Can you use a cluster of smaller drivers for mids on an open baffle?

[pk]

Hi,

In order to achieve high sensitivity, I have thought about using a ”cluster” of smaller drivers for mids on a large open baffle mounted in a square (e.g., 2 x 2 eight inchers [e.g., the JA8008], or maybe even 4 x 4 three inch drivers [e.g., Faital Pro 3fe20]). The idea is to have super-fast, articulate, and clear mids to match the Beyma TPL-150H I use as tweeter, that is, ideally the mids should work from 100-1500 Hz (reduced extension into the bass might do).

Is it just a stupid idea, or? Please enlighten me!

PS: The idea is somewhat inspired by the Megatrend MkIIs from Transmission Audio, although the mids are used all the down down there…

Thanks!

Best regards
Peter

[sreten]

Hi,

In terms of senstivity it will work, but will also give lobing issues.

The physics is simple, an 8" all things being equal = a 2x2 4" array.
However the 8" is more likely to have better controlled directivity.

One quality driver usually beats multiple lesser quality drivers.

All depends on the application and cost of the drivers.

rgds, sreten.

[Patrick Bateman]

Originally Posted by pk

Hi,

In order to achieve high sensitivity, I have thought about using a ”cluster” of smaller drivers for mids on a large open baffle mounted in a square (e.g., 2 x 2 eight inchers [e.g., the JA8008], or maybe even 4 x 4 three inch drivers [e.g., Faital Pro 3fe20]). The idea is to have super-fast, articulate, and clear mids to match the Beyma TPL-150H I use as tweeter, that is, ideally the mids should work from 100-1500 Hz (reduced extension into the bass might do).

Is it just a stupid idea, or? Please enlighten me!

PS: The idea is somewhat inspired by the Megatrend MkIIs from Transmission Audio, although the mids are used all the down down there…

Thanks!

Best regards
Peter

Of course! Efficiency and bandwidth are related, so the use of multiple small woofers is one way to increase the efficiency of a loudspeaker.

Keep in mind that there's no correlation between 'speed' and size of a transducer. Check out the paper from Dan Wiggins for the best debunking of that tired old myth.

[Rudolf]

Originally Posted by sreten

Hi,
but will also give lobing issues.

Just to show the magnitude of lobing issues, this is a 3x3 array of 3" drivers on a 26x26 cm baffle. 1 kHz would be the upper limit imho:

[Patrick Bateman]

Arrays are a lot of work, but the rewards are sometimes worth it if you have the time to do the engineering. For instance, in a 3 x 3 array you can filter the eight drivers on the outside and run the center driver full range, and basically turn the nine drivers into an approximation of a full range. It's the same idea as the old quad loudspeakers, but using cone drivers instead of planars.

And this concept is 'fractal' in nature. For instance, instead of doing 3x3 you could also do 5x5. Dr Smith from Snell is on this forum, and he's discussed this in a few threads.

The Unity horn also works on the same 'fractal' nature, that you can do a 2-way, 3-way, 4-way, etc... I think the Unity horn is a superior solution than the Quad because it adds directivity and improves phase response to the same concepts originated in the Quad full range. When you start messing with these arrays you quickly realize that phase response and directivity go to hell quickly.... This is where the Unity solves a ton of those issues.

Check out the array tool from FRD consortium to simulate these properly. This is a big project BTW, lot of work.

[pk]

Originally Posted by Rudolf

Just to show the magnitude of lobing issues, this is a 3x3 array of 3" drivers on a 26x26 cm baffle. 1 kHz would be the upper limit imho:

Hi Rudolf,

Thanks for responding. Since I was thinking aboit crossing to the Beyma TPL-150H around 1500 Hz, it sound somewhat sensible - at least in principle.

May I ask: How did the 3 x 3 array sound compared to, saym using on large (e.g. 8" or 10") dedicated mid?

Thanks!

Best regards
Peter

[Rudolf]

Originally Posted by pk

May I ask: How did the 3 x 3 array sound compared to, saym using on large (e.g. 8" or 10") dedicated mid?

Sorry Peter,
I failed to mention that this is a Boxsim simulation of some Visaton FRS8 drivers. Personally I would never build wide arrays - only vertical lines. In above array I find the change of power response (beaming) with rising frequency quite horrible.

[Patrick Bateman]

Originally Posted by Rudolf

Sorry Peter,
I failed to mention that this is a Boxsim simulation of some Visaton FRS8 drivers. Personally I would never build wide arrays - only vertical lines. In above array I find the change of power response (beaming) with rising frequency quite horrible.

If you address the array as if it was an approximation of a point source, beaming becomes your friend because you can use it to control directivity. It's a bit of work, but it's posssible. This is the main advantage of a 3x3, 5x5, or 7x7 array.

For instance, in a 5x5 you can do the following:

Center speaker plays 100 to 20,000 hz;
Middle ring (eight speakers) plays 100 to 4500 hz;
Outer ring (sixteen speakers) plays 100 to 1500 hz.

The main ideas are this:

• We are approximating one full range loudspeaker, beginning in the center of the ring.
• As we start in the center of the ring we are full range, and as we get closer to the edges each ring is playing lower
• The interference pattern which is created by the overlap narrows the directivity

It's some serious business, but a number of speakers have done this to good effect. I tried doing it a few years ago, but there are a few drawbacks:

• It's hideously complex, and sometimes "keep it simple stupid" is better
• IMHO, waveguides are a simpler and more effective solution. But they are also much MUCH larger - a 15" waveguide only goes down to 900hz
• To really do this right, you need very specific crossovers and frequency shaping. Doing it with a passive crossover network is no ideal.

JBL has some papers on this - written by Ureda if memory serves



If you want to hear what this sounds like, just go listen to one of those "sound bars" from Yamaha that uses a pile of cheap drivers to approximate a five channel system. They work on the same principles - they use DSP acoustic filters and time delay to control directivity. Don't listen to the super-cheap ones from competitors - 80% of the sound bars don't use DSP and the proper filters, but there are a few that do.



I think the most promising solution is to use waveguides AND an array. Above is a real world measurement of a loudspeaker that does this, from Snell. Note that the directivity is darn near perfect from 100 to 15000hz, except for a blip in the midrange. It's designer is a regular contributor to the forum, and posts on an almost daily basis. If you want expertise on how to do this properly, this is the place.

The use of waveguides AND an array gets you the best of both worlds. Waveguides are an elegant tool to control directivity, but the lower they go, the bigger they get. Arrays do not do it as well IMHO, but as frequencies get lower and lower it's often better to have a solution that is inexpensive, compact, and attractive. And arrays are superior to waveguides in the realm of size, appearance, and cost in most circumstances.

Link: D'Appolito Arrays with Waveguides

If you start to play around with array simulators, you'll notice that you MUST address directivity to get this to work properly. I know it's fun to buy a box of drivers and wire them up with no consideration for directivity, but the results are far from optimum until directivity is considered. You get massive MASSIVE dynamics but the in-room sound is sub-standard until directivity is tackled. Once directivity is addressed, the pieces of the puzzle fall into place and you get a solution which basically sounds like a really good full range speaker.

[Rudolf]

Agreed, Patrick

the way you tackle the array is far more advanced than what I was thinking about.
Complete accordance in directivity as the main issue to look at.

Rudolf

[Patrick Bateman]

Originally Posted by Rudolf

Agreed, Patrick

the way you tackle the array is far more advanced than what I was thinking about.
Complete accordance in directivity as the main issue to look at.

Rudolf

It's a neat trick, definitely, but a lot of work!
I'd love to see a DIY'er give it a shot. We have the tools to do it these days, and all the documentation is readily available, just need to get someone to try it.
The sound of the array is completely different than a a conventional array. Most of the HiFi arrays that I've heard create a presentation that's larger than life, but ultimately fatiguing because it's not realistic, and the imaging is "fake." (The array basically makes EVERY source sound huge and bloated.) For instance, if you play a recording of an orchestra, the orchestra is the size of the wall in your room. But then you play a soloist, and the soloist is as big as the orchestra.

An array with directivity simply sounds like a good full range loudspeaker, but with better dynamics and power handling. (Since the array is simply approximating a single full range driver.)