Open baffle design, 2.5-way

Jump to Latest
D

davirix

Member
Joined 2022
#1
Hi,

I’m planning my first DIY sound system. I’ve read what I could reasonably absorb over a couple of months of afterwork readings, and all things considered I think I’ll go for an OB design, with 2x12” woofers, and one tweeter. Closer designs are Manzanita, Spatial Audio project M4…

Shortlisted drivers, based on basic criteria within my reach, are the Ciare FXI12.50W (https://www.audiovideoparts.com/disegni/FXI12.50W.pdf) and the SS DISCOVERY D2608/91300 (https://www.scan-speak.dk/product/d2608-913000/). I don’t really know how to make sure these two are a good fit apart from a good overlap of the SPL/Frequency SIM charts, which I think is ok. I’m open to suggestions, possibly made in EU, where I will buy them. Some questions:

  • Should I push more the analysis of drivers before purchasing or, as I read somewhere, the only way is to buy and check for yourself?
  • Can the two woofers be wired in series and so be treated as one by the minidsp/amp?
My idea is to have a music server – minidsp flex – 2 amps class D hypex NC122MP with 4 output channels total.
  • Would it be feasible, and a significant improvement, to put a passive crossover low pass filter for one of the two woofers (I’ve seen standalone components for that) not to exceed the total 4 channels planned?
  • If the tweeter has 8 ohm impedance, does it matter whether the woofers are 4 or 8 ohm?
  • In alternative to the 2.5 way system a.m., would it make more, from a sound perspective, to have 2-way main loudpeakers and add an active subwoofer?
No point in describing the listening room as I have to move every few years for work (approx 5x10m right now).

Any insight would be highly appreciated.

David
 
#2
Woofers are midrange for PA. Not enough Xmax for open baffle.
You can connect woofers in series or parallel, but .5 is possible only with parallel. Then impedance of a unit should be 8 or 16R,
Search for woofers with Fs around 30Hz, Xmax more than 10mm.

Dipole 2-way with 12" driver(s) and a dome tweeter can not give smooth response and directivity. If you like music to sound different in every spot in your room, it's ok.
 
  • Like
Reactions: tigeranand
#3
Thanks Juhazi, but one of the references for my design are these papers: http://www.quarter-wave.com/OBs/OB_Theory.html , where the author used a 15" woofer with a XMAX = 3.8mm. If I got it right high xmax means lower sensitivity and higher cost, hence I guess it's a matter of finding the right compromise.

Juhazi said:
Dipole 2-way with 12" driver(s) and a dome tweeter can not give smooth response and directivity. If you like music to sound different in every spot in your room, it's ok.
Click to expand...

Unfortunately I've not been able to audition any system of this kind... so you would discard as flawed the two above mentioned models I got inspired from?

Best,

D
 
#10
You should model the speaker you want to build using software to save yourself a lot of time and money. Of course building and testing is fun. I found that boxsim software, available free from visaton does a decent job of modeling an open baffle speaker if you place pairs identical woofers driven out of phase on the front and back of the cabinet and make the box thin front to back. Boxsim will compute the radiation pattern, as well as the frequency response and driver excursions so you can watch those woofers bottom out. Bass in an open baffle rolls off at 6 dB/ octave at a frequency determined by the baffle width. Woofers roll off at 12 dB / octave below their resonance. So unlike a sealed box you are rolling off at 18 dB/octave. So you need a huge Xmax to get any substantial undistorted bass output below the cutoff for a baffle that isn't several feet wide. See the Linkwitz LX521.4 for example. Take a look at the Carver Amazing speaker to see an open baffle design that actually works and makes some bass. Notice the large number of long throw woofers with added weight to produce a high Q resonance that helped passively EQ the 6dB / octave bass roll off due to the narrow baffle. Xmax is king. Sensitivity doesn't matter.
 
Last edited:
  • Like
Reactions: Juhazi
#13
olsond3 said:
Sensitivity doesn't matter.
Click to expand...
Again, I'm not taking side, but this is the contrary of what King wrote in his paper "... as the baffle size decreases, the support for the bass frequencies drops leading to a requirement that the woofer efficiency must be significantly higher than the 90dB goal function". I feel like I'm diving in a can of worms...

Juhazi said:
King doesn't discuss distortion at all, or low bass equalization. I'm not sure, but his measurements look like nearfield, they don't show dipole loss that happens by cancellation.
OB/dipole is difficult to handle, please study these pages too

http://www.dipolplus.de/
http://musicanddesign.speakerdesign.net/tech.html
http://www.quarter-wave.com/Project10/Project10.html
Click to expand...
Thanks a lot Juhazi, a lot to digest for a DIYer, I'll see what I can process.
In alternative, do you know any good and tested DIY kit/design for a 2-way open baffle system?
 
#14
^No, sorry.

There are too many of those, sharing same basic problem - transition from dipole bass-low mid to wild variations in directivity in upper mid and treble. This will make them very sensitive to positioning in the room and to room acoustics. An audiophile's dream speaker? It is rather easy to create good on-axis response with dsp, but that is not the whole story! Ability to play loud is another challenge.

Actually I have never listened to any of that kind of speakers. Never even wanted to make a test speaker myself! But this makes me thinking about making that kind of test, I will soon get 12" woofers and I have several 4-5" fullrange drivers in the closet! The problem is to make good measurements indoors during the winter.

There are hundreds of diy projects in the fixed thread here. Some are very well documented. I don't remember names, some other members might help!

You can find kits from Madisound, PartsExpress or in Europe from Visaton.

Search for "WAW" here at diyaudio.

https://www.reddit.com/r/diyaudio/comments/c5vh5w
 
#15
davirix said:
this is the contrary of what King wrote in his paper "... as the baffle size decreases, the support for the bass frequencies drops leading to a requirement that the woofer efficiency must be significantly higher than the 90dB goal function".
Click to expand...
This is so the speaker can meet a wanted sensitivity goal.

On the other hand if amplifier power is not a limitation, for example you design a system using line level EQ and will provide whatever amps are needed, then you've taken sensitivity off the table and it's no longer a limitation.

Maximum output is related to Xmax.
 
  • Like
Reactions: Juhazi
#16
....and greater power requirements = greater potential for thermal power distortion among other things, so historically, 'smart money' is to follow the pioneer's/MJK's dictum that 'there's no such thing as too much efficiency', so sensitivity per se is irrelevant since it's just the norm for converting power measurements to its efficiency. 😉
 
  • Like
Reactions: Arez and AllenB
#17
davirix said:
Again, I'm not taking side, but this is the contrary of what King wrote in his paper "... as the baffle size decreases, the support for the bass frequencies drops leading to a requirement that the woofer efficiency must be significantly higher than the 90dB goal function". I feel like I'm diving in a can of worms...
Click to expand...
King said "Figure 1 depicts a 90 dB/W/m goal function, tuned to 45 Hz, which will be assumed for this sample OB design." So that was just his arbitrary "goal" for the sample design. Maybe he had a low power amplifier in mind. I start with the maximum SPL I want the speaker to produce at the low frequency limit. This and the attenuation caused by the baffle size determines the required speaker displacement to produce that max SPL. Displacement is just the cone area multiplied by the cone excursion. So you can get displacement for a given driver by multiplying the driver area S times the Xmax. It it is less than the desired displacement you need larger area and/or larger Xmax drivers. That is how you engineer a speaker. When the drivers have been selected, you look at the required power to drive that driver to its Xmax and that determines the required power for the woofers. Efficiency doesn't enter into it until the last step. Amplifier power is essentially free these days, so efficiency does not matter. When you have separate amplifiers for woofers and tweeters with active crossovers it really doesn't matter. You will notice I didn't mention TS parameters either, as they don't matter when you have digital filtering available. So long as you have the required displacement, power to get there and EQ you can get an excellent flat response.
 
  • Like
Reactions: davirix, Juhazi and GM
#18
The thing King left out was how loud will that speaker play 45 hz with the woofer at Xmax. Because the example is a passive crossover design, matching the sensitivity of the drivers was of some interest. Assuming everyone wants a flat frequency response, no resonances and low distortion what separates speaker designs is the maximum undistorted SPL at the low frequency cutoff and the radiation pattern. Having selected a dipole pattern and max LF SPL, there is still wide or narrow dispersion for the mids and highs to consider. Cone diameter, wave guides and crossover frequencies will let you influence that. Lately I have enjoyed using low crossover frequencies to obtain very a wide dispersion pattern limited only by the baffle.
 
Last edited:
  • Like
Reactions: davirix, Juhazi and GM
#19
Thanks for the inputs, much appreciated. I've started reading the material above referenced, I've downloaded and tried the software programs (edge, basta!), but I am sort of overwhelmed by the level of knowledge required, and confused by the disagreement among what I may consider "experts".
At this point I am torn between going forward with this project despite all the doubts I have - hoping the community will be able to support step by step - or just buy a DIY kit for passive speakers, forget about parameters, curves, distortion and enjoy the music sooner.

D
 
  • Like
Reactions: Juhazi
You must log in or register to reply here.
Top Bottom