My Line Array Project Log

[zobsky]

Hello,

I have decided to create this log to document the design and construction of a pair (or maybe triplet) of line arrays.

I fully expect this project to crawl along depending on how much spare time I have, so please be patient.

The woofer/mid will be madisound foster 5.25". I evaluated by ear, in addition to this foster, a closeout 4" audax aerogel woofer from madisound and the hi-vi A3S from parts express.

The audax was the best sounding, followed by the foster. Sadly, the audax is no longer available. The hi-vis were far too inefficient to interest me, even arrayed. Hence, I'll be using the foster.

The tweeter will be apexjr's cheap tweeter, apparantly audax/harmon surplus.


Picture of the drivers:

[zobsky]

Driver Specs:

Woofer

Foster 2422 264 00504
Paper cone, cloth surround wide band woofer.

8 ohm
87.3dB
Fs 80.7Hz
Vas 10 liters
Qms 4.267
Qes 1.375
Qts 1.040
BL 3.57
Cms 808.7 uM/N
Mms 4.81g
Re 7.2 ohm
Krm 1.155
Kxm 6.832
Erm 0.818
Exm 0.716
Sd 0.0094 sqM
Le 0.57mH @ 1kHz

Probably best used free air.
25W power handling



Tweeter:

Dave(Planet 10) posted specs on this tweeter. They can be found on this thread http://www.diyaudio.com/forums/showt...87#post1177687

Apparently, they have an Fs of about 3Khz. IThey are around 1.5" across the short side of the face ie. aligned across the phase plug and the cut out hole is EXACTLY 1 1/8" .

Here's an impedance plot of these tweeters
http://www.diyaudio.com/forums/attac...amp=1175916108

[zobsky]

My main reference for line arrays is Jim Griffin's excellent paper http://www.audiodiycentral.com/resource/pdf/nflawp.pdf.

I plan to build the array 12 woofers high arrayed as 3 series elements consisting of 4 parallel woofers each resulting in a 6 ohm nominal impedance.

From this handy table ( http://ratch-h.com/arrayimpedance.html ) , to end up with a roughly equivalent tweeter line impedance and physical height, I'd be best off with around 35 tweeters. I'm guessing that I may have to pad the tweeter down somewhat depending on the distance to the listener.


A line array crossover should be steep and accordingly ,I'll be using a 3rd order hopefully crossed over at around 4 KHz. Once the line is built, I plan to use my active crossover to determine what works and what doesn't.

Here's the leap of faith:
This figure from Dr. Griffin's work talks about line length, cancellation etc. I've added in the appropriate lines for my drivers. Because of the cheap tweeters, I can't crossover close to the recommended 1 wavelength corresponding to the woofer CTC distance (around 133 mm or a little more depending on spacing)
However, the woofer (wideband actually) has a rising response and while I'm close to the 2 x wavelength cancellation point I hope this rise will help offset the reduction in SPL due to comb filtering cancellation losses.

[zobsky]

Some renderings of possibilities

Straight baffle
http://img337.imageshack.us/img337/5793/bafflers5.jpg

U-baffle
http://img128.imageshack.us/img128/1998/baffleuik8.jpg

K-baffle (slots shown straight instead of exponential because they were easier to draw that way)
http://img404.imageshack.us/img404/7159/bafflekeq2.jpg

tapered transmission-line (a.k.a Bill Fitzmaurices' TLAH)
http://img404.imageshack.us/img404/1...letlinehe7.jpg
(I also thought of a MLTL variant of this, .. but was told that it wouldn't work, . .

baffle with helper woofer (bass horn)
http://img402.imageshack.us/img402/9...flehornke6.jpg

I'm intrigued by the possibility of a K-slot on the rear of the baffle with polyfill batting behind it todamp out some of the midrange. This seems like a hybrid of U-Baffle, quarter wave T-line , and Karlson rocket ( http://home.planet.nl/~ulfman/images/rocket.jpg ) . I've seem instances of OB karlsons (though never in a line array), .. not much subjective information about the sound.




Opinions, thoughts, criticisms, get rich quick ideas ....?

[thylantyr]

Found this pic in cyber. A line array using those tweeters.
Looks nice for a budget design. Pay attention to details and
the line array should perform very well.

http://home.pacbell.net/lordpk/speakers/93.jpg

[zobsky]

Thanks, .. I saw that a few weeks ago @ http://www.caraudio.com/forum/showthread.php?t=208488 (enjoyed reading the holy war too, .. it's amazing how uninformed some people are, even compared to yours truly ).

If nothing else, one good thing about the tweeter is that it's easy to mount (front mounting / standard 1 1/8" hole, only 2 screws per driver) and the center to center distance is small (the flange only extends to the 2 screw holes).

[bjorno]

http://home.pacbell.net/lordpk/speakers/93.jpg

I think the line arrays are arranged the wrong way considering sharp central phantoms.

The expected horizontal blur is at least as wide as the driver c-c distance, i.e. the M-T c-c, or the width of the enclosure if looking in the time intensity trading window, within 700uS, where the main location cues is provided to the ear/ brain.

The line arrays should physically be swapped, left with right due to the fact that the horizontal azimuth angles are not (psycho-) acoustically constant for high and low frequencies.

It is known that about an angle of 46 degrees is noticed by the hearing for frequencies over about 2 kHz, if its a 60; 60:60 set-up and only frequencies lower than 1 kHz is occupying the 60 degree azimuth angle.

Substantial improvement of location and phantom sharpness concerning the central MAA (Minimum Audible Angle) can easily be noticed if comparing the two different possibilities.

b

[zobsky]

Originally posted by bjorno


I think the line arrays are arranged the wrong way considering sharp central phantoms.
....

b

Good point, .. I totally agree.

[bjorno]

Sorry, My syntax checking in my spell checker has played with me again. This sentence came out totally wrong:

It is known that about an angle of 46 degrees is noticed by the hearing for frequencies over about 2 kHz, if its a 60; 60:60 set-up and only frequencies lower than 1 kHz is occupying the 60 degree azimuth angle.

Replace with:

It’s known that frequencies at and slightly over and about 2 kHz, stereo-dispersed about at azimuth angle of 46 degrees is by the hearing noticed to coincide with a 60: 60:60 set-up.

Only frequencies above 2 kHz or higher is gradually occupying more than this 60 degree azimuth angle when placed at 46 degrees.

For frequencies lower than about 1 kHz Hz in a listening room but higher than about 350 Hz the azimuth will narrow from the 60-degree angle as the frequency goes down.

This means for achieving maximum phantom azimuthally related sharpness, the MMA (Minimum Audible Angle) and obviously the music related CMMA (Concurrent) optimised for a multi way speaker set-up, then the T-M combination in a two-way or multi-way system should be slightly horizontally separated and preferably done with a continuously angular dependent frequency emanating from a membrane like:

http://www.isvr.soton.ac.uk/FDAG/VAP/index.htm

Click at the OSD (Optimal Source Distribution) and then the presentation.

This can be done with a discrete set-up too and provide pinpoint accuracy and resolution for phantoms centrally located in music.

My point is, that if ever micro details exists in music when listening to stereo they should be portrayed also with localisation accuracy.

Only sharp phantoms can be resolved with segregated details the brain ear combination naturally prefers without the otherwise accompanied listening strain that occurs from a blurred phantom center.

The experienced resolution requirements for a speaker system that is mentioned in another tread: Beyond.. is not enough to fool the ear/brain to believe a phantom is real, resolved and precisely located because the auditory stream segregation qualities can only be optimised if phantom location and cross-talk issues is also addressed an taking into account for a 2 speaker based stereo set-up.

Even a dipole set-up (even if only one driver is used for each of the speakers) will have phantoms that are elliptically distributed and smeared in the horizontal plane, thus the possible low MMA is unnecessary corrupted for a person in the sweet spot.

Other research results that is backing up (confirming the effects) this approach when considering localisation for different frequencies and bandwidths:

Sam Ferguson and Densil Cabrera: Vertical Localization of Sound from Multiway Loudspeakers.
…and when phantoms combines with visuals:

Multisensory enhancement of localization with synergetic visual-auditory cues : Martine Godfroy and Corinne Roumes


b

[zobsky]

Number crunching time.

12 eight ohm woofers wired 4 x 3 = 8 /4 x 3 = 6 ohm nominal

35 eight ohm tweeters wired 7 x 5 = 8/7 x 5 = 5.7 ohm nominal

Questions:

• Should I worry about the 0.3 ohm difference (I know this will make the tweeter marginally louder)?
• What's a better way to wire an array, .. wire some drivers in parallel and then series connect these groups, .. .or vice versa?

I've had an 8 x 4 sheet of plywood sawn up at the local Home Despot . The pieces are 4 sections measuring 6 x 1 and a single 2 x 4 . I plan to build a tapered U-baffle first, .. and then play with the K-slot idea I had and also the TL.