tests Zaph did on this tweeter in a just a 1/2" deep-1/2" roundover WG and you can see that the dip is almost entirely filled in. I also suspect that the WG will bring the tweeters overall sensitivity up significantly.

OX20 FR



TC9FD (8ohm version)



Prototype photos:

[Saurav]

The photos help a lot. And it always amazes me how your rough prototypes always look better than my finished versions

[dlneubec]

Thanks Saurav. It is just a matter of hiding the flaws! :W

[dlneubec]

Update:

I have a prototype of the MT portion ready for measuring the tweeter and one pair of the mids (see photos below). I used the section that was cut off from the end of the outside section of the tubes the narrow piece in the first pic) to create a lip that each tube can be connected with. I cut a section out of those strips, just enough so they would fit tightly inside the existing tubes. They were then glued in place inside the central tube and then the outside tubes were slipped over them and glued. I left a little gap between the secions just for looks, though I may fill it in if it causes diffraction problems with the tweeter. You can see the lips inside the tubes in the second photo.

The two assembled tubes are held together with cap head bolt and hex nut with spacers between the tubes to set the distance. I have spacers that are both 1/16" and 1/8" so I can stack them to adjust for the most effective distance between the tubes, from a performance standpoint. That is one of the sets of measurements I intend to take, comparing the results with different spacing between the tubes. A second set of measurements I want to do will be with the tweeter in a different in to out position, which is also adjustable (from the inside).

One other thing I intend to try is to place a v-shaped wedge of wood or plastic along the tube opposing the tweeter. This will be an attempt to reduce reflections off the opposing tube right back at the dome. I have two different sizes of these at this point, one is about 1/8" deep and the other 1/4" deep, requiring different spacing between the tubes.

For the purposes of testing, I made only the MT section and a prototype arm and a base to hold it up. The big bass bin tube that holds the woofers will eventually be used to anchor the design, but I will worry about the woofer section later, when I'm comfortable that the MT section will work, especially the dome tweeter.

I have some foam rubber inserts to form vertical sides of the waveguide that I may try as well, but those are pretty roughly cut. There has to be a better way to do them.

I hope to start some measurements tomorrow.