yes indeed - You are right
I forgot about that specil aspect of pseudo-infinite floor-to-ceiling array but we assume perfectly reflective ceiling and floor?
and a tweeter - not a supertweeter?
Last edited:
Narrow line is the best.
(340 m/s) / (20000 Hz) = 17mm = 0.67" << 2"
Two inches is the minimum possible distance between the centers of two drivers. That is three times longer than the sound wave at 20 kHz, which is the top frequency of the audio band.
Yes, because:
(340 m/s) / (20000 Hz) = 17mm = 0.67" << 2"
Two inches is the minimum possible distance between the centers of two drivers. That is three times longer than the sound wave at 20 kHz, which is the top frequency of the audio band.
How about using the othe 48 drivers facing backwards for BSC?
The Audience A3 drivers are used like this in the ClairAudient Line Source Array Loudspeakers: audience : loudspeakers
The Audience A3 drivers are used like this in the ClairAudient Line Source Array Loudspeakers: audience : loudspeakers
ok, that' s theory based upon some measurable problems and an assumption that they are audible
yet somewhat people enjoy arrays made of small fullrangers without tweeters, tapering, curving and so on
examples:
Advantages of the IDS Technology
Parts Express: Project Showcase
audience : loudspeakers
that's because sense of hearing in echoic conditions works differenty than a measurement microphone in anechoic/quasi-anechoic conditions
best regards,
graaf
Construction isn't a problem, I build staircases for a living, a curved staircase is more complex than a curved box.
anyway, I am afraid that sweet spot would be even smaller in case if such focused "10x10"
but it would be much cooler, that is for sure!
I will post some measurements of "4x4" with 16 four-inchers as soon as I find them
best regards!
graaf
But I like lots of chillies
I like Your attitude I must say
if a thing's worth doing at all it's worth doing right!
Two stereo speaker sets
That's great, because you can build four speakers and still have eight spare drivers left, if some of them get damaged in the process.I'd like to use all the drivers but I only have 8' ceilings
You could make a clone of the "sound bar" Home Theatre speaker (can't remember the manufacturer at the moment) that uses over 200 small speakers and DSP to generate steerable "beams" of sound and so provide "virtual surround".
Mount the drivers in a 5x40 horizontal array. Build an array of 200 amplifiers (!) so each driver is individually powered. Drive the amplifiers from a 200-output (!!) DSP processor. Write algorithms to generate any sound field you want, not just "virtual surround".
More realistically, you could build a couple of concentric arrays. (Multiple concentric rings of drivers around a tweeter.) Connect a crossover / delay line between each ring of drivers so that the radius of the radiating drivers progressively reduces as the frequency rises, plus have the outer rings delayed compared to the inner rings to form a virtual spherical wavefront. (Quad did something very similar with their ESL-63 electrostatic loudspeakers.)
Tweeter in the middle.
6 drivers in a ring around the tweeter.
Crossover at about 3 KHz from the ring to the tweeter.
Add a second ring around the first ring. Feed that from a LP filter, so it gets everything below 1.5 KHz or so.
Add a third ring around the second. Feed it everything below about 1 KHz.
Add a fourth ring. Feed it everything below about 800 Hz.
Add a fifth ring. Feed it everything below about 650 Hz.
That's 96 drivers, in a hexagonal array some 20 inches across.
Shutting up now...
Mount the drivers in a 5x40 horizontal array. Build an array of 200 amplifiers (!) so each driver is individually powered. Drive the amplifiers from a 200-output (!!) DSP processor. Write algorithms to generate any sound field you want, not just "virtual surround".
More realistically, you could build a couple of concentric arrays. (Multiple concentric rings of drivers around a tweeter.) Connect a crossover / delay line between each ring of drivers so that the radius of the radiating drivers progressively reduces as the frequency rises, plus have the outer rings delayed compared to the inner rings to form a virtual spherical wavefront. (Quad did something very similar with their ESL-63 electrostatic loudspeakers.)
Tweeter in the middle.
6 drivers in a ring around the tweeter.
Crossover at about 3 KHz from the ring to the tweeter.
Add a second ring around the first ring. Feed that from a LP filter, so it gets everything below 1.5 KHz or so.
Add a third ring around the second. Feed it everything below about 1 KHz.
Add a fourth ring. Feed it everything below about 800 Hz.
Add a fifth ring. Feed it everything below about 650 Hz.
That's 96 drivers, in a hexagonal array some 20 inches across.
Shutting up now...
Sounds like a very good idea to me!
But probably the crossover between the tweeter and the inner ring should be somewhere higher - at least 6-8 khz or maybe 10-12 - let's take full advantage of the full ranges...
And the lower crossover points of the other three rings can be so adjusted that the system to get around 30 HZ at some decent level -3 or -6 db...
I figured out that the total cone area of 100 drivers is around 1060 cm3 which equals the cone area of a driver with diameter around 36.7 cm (14.44 in) and note that this is effective area and effective diameter. 36.7 cm effective diameter will have a driver with about 40+ cm total diameter which is around 16 inches...
If the xmaxX2 (xmax-2.5mm) is 5mm then the maximum displacement will be around 500cm3 (30.5 cu. in.) which is about twice than the capabilities of a decent 12 inch bass driver... And in turn that displacement gives the opportunity of reproducing 30 hz with around 93-96 db-SPL (74-75 phon)
And that is great!
Even if the xmax is 1.25mm (2.5mm total membrane travel) we still will have
250 cm3 total displacement
- enough for 90 db-SPL (70 phon) at 30hz - another question is what will be the power in the low frequencies needed to achieve that, but a decent equalizer will handle that Even if the mean system sensitivity is 96 db you'd still have the ability to achieve 30 hz at -6 db
Another issue is the impedance of the speakers - I'd sacrifice the linearity for power and go for 2-3 ohm in the 0-50 hz region... and a powerful SS amp.
Best regards!
You have to be a bit careful where you put the crossover frequencies.
If any of the rings are fed frequencies that correspond to less than about 1/4 to 1/2 wavelength compared to the ring diameter, you will get comb filtering. Also, as the frequency rises, the directivity pattern of any given ring will narrow. The frequencies I gave above should ensure each ring cuts off before it starts to beam excessively. To use a different frequency set, pick a frequency that corresponds to a desired wavelength for the inner diameter and work outwards from there. Ring diameters are 4, 8, 12, 16 and 20 inches. The frequencies will be the reciprocal of the diameters. If you cross the inner ring to the tweeter at 6 KHz, then the frequencies will be about 6 KHz, 3 KHz, 2 KHz, 1.5 KZ and 1.2 KHz.
The tricky bit comes in driving the beast. If you run all the drivers fullrange, they'll handle at least 500 watts per channel. If you try to drive the rings separately, the range of frequencies involved imply an active crossover / multi channel amp system.
Each ring has a multiple of 6 drivers. Each ring will need its own channel. Each channel will need one or more amp modules. Assuming 8 ohm drivers wired in series-parallel to provide 12 ohms per set of 6, and amp modules capable of driving 4 ohm loads, 7 stereo amp modules would be required for the stereo pair. (Ring 1 = 12 ohms, ring 2 = 6 ohms, ring 3 = 4 ohms, ring 4 = 2x 6 ohms, ring 5 = 2x 4 ohms.)
At 5 watts per driver, and assuming 12 ohm wiring per 6 string, amp modules rated at 90 watts into 4 ohms will be required.
There is one remaining problem: only the inner rings handle high frequencies, but since there are fewer drivers per ring, they will each have to work harder to keep the frequency response even. For example, the first ring has 6 drivers and the second ring has 12. The drive to the first ring will have to be boosted by about 9 dB for frequencies above the LP cutoff of the second ring.
It's all starting to look rather messy...
If any of the rings are fed frequencies that correspond to less than about 1/4 to 1/2 wavelength compared to the ring diameter, you will get comb filtering. Also, as the frequency rises, the directivity pattern of any given ring will narrow. The frequencies I gave above should ensure each ring cuts off before it starts to beam excessively. To use a different frequency set, pick a frequency that corresponds to a desired wavelength for the inner diameter and work outwards from there. Ring diameters are 4, 8, 12, 16 and 20 inches. The frequencies will be the reciprocal of the diameters. If you cross the inner ring to the tweeter at 6 KHz, then the frequencies will be about 6 KHz, 3 KHz, 2 KHz, 1.5 KZ and 1.2 KHz.
The tricky bit comes in driving the beast. If you run all the drivers fullrange, they'll handle at least 500 watts per channel. If you try to drive the rings separately, the range of frequencies involved imply an active crossover / multi channel amp system.
Each ring has a multiple of 6 drivers. Each ring will need its own channel. Each channel will need one or more amp modules. Assuming 8 ohm drivers wired in series-parallel to provide 12 ohms per set of 6, and amp modules capable of driving 4 ohm loads, 7 stereo amp modules would be required for the stereo pair. (Ring 1 = 12 ohms, ring 2 = 6 ohms, ring 3 = 4 ohms, ring 4 = 2x 6 ohms, ring 5 = 2x 4 ohms.)
At 5 watts per driver, and assuming 12 ohm wiring per 6 string, amp modules rated at 90 watts into 4 ohms will be required.
There is one remaining problem: only the inner rings handle high frequencies, but since there are fewer drivers per ring, they will each have to work harder to keep the frequency response even. For example, the first ring has 6 drivers and the second ring has 12. The drive to the first ring will have to be boosted by about 9 dB for frequencies above the LP cutoff of the second ring.
It's all starting to look rather messy...
I'd like to use all the drivers but I only have 8' ceilings
With 200 buy out drivers, i'd suspect a few would be enuff different than the others to be discarded. The spec says 2.25" bezel -- 8' leaves room for 42 drivers.
You could make a wide shallow bipole. 42 drivers on the back & 42 on the front = 84/side Or angle the back, and put 2 splayed arrays of up to 29 drivers.
Actual numbers in easch line will have to consider ways of wiring them and ending up with reasonable impedance.
dave
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.