3-Way Open Baffle Straight Line Array (Floor to Ceiling)

Moving on from a sealed fullrange line array made with SB65WBAC25-4 times 36 floor to ceiling is my next project. I like the current line array system, but I also like my 3-way open baffle and horn system so what to do? Full length 3-way open baffle of course!

- Single driver from 400hz to above 7khz. (I think) one the things I find pleasing from the horn and existing OB system is radiation pattern control in that range. See Geddes. Also Linkwitz took his system from about 1khz to 7khz with a single 4" driver with a focus on radiation pattern. I'd like to push that down a bit with the Line Array application.

- I like the sound of Open Baffle bass! In my room it works and I don't have space for multi subs. I've had 4 sealed sources in my room and it works to smooth out low frequency modes; however, I don't like the room pressurization with sealed subs. Something about the presentation is slightly different. Since I don't listen loud and prefer low volume listening I've found the presentation of OB from 30hz to 500hz more pleasing.

- Floor to ceiling straight array. Trying to follow Jim Griffins white paper but will not be implementing shading at this time.

- Minimal baffle width. Trying to follow Linkwitz and John K. philosophies for baffle width. Trying to stay below the baffle peak in all drivers is difficult. Some compromise are in order of course.

- One thing I wanted to try is a space between the midrange and woofer line in hopes to get some horizontal OB control with minimal baffle. Not sure how to perform accurate measurements for this in room. Would be nice to get some side cancellation so the midrange line doesn't "see" the woofer baffle as support. Not sure this will be feasible. According to Jim Griffin white paper spacing should be less than 1/2 a wavelength. I think we can get close to 1/4 wavelength at crossover but I'm oscillating on the woofer array right now.

Tweeter Array:
Dayton PTMini-6. 24 high.
Since they are sealed will have to make the array with front and rear facing units. That doubles the cost of the tweeter array so we will make a front firing array first and see about the full monty later. The nice thing about these drivers and this application is the driver width = 23mm. So a full dipole line array could be constructed in a 1" by 1.5" package including all wiring. That is mechanically very nice. This also allows a type of plug and play tweeter array to use with my existing LA for sport. Also in theory the radiation pattern should be nice, although since we are using side mounted midrange we might not be able to take full advantage of the radiation pattern :(.

Construction will be done using 1/2" baltic birch with tweeter mounted using double sided sticky tape from home depot. This worked out well on the test baffle. Simply drill holes for the wires/quick connects and go one your way. Allows for nice spacing with minimal effort. Also one can pre-assemble the wiring harness for 6 in parallel and 4 in series. Making roughly a 4 ohm load. Can gain a bit of sensitivity that way. Construction will not be as excruciating this way and might get some damping due to the tape soft mount. Ceiling mount so they don't fall over.

Midrange Array:
NEO8-PDR times 10.
I know not really exciting here but it really is one of the best choices for this application. Will have to go with 2 in parallel and 5 series. This looses some sensitivity but no choice here. Cannot fit 12 in my room.

Construction will be done with 1/2" by 1/2" aluminum square bars. This allows for minimal baffle and hopefully enough strength to stay somewhat straight! Foam damping behind the driver and routed out depressions for the NEO8 rivets in the back. This will be a lot of work for me but not too bad. Probably need a hole template. Ceiling mount so they don't fall over.

Woofer Array:
Would be nice to stay with 1/4 wavelength at XOVER and use 10" drivers times 8. That said I haven't really liked the any 10" OB drivers I liked for bottom support. Thinking 15PR400 would be something but no minimal baffle then and CTC spacing is a bit much. Some thinking to do........taking suggestions. I like the low QTS professional woofers with EQ better than the high QTS drivers I've tried. That said the GW-1858 was nice but missing the clarity of the 15TBX100 I used. I was using the GW up to 500hz so in this application the lower QTS sounded better. If low bass under 100hz, probably no difference. Since we are looking at 350hz to 400hz XOVER we should be careful about this or perhaps some major disappointment and expense.

More to come later. Tweeter and Midrange drivers on order!
You find the sb65 lacking in the treble??

I guess I'll have to fess up. Yes mostly and a bit no. You can eq the extreme treble to satisfy (so no) but it ends up being a very non-efficient speaker because you don't get the same gains in the treble as the bass.

I like the treble presentation better with some of my other speakers. Perhaps the LA form is more picky about absolute eq but I have trouble with long term satisfaction of the SB drivers in a full range line array application and I can't stop fiddling with treble levels. I don't do this with a number of other designs. That said the midbass is pretty smooth, better than a large number of my other speakers. There is an ease about it that is not easily matched.

Perhaps I should try Jim Griffins suggested shading or implementing a CBT but the one thing I wanted with that build was to use them fullrange to 30hz and no crossover. This goal was met but there are other issues that require solving for my long-term satisfaction. Not sure how much further I want to go with that build. I like the cabs and it was a lot of work. I don't want to change the enclosure and I don't have facility to do the shading properly at this time.
Thanks. Excited to see your project come along. My ob line array should have somewhat similar polars as yours above 500hz although mine won't be so wide at the very top treble. But in general, wider dispersion in the treble, narrowing up as you get lower, im very curious how this power response will sound since it is the opposite of most speakers. My hunch is that the very high perceived resolution of horns will be missing because the response will have to be reduced in the treble to counteract the rising power response in the treble.
...But in general, wider dispersion in the treble, narrowing up as you get lower, im very curious how this power response will sound since it is the opposite of most speakers....

Yes, that is actually one of the items I was going for. It is exactly opposite to what I'm running now. Will be a neat change. I'll give some subjective prose and rudimentary frequency measurements when they are up and running. Hard to measure fixed dipole arrays in room with any sort of accuracy but some will be better than non!
Some pics. Box full of drivers, aluminum square rods and some 1/2" baltic birch. Will try to start tweeter array first.


  • IMG_20191230_194141.jpg
    196.5 KB · Views: 788
  • IMG_20191230_194547.jpg
    138 KB · Views: 786
A good day! All the midrange and tweeter drivers have arrived. Time for some busy work. Will try to assemble the Tweeter arrays first. Wiring harness and maybe drill template to come.


  • IMG_20200108_184229.jpg
    361.2 KB · Views: 711
  • IMG_20200108_184239.jpg
    152.5 KB · Views: 704
If you can live with just 9, you can connect them sets of 3 in series, and then the three sets in parallel. Same impedance as original, greater sensitivity.

Wow, this did not even register with me until you mentioned it :). Might have to think about how to best experiment with both. With 9 you get just over 70 height coverage with 10 you get just over 80% (8 foot ceilings). Not sure I'll need the sensitivity. I do like the +80% coverage idea but will have to think on it more, thanks!
Finally got some time on these!! As per first post, 1" piece of wood cut. Drilled holes for wires. Bent the leads 90 degrees and used strong sticky tape for driver mounting. All in all pretty straight forward.

6 drivers in parallel and 4 sets in series. Should be nominal 4 Ohms, measured 4.1 ohms.


  • IMG_20200222_155253.jpg
    49.8 KB · Views: 627
  • IMG_20200222_155308.jpg
    54.6 KB · Views: 260
Tested 2 Tweeter arrays with my existing fullrange driver line array. The tweeter array adds what was missing. Very different presentation in the treble. The wide dispersion changes/makes the sounds easier to listen too but also with more treble detail. Sounds better when standing up too.

Now for the midrange panels!

P.S. According to distortion sweeps these are good to about 4.5kHz.


  • IMG_20200223_133840.jpg
    160 KB · Views: 424
Last edited:
Progress is slow. Finished drilling the half inch aluminum brackets for the midrange sections.

After measuring out the holes, looks like 12 drivers (NEO8-PDR) per side can be accommodated. Currently I only have enough for 10 per side so I'll first make up the wiring harness for 10, 2 in parallel and 5 in series for around 9 Ohms. The 12 config sets up better for 3 in parallel and 4 in series for around 4.8 ohms. Some duplicate work but it will be a while till I can acquire 4 more drivers.

In the spirit of minimal baffle drills were made half way though the brackets to all for mounting of the NEO8 with rivets. Pics show a bracket and the cleared out sections for the rivets before 1/8" foam installed. I did not want to front mount the drivers as that would increase baffle width. Would it matter in the end? Probably not but drilling a few more sections was easier than using 1/8 inch strips with backing material. Would have be close to the same number of drills so decided on the cleaner look.

Driver center to center spacing is the 7.875" driver length plus 1/16". Tried to get them closer on a test baffle but the frames eventually push out the mounting hole spacing a bit. If only doing 4 drivers it might be doable but more than that require some extra spacing.

Fingers crossed on the driver mounting. Must sort out ceiling mount before installing the drivers.


  • IMG_20200314_111042.jpg
    49.1 KB · Views: 218
  • IMG_20200314_111054.jpg
    40 KB · Views: 366
1 Midrange assembled. 1/8" board on front, 1/16" foam on back. 4-40 screws with nylock nuts. Feels good in hand, hopefully not to much resonance when the last of the drivers are installed. Might install some dummy wood for stiffness.


  • IMG_20200314_223454.jpg
    100 KB · Views: 194
  • IMG_20200314_223504.jpg
    137.1 KB · Views: 222
  • IMG_20200314_223512.jpg
    132.4 KB · Views: 242
  • IMG_20200314_223520.jpg
    135.9 KB · Views: 216
Tested 2 Tweeter arrays with my existing fullrange driver line array. The tweeter array adds what was missing. Very different presentation in the treble. The wide dispersion changes/makes the sounds easier to listen too but also with more treble detail. Sounds better when standing up too.

Now for the midrange panels!

P.S. According to distortion sweeps these are good to about 4.5kHz.

Have you thought about trying them like this? I appreciate that your SB drivers are pretty small but it would be interesting to see what it does to the full range line having the drivers in front.


Have you measured any of the setups yet?


  • ribboninfront.jpg
    69 KB · Views: 931
Have you thought about trying them like this?

Is that your build too fluid? Looks great!

I did not think about mounting them like the pic as the Xover was quite high and the SB line array had different goals. This was a quick add on.

That said maybe its worth a quick investigation as with the offset tweeter array and high Xover there should be problems with off axis and other measurements. Even though I know there are problems I do like the presentation better than with just the SB drivers. I did a quick tweeter integration with the SB array so measurements were short and I was eager to hear the results. So much so I've just left them up and been happy to not investigate their flaws. If time permits I might bet back to them and do some better sweeps but got the bug to do the OB LA project. With the OB project there was not way of mounting the arrays like that, with the SB sealed array that would make more sense.