Murphy Corner Line Array (MCLA)

[zeus]

Re: Murphy Corner Line Array (MCLA)

Would a 27 driver array using the Dayton ND-4s be appropriate for a room with a 9 foot ceiling? The 27 drivers could be divided into 3 sets of nine drivers in series, for an approximate impedance of twelve ohms.

[John L. Murphy]

Re: Murphy Corner Line Array (MCLA)

Would a 27 driver array using the Dayton ND-4s be appropriate for a room with a 9 foot ceiling? The 27 drivers could be divided into 3 sets of nine drivers in series, for an approximate impedance of twelve ohms.

Hello Zeus,

Yes, assuming that the ceiling is parallel to the floor. I also assume that you are referring to the 4" ND drivers...the ND-90-4 or even the ND-91. 27 drivers seem ideal for the 9' height and the impedance works out fairly well. You would want to use a fairly high powered amp to have plenty of voltage swing headroom for the higher impedance. The good news is that the power amp will only be lightly loaded with little stress resulting in lower distortion than with its rated load.

If you need very high sound levels or are using it in a very large room (like a club) you may want to reconsider the 12 Ohm configuration and shoot for a lower system impedance. It just depends on your needs.

Good luck with your project. Please keep us informed of your progress if you would.

Regards,
John

[Tin_Ears]

Re: Murphy Corner Line Array (MCLA)

Bill's were the PE Pioneer NSB Buyout . I have enough of those drivers to build 2 corner-loaded arrays and a center channel. How do you think these would perform in an eight foot tall corner-loaded array? One corner of my room has only 4 inches of space so I'd be limited to that width. Just to get the Qtc down to 1.0 would require a box 4x10x96 with internal dimensions of about 3x9x94 which is about .065 ft3 per driver after subtracting the truncated 45 degree corner for the driver baffle. I guess my biggest concern would be the asymmetrical shape of the box, being 4 inches across the back and 10 inches on the side. The center channel would be mounted at the joint between the front wall and the ceiling.

[Ron_E]

Re: Murphy Corner Line Array (MCLA)

Ron,

Yes! The SPL increase due to corner loading is greatest in the bass range. So corner loaded sub-woofer arrays would make sense to me...unless they are competing for the location with a full range line array. Then I would let the full range array have the corner location.

Regards,

John

Is there a way to estimate how much gain I would get with a sub only array? My listening room is 12'x10'x7.5' with the speakers on the 10' wall. With a slightly staggered line I could get 16 of the Aura NS6-255-8A woofers in each corner. http://www.parts-express.com/pe/show...number=299-030 I modeled them in Winspeakerz in 5 cu.ft. with 600 watts and this is what I came up with. The longest dimension in the room is 17.4' so I included room gain set at 32 hz.



I plan to run them stereo and I have Ultimate Equalizer to take care of any EQ. http://www.bodziosoftware.com.au/ The rest of the system will probably be 3 way with ribbon tweeter, planar mid, and woofer. I assume I can integrate all of this together since this is a dedicated listening room with fixed listening position. Any input is much appreciated.

Ron

[Catfish]

Re: Murphy Corner Line Array (MCLA)

Now this is a very interesting project indeed.

I was one of those audio professionals to whom Mr. Murphy refers as having heard his original line arrays in his living room. Twenty something years ago, I worked at an audio electronics company where John was working as Chief Scientist. I used to annoy him with questions every chance I got and he was always gracious to answer and explain with patience. I credit him with giving me my start in what has been a wildly successful career in pro audio product design and manufacturing. John has the heart of a teacher and it doesn't surprise me to read that he's placed his design in the public domain so others can enjoy and build upon his work.

Those original line arrays were something. I was skeptical at first, but hearing was believing. Of course, those cabinets looked like coffins and took up way too much living room to ever pass the "wife test." Using the corner effect to simulate increased cabinet volume is a novel idea. A bit of designer grill cloth and these could disappear into the architecture. I'm curious if these could be built in smaller segments to give them some modularity. Could one, for example, make two 4-foot arrays and stack them vertically to get the same result as a single 8-foot array? This might make them more commercially viable for portable sound systems and even cabinet makers to build and ship via UPS.

The need for frequency shaping appears to add a level of complexity to the project. I suppose your average home-built speaker is expected to deliver correct response from flat amplified audio. Call me a gear-snob, but that Behringer EQ isn't shouting "audiophile" to me. Maybe a more elegant solution would be to build the equalization into the speaker box. Okay, passive components working with powered signals are probably not realistic; it's difficult enough to build a passive crossover.

So instead, maybe the MCLA becomes a powered speaker. Powered speakers offer many benefits over passive speakers. The amplifiers can be matched perfectly to the drivers. The internal wiring can be known and compensated. Power limiting and protection can be optimized. Low level audio can be sent to the speaker cabinet eliminating long speaker wires which can be lossy and sometimes introduce instability. And of course, in this case, the front end to the power amp can have the frequency-shaping circuitry. Maybe a few dip-switches to flip if the MCLA is moved from a corner (eighth space) to a wall (quarter space), etc. Or some trimmers for calibrating speaker-to-room response.

Conversely, if powering is not desired, one could still make the electronics box and insert it into the signal chain between source and power amp. Maybe we could design this box and publish the schematic and bill of materials in the same way that John published the MCLA design so the more adventurous hobbyists could build their own. Just thinking out loud here.

Anyway, cool project. And to the poster who feels like tossing in the towel on the project, hang in there. I suspect with some frequency shaping tools, you're going to be very happy with the results.

Mike Dosch aka Catfish

[richidoo]

Re: Murphy Corner Line Array (MCLA)

I think dmontella is gone, but I have some ideas about where he went wrong.

There is too much insulation inside the box. Experts like Vance Dickason say to stuff those sealed boxes full to simulate larger volume and lowerQ, but there is a trade off. We have active amplification and electronic EQ so we can manipulate the Q digitally. The excess stuffing robs microdynamics and detail converting them to heat inside the box. With sealed speakers you apply only the minimum internal acoustic damping necessary to prevent internal echos and sound coming back out through the speakers. The box size is HUGE for each individual driver, so I think internal echo is pretty well diffused temporily anyway. Thin absorption like that 1/2" stick-on foam from PE on just one internal side, and maybe 4" of FG stuffing at the top or bottom to reduce lengthwise LF echos. The rest should be hard wood internal surface.

With regard to the tubby bass compared to his IB subs, I question the wiring hookups inside the box. When this many connections are made they must be soldered and large gage hookup wires used so that the source impedance of the amplifier stays low and speaker damping remains high. Damping is where the bass detail lives (and low phase error, as John has discovered.) Put a resistor inline with one speaker cable connection and you will hear all of your bass detail go out the window. I have built speakers using the 14ga stranded speaker wire in the clear insulation, available from Home Depot, labeled RCA brand. It is fine for HTiB, but it is awful for hifi. Looks like clear insulation on the hookup wires so maybe RCA.

I own Behringer DCX246, and I agree with Catfish that the Behringer stuff is terrible, poor reliability and poor sound quality from audiophile perspective. But if it's working it's not so bad as to cause the stated problems. It is not muffled treble, nor warm bass, more like gritty and cheap sounding.

The off axis attenuation is to be expected. All speakers beam the high frequencies in proportion to the size of the moving surface. This is why real tweeters are small. Full range drivers always have this compromise. You have to sit in the sweet spot. That these have dispersion to 25 degrees is excellent.

Sonos does have coax digital output which is high in jitter, but might solve the optical input issue John warned of.

I was tempted to build these with 9 foot height but a local high end speaker manufacturer told me that I would not be happy with the high freq. I tend to pad down the tweeters anyway so I am still very interested to hear them. I have a Rane RPM26z that could do the EQ. Or I could do it with a PC using AudioLense. Or, just put a real tweeter at ear height, and 1st order passive filters between the array and the tweeter at 5k, or wherever the off-axis testing suggests. SB Acoustics dimple tweeters are very inexpensive and very good sounding. I'll have to look on PE catalog for something similar. Certainly a horn loaded compression driver could provide all the treble energy you could want, with excellent dispersion, and PE now sells the excellent SEOS horn.

Hmm, this is starting to get me excited again!
Rich

[john3v]

Re: Murphy Corner Line Array (MCLA)

Hello John, v01d,

50 of NS3 are traveling my way. I will have time for this project probably around the end of January. No Behringers in the path, will do EQ with free tools on my PC as it is my only source these days. I may not be able to place them in the corners, so probably will go with simpler cabinets. Most likely modular design (3 or 4 separate boxes per column) so I can move them around easily.

I will try them with 300b and Aleph-3 amps. For now the only reason I am going with my own line array - anything else with that sensitivity seems to have deficiencies I like way less than compensating the high slope.

Will drop a line how it turns when done, could be another 6 months from this post though

I wonder if anyone else has managed to build the array and achieved good results? Especially with NS3.

Cheers.

[oneplustwo]

Re: Murphy Corner Line Array (MCLA)

Anyone have any new builds to report on? I'm considering the concept but with Tang Band TC9 drivers and doing EQ with miniDSP. They'd be driven by Aleph X monoblocks and perhaps get bass supplemented with infinite baffle subs.

[russ_L]

Hi All- still interested even after several years to going down this path. I'm sure we were all disappointed in dmontela's poor results. And thank you John and Rich for suggesting some possible solutions to the poor results. But unless i'm missing something 22 drivers per side?? That's 11 x 2 yielding 44 ohms, or 2 x 11 yielding 1.45 ohms!! dmontella is doomed to fail in either configuration, or again am I missing something?

Russ

[billfitzmaurice]

11 x 2 yielding 44 ohms, or 2 x 11 yielding 1.45 ohms!! dmontella is doomed to fail in either configuration, or again am I missing something?

Series, parallel. You don't use one or the other, you use both.

[Sydney]

... again am I missing something?
Russ

Besides wiring for safe amplifier load - Most nowadays would implement shading of array segments.

[russ_L]

Don't know what you're getting at Bill. Two rows in parallel, each row with eleven drivers resulting in 44 ohms. OR, eleven rows in parallel, each row with two drivers resulting in 1.45 ohms. One or the other, both terrible. That's my point. dmontella had no chance.

Russ

[Jim Griffin]

Russ, I don't know your project so let me outline a line array which has 25 (4 ohms impedance) drivers in each side. I want to wire them so that I attain a 4 ohms equivalent impedance for the composite of 25 drivers per side.

One way to do this is to group the 25 drivers into groups of 5 drivers each so that I have 5 sets of 5 drivers. For this example, number the drivers 1 to 25 from ceiling to floor. Connect the + terminal on driver #1 to the amplifier which powers the array. Wire the - terminal of #1 to the + terminal of driver #2. You are wiring these drivers in series so continue this pattern through driver #5. Now connect the - terminal of #5 to the - terminal of the amplifier. You have a string of 5 drivers connected in series (an impedance of 4 ohms X 5 = 20 ohms. Now wire drivers # 6 to 10 in series and connect to the amp the same way as you did for the first set of 5. Repeat for drivers #11 to 15, then repeat for #16 to 20 and #21 to 25. You have yielded 5 strings of drivers which have 20 ohms impedance each. Now the 5 strings of 20 ohms are connected parallel across the amp so you have a 20 / 5 = 4 ohms impedance load across the amp when you are finished.

The second way you could connect the drivers is to connect the + terminals on each set of 5 drivers together to yield an impedance of 5 / 4 = 0.8 ohms for each set. Then connect the - terminal of 5 each set to the + terminal of the next set to eventually yield a total series of 5 of these sets of 0.8 ohms. That yields a composite of 5 X 0.8 = 4 ohms.

Either way you can use series connections of the individual drivers and then parallel them to create a 4 ohms total. Or parallel connect the drivers in each set and then series connect them across the amp to yield the same 4 ohms for the composite.

All this can easily be visualized quicker that I can explain it.

My white paper explains line arrays.

http://www.audioroundtable.com/misc/nflawp.pdf

[billfitzmaurice]

Don't know what you're getting at Bill.

Use this:

http://www.colomar.com/Shavano/impedance_proc.php

[russ_L]

Guys. I know how it works. I'm just trying to point out that dmontella had two ways to wire the twenty-one (21) drivers. One way results in an impedance of 44 ohms. The other way results in an impedance of 1.45 ohms. And this most likely caused the poor performance he observed. John Murphy and richidoo missed this issue when recommending possible causes.

I'm just looking for someone to agree with me; yes that could very well be the reason.

Russ