Omnimatrix: low budget, big sound?

[Pete Schumacher]

Re: Omnimatrix: low budget, big sound?

Now that design, I would name it the Bikini

Toucans, bikini . . . kind of goes together.

[Saurav]

Re: Omnimatrix: low budget, big sound?

Maybe he should just call them the Mai-tais.

[dlneubec]

Re: Omnimatrix: low budget, big sound?

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.

[donradick]

Re: Omnimatrix: low budget, big sound?

Thanks for the update. I've got my cart loaded with the drivers as soon as the choices are finalized. Another ambitious effort from Bloomington!

[lunchmoney]

Re: Omnimatrix: low budget, big sound?

Wow, so cool.

Toucans!!!

[dlneubec]

Re: Omnimatrix: low budget, big sound?

I have my measurement gear up and working again and did a lot of measuring last weekend. I'm still trying to find the best solution for the tweeter mount. It appears that some sort of v-shaped mount on the opposing tube provides the best overall response. I took the best measurements of my attempts so far and worked up a 2-way crossover based upon measurements from just a pair of the midranges (adding the pair twice in the model to equal the output and impedance of 4) and have a pretty decent looking response, but I want to continue to experiment with the tweeter mount to try and find the best possible solution.

I still suspect that a ribbon might be the best solution. I've also been experimenting with controlling the vertical off axis with a dome tweeter and have had some really decent looking results, at least up and down 15º.

Here are some preliminary graphs I can post. None of these are normalized to real world spl levels, so you can ignore that aspect of the graphs.

Here is a 2.5Khz or so crossover model with Reverse null and impedance plot


Here is the phase tracking from the crossover above


This is horizontal at 0-15-30-45 off axis. The one that is tailed for HBT is the on axis measurement used for the crossover above.


As rough as those plots look, here is one with them all summed together. You can obviously see the boost provided by the waveguide effect.


Here is the tweeter on axis horizontally, but at 0-7.5 and 15º up. I should say that all these have a narrow strip of felt above and below the tweeter in a V-shape which seems to improve the vertical response. All the tweeter measurements were at about 1m.


Here is what the woofers response looks like, all 4 in this case were on. This looks pretty good, but it is on axis only and the problem areas will show up when I do off axis.

Impedance, impedance phase (drivers include a 47uF cap high pass)


Frequency Response on axis at the tweeter height, centered between the tubes, 1m.


This weekend it is back to more experimentation with the tweeter mount and opposing tube wall treatment. BTW, it looks like the curve on that Fostex ribbon tweeter is almost a perfect match to the 4.5” dia. tubes, so that may be worth a look in the future.

[dlneubec]

Re: Omnimatrix: low budget, big sound?

I’ve made some good progress on the dome based design. It has been a real challenge, with a lot of DIY to it. The main challenge, as expected has been to get a good response from the side firing tweeter. Though I haven’t taken any photos yet, I ended up with the optimum position between the tubes at ¼”. The treatment on the opposite tube that has worked the best is a combination of elements. The first treatment covers about 2”, centered on the tweeter. About a 3/16” deep, 3/8” wide V-shaped piece is centered on the tweeter, running vertically. This acts as a deflector to push sound toward the front and rear. This deflector is flaked by 1/8” deep x ¼” wide felt. From the measurements, Id say, this acts to reduce reflections of the very high frequencies back and forth between the tubes. To even out the response even more, I added to other strips of 1/8”d x ¼” w x 2”l pieces of felt to the tubes, alternating from one tube to the next. The first strip is adjacent to the tweeter on its tube, at the front edge of the round over from its opening. The second strip is back on the opposing tube but placed further along the edge of the tube. What you end up with is 3 strips of felt, alternating from on the opposing tube, beside the deflector, to the tweeter tube, at the edge of its opening and then back to the opposing tube, each at different distances, therefore, I think, catching reflections going back and forth at various intervals from the mouth of the waveguide.

The second main treatment is intended to provide a vertical element to the waveguide. This is done by placing to somewhat rounded V-shaped pieces that extend from the top and bottom of the deflector/felt treatment to the top and bottom of the tubes. These piece come to a point at the edge of the deflector/felt treatment and then wide to about 1-1/2” at the top and bottom of the tubes. They are made of ¼” closed cell foam. Ok, I’ve decided to take a couple quick photos to help you to understand what I described above. In the next post, I’ll move on to a bunch of measured results.

[dlneubec]

Re: Omnimatrix: low budget, big sound?

The measured results below represent only the MT portion of the project. I wanted to be convinced of the validity of the MT portion, in particular the dome tweeter, before worrying about the Woofer/3way effort. I can say that I’m quite pleased with the results so far and will shortly move into constructing a bass bin and working on the 3way design. The first set of on axis measurements were taken at 2 meters, with the mic at the LP of 39”, which is actually even with the top of the tubes. I want to keep this design on the diminutive side and did not want to make it another 4” taller in order to get the tweeter exactly on axis. The vertical and horizontal off axis measurements were taken at 1 meter, due to the difficulty of doing that at 2 meters. I will publish them in the next post.

As far as how the MT sounds, I set it up with a biamped setup using a Dayton RSS265HF subwoofer, set to crossover actively at 180 hz. The high pass on the MT is currently set at around 200hz, with the MT crossover happening at around 1900hz or so. The model shows it at around 1800hz, but that is using older measurements before I optimized the tweeter configuration. The sound is quite amazing for such small, inexpensive drivers ($8/mid with a $16 tweeter; total per side = $48). There is a substantial depth to the soundstage and a naturalness of the midrange that is surprising for a single speaker. I’m very anxious to see how it sounds with a well integrated woofer and then with a second speaker. These will play very loud without noticeable strain as well.

Here are the mid and tweeter measurements with and without crossover, so you can see the attenuation that has taken place in the crossover. Note that the mid is down 8-9db at the crossover, while the tweeter is down around 22db at the crossover, after equalizing the boost provided by the waveguide tubes.

Midrange with and w/o Crossover


Tweeter with and w/o Crossover


The tweeter response is still fairly raged on axis, but you will when you average the responses of the horizontal and vertical off axis, it looks pretty good. I have not found any of the response anomalies to be audible so far in some pretty extensive listening sessions.

Here is the measured response of both drivers, individually and together along with their acoustic phase response. You can see that the phase tracks very well from around 1khz to 6khz. This is followed by a graph of the driver responses and a wide, deep reverse null. Finally is a plot of the on axis response at 2m, with a large gate of 25ms, which essentially incorporates the room response.







Here is the measured impedance and impedance phase:



Next up are some off axis response and comparisons.

[dlneubec]

Re: Omnimatrix: low budget, big sound?

This post includes some vertical and horizontal off axis response measurements, some merging of those responses, comparisons to the on axis response and comparisons of vertical and horizontal off axis summed results.

First graph is a vertical response at 0º, up 7.5º and up 15º. Above the individual responses is the summed response of the three measurements below. Very flat from 300hz to 9khz, the meat of what we hear. That is followed by a graph comparing the on axis response to the summed vertical off axis response. They are quite similar.





Next up is the horizontal off axis responses at 0-15-30-45 degrees, with the off axis summed measurement above. You can see that at 45º the upper end gets quite ragged, but is pretty good out to 30º. I’m not sure you can really call this speaker truly omnidirectional. The response pattern, due to the tweeter waveguide directivity and the off axis cancellation due to dual, side by side mids really don’t allow for the same response through 360º degrees. That said, it might be more bipole-like in that the front and rear response out to 30º is pretty good. Notice however, that the on and off-axis response when summed is surprisingly flat, until the rise in tweeter response at around 13khz. That is followed by the horizontal and vertical off axis summed responses plotted together for comparison.





Next is shown the horizontal and vertical off axis summed responses, with , in turn, their summed response above that. This should be pretty indicative of the in room power response, at least out to the range I took measurements for. Finally, I have a graph that compares the summed vertical and horizontal on and off axis responses plotted with the on design axis response, which again, are quite similar.





That tells the story of where this project is at this point. I’m quite happy with the results and if the woofer/bass bin integration works out well, then I may proceed with rounding out a pair with the dome tweeter, saving the ribbon options for a more expensive future version of this design

[critofur]

Omnimatrix: small tubes, big sound...

I like: "Omnimatrix, sound that dominates your room!"



Dan! Great of you to post all those measurements! It's wonderful to have so much information about interesting projects. Thank you. I hereby officially grant this thread a five star rating!

-------------

I'll want to copy/clone your/Raal's design also, if I can find the time... (Or rather, if I wasn't so inefficient; spending all my free time browsing the web and forums!)

I had ordered some of those tweeters earlier and they were just wrapped in cardboard! No protection for the dome, it was squished flat. A quick call to PE and replacements were sent right away. This time, they are WELL packed in boxes.

I may use the ND105-8 x 4 wired series/parallel to net 8 Ohms nominal, if I get around to duplicating your efforts.

I really like using cylinders for speakers. Both aesthetically and for the practical benefits gained. If only... I could daydream about cylinders made from laminated stainless steel + aluminum with a thin damping layer in between. I have done a little testing of this combination (ss + al w/damping in a curved enclosure) it yields an awesome combination of stiffness/deadness/compactness far beyond any existing speaker design I've encountered.

[donradick]

Re: Omnimatrix: low budget, big sound?

Dan: Looking very good! For some reason, I have a strong desire to build these. The low cost of entry to a good omni design, and the simple cabinetry are very attractive.

[dlneubec]

Re: Omnimatrix: small tubes, big sound...

Thanks, Critifour.

It's been a fun project, especially since I enjoy tinkering, since there were tons of experiments with treatments of the tube to tweeter interface, the depth the tweeter resides in its housing and the distance between the tubes.

Bear in mind that the size of the tubes is a key ingredient. As they get bigger and the mids further apart horizontally, the more you have off axis comb filtering issues, which in turn determines the crossover points. The same goes for the vertical separation like any MTM. What I'm getting at is you need pretty small mids, with low cabinet volume requirements in order to make this work. In this case, these are about 4-1/4" ID tubes, with a little less than 8" between the baffles, for around 3-3.5 liters of space for 2 drivers. I'm guessing the NF105's would not be a good choice. Perhaps something like the ND65 or RS75, or perhaps one of the Fountek's, FR88 or FR89 would be possibilities, though I have not looked at them. These little Vifa TC9's really sound great.

I think the little ox20's are going to be great in some waveguides. I tested one without baffle in a little 2-7/8" dia waveguide and it really looked good from about 3kkhz on up. I have a couple extra's so I will eventually try them on a baffle for some more low end support in those little waveguides, plus in some larger ones, like the MCM 6-1/2".

One cool thing about this design is the opposing push push driver arrangement and how it does a great job cancelling out vibrations. There is very little vibration you can feel with your hands, and these tubes have just 1/8" thick pvc walls, though there is a rib in the center and at each end, pluse the overlapped lip between the tubes. The same will be true of the bass bins, which have opposing push push woofers.

I like: "Omnimatrix, sound that dominates your room!"



Dan! Great of you to post all those measurements! It's wonderful to have so much information about interesting projects. Thank you. I hereby officially grant this thread a five star rating!

-------------

I'll want to copy/clone your/Raal's design also, if I can find the time... (Or rather, if I wasn't so inefficient; spending all my free time browsing the web and forums!)

I had ordered some of those tweeters earlier and they were just wrapped in cardboard! No protection for the dome, it was squished flat. A quick call to PE and replacements were sent right away. This time, they are WELL packed in boxes.

I may use the ND105-8 x 4 wired series/parallel to net 8 Ohms nominal, if I get around to duplicating your efforts.

I really like using cylinders for speakers. Both aesthetically and for the practical benefits gained. If only... I could daydream about cylinders made from laminated stainless steel + aluminum with a thin damping layer in between. I have done a little testing of this combination (ss + al w/damping in a curved enclosure) it yields an awesome combination of stiffness/deadness/compactness far beyond any existing speaker design I've encountered.

[Wolf]

Re: Omnimatrix: low budget, big sound?

Are you using the tubes you got from me for the bass-bins?
Just curious,
Wolf

[dlneubec]

Re: Omnimatrix: low budget, big sound?

No, though I took a good look at doing so. Even the smallest ones were just too massive at 15.75" dia. Otherwise they would have been perfect. I will probably be going with 12"Id or 14" OD tubes. I can get 12" ID pcv pipe, but all the drivers I'm considering want larger enclosures for dual woofers, so I'm expecting to use 14" od plwood half-rounds. I may very well make them in an upside down U shape with a half round top instead of circular tubes, so that I can get more volume in the bass bins and some room for a port. That can get me up to about 58 liters in a 2' long tube (baffle face to baffle face) which gets me into the low 30's F3 with a sensitivity in the 91db range, depending on what driver I use.

The 4 parallel 16 ohm mids have a sensitivity of around 94-95db and with the current attenuation of around 4db, that should be a pretty good match with the dual woofers at 91 db or so.

[dlneubec]

Re: Omnimatrix: low budget, big sound?

After some test listening with an active crossover for the woofers, I've decided to go with the 12" Dayton Buyouts I picked up a couple years ago for $20 each, in a sealed tube. With in room/boundary gain factored in, the graph below is a simulated output with the low pass shown at 200hz. This is a high QTC arrangement, with the box highly stuffed. F3 is predicted at 31hz, but f6 is at 19hz and it never gets down to F10 according to this model. This model figures on 25% room leakage. You can see the normal response without the room/boundary gain figured in, so I'm getting a substantial low end lift with the woofers this close to the floor and rear woofer figured at about 18" from the front wall. I plan to use a C-core coil on the woofer for an extremely low dcr to attempt to get the sensitivity (with room/boundary gain included) up to the point they will mate well with the mid/tweeter combo without additional attenuation. I will be measuring these soon and will be looking at the 3way XO once the measurements are complete. These are really sounding very good. I liked the dual 12" sealed low end as compared to the ported pair of RS180's. You can see I'm getting over 110db at 150w without reaching excursion at any point.



Here are some photos of the project. I've now added a 24" long 12" pcv (SDR35 sewer pipe) tube bass bin. The upright section is made from 1-3/8" thick red oak table tops that you can now get at Menard's in a 24" x 60" size. I've added the hinged top section so that the MT section can rotate a bit off axis from the woofer/upright support. I will also be starting on the build of the second speaker before too long so I can evaluate a pair.

I'm not sure on the finish at this point, but I'm leaning toward a very dark/black type finish on the wood portions, perhaps high gloss with a contrasting flat, off-white finish on the tubes. I really like to soft look of the sanded pvc bass bin tube that you see below. The passive XO will be housed in the bass bin and I will hide the wires to the MT, probably in a routered slot down the inside curve of the oak.