Large scale open baffle

I'm working on the use of a large-scale dipole to use as a "music speaker system" in an installation where there is a center-mounted "speech-optimized" loudspeaker.
The point to this question is: Has anyone in the DIY community ever tried anything like this on the scale of an auditorium seating about 600 people?
The application is in combination with the existing center speaker as part of a left/center/right idea, where the left/right are either side of the dipole...
Some of you may have figured out what I'm up to - it's a case where I'm doing the job at cost to help out the congregation. We have tested a prototype, on a smaller scale, on the platform, and the ability to cover what was a difficult coverage seating plane is SPOOKY!
It should be noted that left/right imaging is not being sought, here - just a sense of "difference" in the listeners interaural sensate observation.
I'm hoping other "pro" consultants with "enthusiast" leanings might be willing to share thoughts, in the "Synergetic Tradition".
 
We are receiving the drivers, as I write. The whole lot is apparently to be in two shipments. We have a few procedural points to observe: 1. Each driver will be run through the LMS "rub and buzz" to assure they're viable. 2. Each piece will get its own documented impedance curve in free air, and on a baffle designed to avoid exact multiples of dimension in any direction. 3. Each device will be documented for frequency response. Then, 4. We will start grouping the drivers on various baffles to observe how that "figure-of-eight" pattern "lands in the seating...
We are testing with the baffle(s) on the platform, to simulate the boundary behavior of the ceiling. As the testing continues, microphones and other sources will be introduced and the effects of the polar behavior on things like intelligibility and feedback noted - also note: The existing center-mounted speaker is speech optimized and the dipole is intended for music playback and some other hoped-for effects.

More to come...
Cheers
 
A side note: This whole experiment is a result of the idea that appearance trumps acoustic merit.
An very wide seating plane resulted from the length/width ratio created by a "value engineering" effort on the part of the major investor in the construction project, combined with ignoring what a cube produces in modal behavior. One would have required an "exploded array" to cover the seats. The other lead to some really excessive corner-terminated low-mid frequency modes with almost no glass to help leak it out - even though part of the standing waves were moved around a bit. The modes will have to be tolerated. The seating was dealt with by the design of a custom 130X52-degree horn driven by a Renkus CDT-1, with a pair of 10" Emminence woofers in a D'Appolito-style box. The result was a horizontal directivity match across both crossover points. Vertical pattern behavior was maintained down to about 300 Hz. The whole thing is tri-amped and processed by a Behringer DCX-2496. I've seen commentary on the last item in this forum. I, personally am not impressed by the user interface - while it does appear to work quietly...
Add to all that, the dipole, and it becomes another example of what happens when someone says: "We can deal with THAT problem..."
We shall see...

Stay tuned...
 
For those that might still be watching this thread, the curve off the back of a Galaxy S5N8 starts to dive at about 4k.
Also, as many have noted, shooting polars on an open baffle is almost useless - after about 20 degrees rotation that back-side energy starts to sum with the front - with the inevitable phase issues added to the "mix". Listening to a single driver along the "edge-axis" of the panel still sounds intriguing.
I'd like to know more about how the voice coils are being wound. It almost looks like a turn-counter is used. Whatever comes up for a DCR and the resulting Free Air Resonance seems to wander a bit, from driver to driver... The sound is nice, and averaging will help.
Stay tuned...
 

Sangram

Moderator
Paid Member
2002-09-25 11:01 am
India
Though this is well beyond my technical ability, I do have a small line array mounted on an OB. My observation is that best results are only achieved when the back and front have similar frequency response, which is very problematic at high frequencies primarily because of dynamic drivers' physical design. I did manage to counter some of the issues with a rear-firing tweeter but that only fixed it for one seating position (mine).

There are many advantages I see from this arrangement such as very high usable sensitivity and lowering of distortion - and averaging of drivers - specially cheap ones, but the polar response (specially vertical in my case) is still less than desirable.

I will be watching your thread keenly. Eager to learn more about what you finally end up with.
 
Turk, I own a Driverack PA. The client insisted on the DCX. My brother has done a fair bit of contracting using various versions of the Driverack to good effect. Even the client has admitted the next processor will be a different item!

Sangram, it is likely that the way we are doing this project, we may go the other direction.
I'll try to get some diagrams up so you can see what we are doing.
We are lining up two rows of drivers, one in front of the other, one set facing one way, the other in the opposite direction.
What it appears to be getting us is additional energy below 4k based on the rear-firing driver's output, added to the front firing drivers on-axis energy. So far the judgements on the listening quality are empirically based - when it gets ugly, we sit down and figure out why, supplementing our decision making with measurement data.
We are not under deadline, so have the luxury of taking our time. In addition, we are waiting for delivery of the rest of the drivers!
Part of the idea going forward might be to try and shift the effect of the so-called Baffle Step by selctively damping the backside output to make the rear firing energy sum with the front of the other row to mitigate the low (relatively low) frequency rolloff. Polars will follow to see what the ramifications are. Remember, the angle of observation will always on both sides of the figure-of-eight radiating pattern!
The function of this panel is to serve as a difference source, using surrounding walls to reflect acoustic energy out to the congregation. The effect on the platform for the performers is a factor we'll start to study - how it will interact with the monitors - or even replace their function, is up for research...
It's getting to be a lot of FUN!

Stay tuned...
 
I am working with an existing system with three different Behringer models, at varying output power levels.
I haven't had problems, other than knobs not corresponding with similar positions on the same model, etc. Since I set the gain structure using measurement that isn't a big deal. The more aggravating issue is the DIP switches used to configure the input/bridging/pairing schemes. Those switches need a magnifying glass to be seen and read accurately!
If the client had Crown or other brands in that class, I'd be just as happy.

They say you can't "hear" an amplifier. What I'm hearing with these things doesn't impress me.
Caveat Emptor...
 
Progress Report:
The drivers are now being selected for impedance curve matches/resonant frequency similarities.

Here is a question for the group: What about using a Zobel? The array is not going to employ any other driver. It's low frequency augmentation will come from the "Mid" speaker (which also serves as the "Speech" source).
A Zobel when there is nothing but the correction of the impedance tail?
Let's start what could be an argument: Will there be an audible result?
Testing begins next week to determine the aggregate polar behavior - comparing the use of one row of drivers all facing the same direction, vs. a second row along side and behind facing the opposite, providng better high-frequency dispersion - while still being out of polarity.
Bring on the experts on Zobels!
 
An array panel was raised today. It was hung about 16 feet above floor level. The mic was moved from center aisle to outer seats along the front row, on either side to sample how the "front side" measured compared to the "back side" in corresponding seats. Tomorrow we do some simple EQ to make the curve easier to track, from inner to outer seats.
This is different than doing a polar response, as the distance from seat-to-louspeaker is going to change. We are looking at the change in level, along with the roll off in high frequency response as we move from seat to seat.
What we saw this evening was encouraging enough to proceed with the double-row panel, based on the rolled-off high frequency response from the back side of the array.
We will also bring "test" music to the venue and "play" the panel in Difference Mode (basically sideways) to see what reflections from the platform/stage walls do to whatever imaging might result.
The next step, probably next week, will be to finalize which version of the two-row panel we like best and start cutting plywood and stretching grille cloth.
There is evidence of a very narrow "null" down the center aisle, but the seats next to the aisle have surprisingly listenable response curve,ergo, time to listen to music...

More to come!
 
I just came from the venue, where I spent most of the afternoon. Currently we have been working with a piece of OSB with five drivers on it. We hung it in the approximate location we intend to put it, to allow us to see how the acoustical environment affects the imaging - which heretofore we were uncertain about - and to see what the front/back differences were like out in the main space.
If you look at the Galaxy S5N8 5 1/4" driver on the Parts Express or other webpages you can see the likely impediment for rearward energy radiation. Add to that the hole in the 3/4" panel surrounding the rear frame and you end up with a ring radiator. Our documentation of the resulting behavior has been pretty casual - we went out into the seats and "walked the space."
We found, by just listening that the sphere of energy thrown from the front of the driver sounded different than that thrown to the rear.
But, not much.
There was enough high frequency in the rear "balloon" that the expected Difference information was discernible. When the "mid" speaker, the existing center mounted speech-optimized device, was added with some care in balancing the levels - the spooky clearly directionally noticeable information began to form out in the space. This "cloud" of sound seemed to form around and above, as well as behind the listener.
This is unlike the image you get from loudspeakers placed ahead and above the seating plane in the traditional Left-Center-Right kind of arraying often used. Amorphous isthe best description i can come up with at this point. Once I get a more finished and representative device built I will work on posting some pictures. Today was fun and exciting, but we have a ways to go and there are some corners in the route to completion. I like to look around them and make sure we're still going in the overall direction we intended!

More information follows...
 
The problem with this kind of project is that it is inevitable that there are protracted (!) periods of observation required while each contributor drags in their reference material, in order to assure all the participants that a consensus is being developed as each step is taken.
The bigger problem is that the process is getting to be more fun than efficient!
This week I am documenting here that the goal is to test the double row panel and get it and the respective drivers into the shop for final construction while the issue of low frequency up to about 250Hz is addressed - preferably in a dipole whose directivity will "knit" with the panel in each opposing sphere.
We experimented with using the low frequncy potential in the center speaker to "fatten things up" a bit - without the result for which I was looking. We could have driven the center woofers a bit harder, compromising the intelligibility goals toward which the design of the center was being driven.
A question to the group:
I'm thinking about four OB 15" woofers mounted"face-to-face" in a pair on either side of the respective panel, driven so-called Isobarik -style. While the budget is geared to meeting needs, wants will be entertained and the usual "value-engineering applied, at least to the point where the Designer steps in, and reminds the folks why we're pursuing this project less than two years after the original installation!
Suggestions are welcome, along with the accompanying engineering justifications, respectively! :)
Pictures are being taken and will follow once the issues "in the air" have been further addressed.
Your patience is appreciated, mine is getting frazzeled by the desire to hear how it all comes out!

Cheers!
 
The latest:
We ran a 100 Hz test tone through the face-to-face Eminence Gamma 15's we're using as the low mid component in this array. The modal behavior of the venue is bizarre! Every six-to-eight feet you walk on the sanctuary floor there is a major shift in the perceived level of the tone...
We did, however, note that, when playing music through the woofers the modal pattern is not so obvious.
The ten-driver panel was raised to the position where it will finally hang. There will be some separation from the woofer set-up which will be closer to the ceiling.
The face-to-face array of woofers is also operating in "push-pull" dipole mode. We do have some options to move the woofers up and down - some.
Because of the scheduling conflicts of two people working on this project there are very narrow time windows in which the work is being done. Still, it proceeds.
Stay tuned - the pictures are being taken as we go...

Cheers!
 

turk 182

Member
2012-10-26 3:03 pm
LAMgeezer
being that your dipole is going to behave omni like your modal behavior could simply be the product of distance to boundary reflections a simple change of position and running the same tone during your next window of opportunity would confirm that.
i guess i still have to be patient for pic's...
 
We rotated the woofer array and moved it around the platform a bit to be sure. Many of the dimensions of the space are either identical or exact multiples of each other. Tuning and regulating the organ pedal division is kind of a bizarre exercise, as you might imagine.
There is one additional feature - all the glass (what there is), resonates far above the frequencies where it might "leak" and provide some relief.

Oh well...

Turk, hold on!

Cheers