Build Thread: OBEOS

[Tytte71]

This project initially started on a Norwegian forum with the goal of improving the renowned Acoustic Research AR-9. I believe I found my way around this and one of the improvements was to achieve better dispersion control.

Since then I have been through a process understanding more about the importancy of dispersion control and the effect waveguides and OB have on this matter. Slowly the original project started to fade out for the benefit of a simpler and more efficient design. In this process lots of help have been given from Jzagaja and Snickers-is (can not thank them enough for their patient and willing to share their knowlegde) and from people (like StigErik, Gedlee, Linkwitz - and lots of others) that have made theory and practical measures available on the Internet and especially on this forum.
I have a lot to learn in order to fully understand all aspects and theory behind loudspeaker design and my consolation in this context is that the world is not actually overflowing with people that really know and can this stuff

I would like to share my thoughts and design ideas in order to (hopefully) get creative discussions.

The design and boundaries

The premises for the project is to make a relatively compact system with dispersion control all the way down to the "sub frequencies". Though designed for high-end home hifi, the system shall have the capacity of playing extremely loud without a hint of dymanic compression and nasty distortion. The sub frequencies will be covered by use of four distributed "compact" subs. The main speakers will be made as a two-way design. The listening distance can vary between 3 - 4 meters. The living room (29m^2) is far from ideal with open space to kitchen, 2'nd floor and hallway (have not yet been able to achieve realistic bass level in this environment)

For electronics I will use two Lyngdorf amps (TDAI2200 - running them without RoomPerfect) for the main speakers and IcePower 1000ASP modules for the subs. The amps will be controlled by my HTPC with Lynx AES16 running JRiver. XO and DRC filters is created with the Juice Hifi Audiolense 4XO software.

A compression driver with an elliptical OSWG (Horns) will cover 800Hz up.
The Elliptical shape enables a sleek design and allows for closer c-c distance between tweeter and midwoofer than a circular OSWG. I will loose some in vertical dispersion control, but I do not expected it to be of great significance. I will also expect the horisontal dispersion to be in the range of the sonogram shown for Horns but with extended coverage down to 800Hz.

The tweeter will cross over with a 12" midwoofer suspended in a open frame using no baffle (the design will allow for later mounting of open baffle og cabinet if necessary). The first dipol peak have its maxima just above 800Hz. The midwoofer will be used down to ca. 100Hz.

The four subwoofers will be placed into the walls in front and at side of the listening position, hoping to cheat the worst room modes. The subs shall be able to play all the way down to and below 20Hz without moving outside its linear stroke range at 115dB.

Driver selection:

Selecting optimal drivers is really difficult and my experience with professional drivers are next to zero.

Tweeter:
For the tweeter I have been recommended using the 1" BMS4550, which have good reviews and which BMS recommends to be used 800Hz up.
But I can also see that BMS has some new 1" and 1,4" drivers that could be relevant candidates. Then I have also concidered the Radian 745NEOPB. Nothing concluded.

Midwoofer:
I have concidered using AE TD12M, Ciare NDH 12-3, 18Sound 12NDA520 or a Beyma variant.
Nothing concluded here either, but the Drivervault site (drivervault) indicates very good performance for the TD12M (midbass dispersion control in the kHz range is of minor importance here). I am also very curious about the Ciare.

Subwoofer:
So far I have found the FaitalPro 15XL1400 to be a strong candidate when mounted in a ~60-70l BR box.
Looking a typical hifi drivers, the TC sound Epic 10 and 12 might also be good candidates.
I am not keen du use any kind of dipole subs.

The sub/midwoofer combo shall deliver that delicious chest pounding punch that I so very much appreciate

Drawings and curves will be made available very soon.

Hope someone find this project interesting.

Best Regards

[Tytte71]

Experiences with the drivers listed or recommendations for selection would be highly appreciated

[keyser]

This sounds like an exciting project!

You say you want to use a 12" midwoofer in dipole mode and a horn for the upper mids and highs. When I decided to build a speaker with a waveguide, I've contemplated using a dipole for the mids too, but ultimately I think this is not a very good idea. It's impossible to match the directivity of a dipole and conventional waveguides. At the dipole peak directivity of the woofer is probably about 5.5 dB's. The waveguide will have directivity closer to ten. What's more, you'll need a lot of amplifier power if you want to get very high sound levels, because a dipole is so inefficient. My solution to the problem can be found here: 2-way: Waveguide + Cardioid-like

[Tytte71]

Thanks for the very interesting post Keyser!

My construction will allow for flexibility and testing various solutions. If the 12" is not optimal, I can at any time decide to try a smaller... or bigger. The frame will be surrounded by an acoustic transperant material.

I see your point in difference in directivity between the dipole and WG, but simulating the nude 12" in Edgde shows very similar dispersion to the WG (horizontal axis) at crossover. The WG is expected to have 10dB attenuation at 60deg and so is the 12".

Opposite to mounting the midwoofer in a sealed cabinet, I will anyway achieve greater dispersion control down in frequency, even though not 100% matching the WG. My goal is to utilise nude driver's cardiode characteristic to get an even power response. It was in this context very interesting to read about your solution and the solution on http://www.diyaudio.com/forums/multi-way/142691-adventures-cardioid.html). I'll definitely have to try something like this.


By the way - the tweeter cabinet will be made separate to the midwoofer mounting.

[Tytte71]

A picture of a custom EOSWG from Jzgaja that I will use. Exterior size of the cabinet 450mm (w) x 300 (h) x 400 (d). Will be placed on top of the midwoofer frame. Throat size 1".

[keyser]

You are right about the approximately 10 dB's drop in level at 60 degrees at the crossover frequency. At the front you can probably get a pretty good match. But don't forget about what happens at the rear. There you get up to 6 dB's of extra output. Don't forget about this, because if not accounted for it will mess up the power-response at the crossover.

Unfortunately I don't have good simulation software for directivity (Tolvan edge is usable, but ultimately limited), but if I had to guess, I'd say a closed box with minimum sized baffle and an unbaffled woofer with a driver of the same size have about equal DI at the frequency of the onaxis dipole peak. All in all I don't see what is gained by using a dipole - but the cons are obvious.

[Tytte71]

Originally Posted by keyser

You are right about the approximately 10 dB's drop in level at 60 degrees at the crossover frequency. At the front you can probably get a pretty good match. But don't forget about what happens at the rear. There you get up to 6 dB's of extra output. Don't forget about this, because if not accounted for it will mess up the power-response at the crossover.

Couldn't the back radiation issue at the dipol peak be solved by use of muffler (LPF) in combination with the midwoofer size selection?

Originally Posted by keyser

Unfortunately I don't have good simulation software for directivity (Tolvan edge is usable, but ultimately limited), but if I had to guess, I'd say a closed box with minimum sized baffle and an unbaffled woofer with a driver of the same size have about equal DI at the frequency of the onaxis dipole peak. All in all I don't see what is gained by using a dipole - but the cons are obvious.

I also use the Edge app to get some indications. Only real life measurements will tell the truth
As I understand, the radiation of a driver mounted in a narrow baffled closed box (more or less) will be determined by the driver's dispersion pattern down to the baffle step freq., corresponding to the dipol peak maxima.
Edge shows a dipol peak maxima at about 860Hz when suspending the 12" without baffle. I expect the 12" to have some natural dispersion control below this frequency. But maybe I should look for a smaller midwoofer to increase the dipole peak frequency?
Anyway Edge (again ) shows decent dispersion control also beyond the dipole peak for the 12" (presume that is the simulated driver response).

I am trying to achieve a more even power response than a closed box. The closed box will gradually start radiating 4pi at relative high frequencies. This is well illustrated on the Gedlee Abbey. The OB variant will radiate gradually into a 8-figure, reducing the effect on early reflections at lower midrange frequencies. If I gain anything on this? I really don't know - and I ehhh don't according to many .

I am not really afraid of loss in power capacity of the OB when limiting its lower operation range. At 100Hz it should still be possible (again, simulations) to achieve a "system sensitivity" of about 90dB/W and with a sound level of 110dB/100W/100Hz well within the driver's xmax. A second midwoofer could be added to increase capacity in the lower frequency region for the actual system (taking into account fs and Qts).

My experience from a "similar system" is from a Lyngdorf dipol system (DP1/BW1) that I once had. In this context I did not find the capacity of the midwoofer to be the limiting factor, but rather the lack of capacity in the "subwoofers" and the far from ideal dispersion in upper-mid/high frequencies. Playing louder than ~100dB with this system made no sense.
It should be mentioned though that the x-over between the 6,5" Seas mid and the sub was ~350Hz.

[Tytte71]

A short update... The tweeter cabinet and midwoofer frame is taking form, at least at the drawing stage . A midwoofer, two front legs and a reinforcement plate will be added. Reading Keyser's building thread and its referenced threads has inspired me to try a muffler (maybe a figure cut foam) around the midbass driver or add damping material inside the frame (sides and front will be covered with some kind of mesh or cloth). Colours of the speakers are not yet decided.



Warm recommendation for Ciare NDH-3 has been given by Al at Usspeaker... concidered to be the smoothest midrange in their portfolio.

The compression drivers alternatives have added up to:
1. BMS4550
2. Radian 475PB
3. Beyma CP385ND (highly recommended by Usspeaker)
4. Faital Pro HF10AK

Any proposals from anyone that have compared any of these drivers or have strong preferences, are highly appreciated.
Must admit I'm a little reluctant to use the Beyma unit with regards to Beyma's recommended crossover frequency - Do I have to be?

Best Regards

[Tytte71]

I was thinking of something like this. The midwoofer does not show on this picture. Have so far skipped using legs in front.

The inside of the metal mesh grille can be stuffed with acoustic dampening material, but I think I will try making a muffler mounted on the back of the driver first. What concerns me a little, will the metal mesh have negative influence to the midrange?

I am also a little concerned that the tweeter cab can cause some diffraction issues to the midrange. The leg should not have any influence at 800Hz down I believe.

Think it starts to look okay, but not quite satisfied with the foot - does not match the rest of the lines I feel. Input anyone?

[boris81]

It's looking great so far! I like this a lot.

I think the biggest issue with the metal mesh will be keeping it from vibrating. That may or may not be easy to do. I can't give you advise on baffle diffraction issues but dipoles seem to not be affected by it much.

You will achieve the best stability if you mount the woofer by it's magnet to that rear brace. Then maybe a minimal baffle can hang by the woofer's flange on the front (detached from the rest of the cabinet). You can experiment with stuffing in the back to get the dispersion you need. Are you looking for a dipole or cardioid dispersion from the woofer?