Murphy Corner Line Array (MCLA)

[russ_L]

Whoops, my bad. The typing fingers got ahead of the brain. That should be twenty-two (22) drivers, not 21.

Russ

[John L. Murphy]

Russ is right. There were only two ways dmontella could wire 22 drivers. For 8 Ohm drivers to have the same drive voltage the only choices would be 44 Ohms or 1.45 Ohms as Russ has told us.

This is a good point and might have contributed to dmontella's dissatisfaction with what he heard. With a 44 Ohm load he may have been into heavy clipping at his power amp just to get the system to a reasonable listening level. If it were configured as a 1.45 Ohm load, then the power amp may have been experiencing spurious oscillations resulting in distorted sound.

But I mostly think that dmontella didn't care for the sound because he never voiced the array with an appropriate EQ. His listening impression appears to have been based only on the raw array response. It's too bad he did not complete the project by applying the required EQ correction.

I originally concentrated on voicing the arrays to achieve a flat, spatially averaged, in-room frequency response just to show what could be achieved. But that flat response was unlistenable on most recordings. Early on I used the standard X-Curve cinema voicing to reduce the treble. After trying out many different voicings I have more or less settled on a variation on the "10 dB tilt" demonstrated by Sean Olive to be the most preferred in-room spatially averaged frequency response (Sean E. Olive, John Jackson, Allan Devantier, David Hunt, and Sean Hess, “The Subjective and Objective Evaluation of Room Correction Products,” presented at the 127th AES Convention, New York, preprint 7960 (October 2009)). Basically Olive demonstrated a listener preference for an in-room spatially averaged response that is very smooth but tilts downward from low to high by about 10 dB with the response at 40 Hz about 10 dB higher than the response at 20 kHz. See the top curve in the attached image for the most preferred frequency response in Olive's carefully controlled study. Remember, these are spatially averaged in-room responses.

Had dmontella equalized his corner arrays to achieve Olive's "10 dB tilt" his reported listening results would certainly have been different.

Regards,
John

[wogg]

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

AS I lurked on this thread I was also convinced you we're seeing the possibilities with series parallel, in which there are a whole lot more ways to wire than just 2. However...

Russ is right. There were only two ways dmontella could wire 22 drivers. For 8 Ohm drivers to have the same drive voltage the only choices would be 44 Ohms or 1.45 Ohms as Russ has told us.

The key words to have the same drive voltage were not included, and change the options dramatically.

I'm also going to guess that John is on point with this:

But I mostly think that dmontella didn't care for the sound because he never voiced the array with an appropriate EQ.

That sort of thing really needs active filter voicing to work.

[Sydney]

Consider this thread dates back to 2009; Since then there have been advances in Arrays and related electronics.

[John L. Murphy]

<snip>
That sort of thing really needs active filter voicing to work.

Yes, exactly. The corner-line-array is based on using an array of full-range drivers with active EQ customized to achieve a specific targeted response based on spatially averaged in-room measurements. My prototype arrays using 24 drivers per array achieves the response shown below when targeting a flat response. But nobody likes speakers with a flat frequency response (see my previous post above) so I actually target a response close to Olive's "10 dB tilt" but with an extra nudge in the bass below 200. This is the voicing that I currently prefer in my own personal listening tests.

You can see the MCLA project web site here: https://trueaudio.com/array/

Regards,
John

[Neil_C]

Hi John

I am very interested in building the MCLA. My background is that I am a semi-pro church organist and recording engineer, producing sample libraries for the Hauptwerk Virtual Pipe Organ. I am looking for speakers which will adequately reproduce the scale of a large pipe organ in my 15' x 12' studio, which I can equalize (with my MiniDSP 2x4HD) both for my seating position at my organ console, exactly equidistant between the corners of the "long" wall and 4' out from the front wall, and also for a couch listening position further into the room for normal music listening.

Your design makes perfect sense and would be a great fit for me as the theory deals with the corner reflections Siegfried Linkwitz talks about, but without using dipoles which would dominate my studio space and leave no room for my substantive music hardware.

My problem however is similar to that of Dmontella; I can only fit a maximum of 23 driver in my floor-to-ceiling space - in order to meet UK building regulations for my garage conversion, the floor had to be raised to allow installation of insulation.

I have thought of several possible solutions but would welcome your advice on which would be the least detrimental to the sound quality:

1) Use a shorter array which would allow a sensible system impedance - if the array was shortened, would it be better to have the top drivers near ceiling level or the bottom drivers at floor level? The room has (insulated) dry-wall walls and ceilings with a concrete-joist-particle board floor with thick carpet. I am very sensitive to room acoustics and the confusion they bring to the party so for years I have done most of my mixing (choral and organ recordings mostly, some recordings of my progressive rock band) on headphones but this is a very flawed and solitary solution to the problem.

2) Would it be possible to replace the missing 24th driver with a high dissipation resistor, given that the individual driver power requirement in the MCLA is pretty low?

3) Would it be preferable to use a 2 1/2" driver rather than the 3 1/2" (and in an impedance-friendly configuration)? I recall your saying that the 2 1/2" Dayton driver would result in closer acoustic centres for the drivers, which would be an improvement. Any differences in the required equalisation are not an issue, as I have used the MiniDSP 2x4HD extensively with my current system, allowing very tight integration of my Decware Housewrecker sub (they weren't joking) with various passive and active near-field monitors, currently some Roland MA15-D micro-monitors.

Kind regards
Neil

[John L. Murphy]

Hey Neil,

I think you would be happy using the MCLA for your monitoring/mixing application. Using 31/2" drivers I would fall back to 20 per side and use 5 parallel strings of 4 drivers to achieve (4/5)x8 = 6.4 Ohms net load. Intuition tells me to not center the array vertically but rather have slightly different gaps at each end. Ultimately what matters most is the voicing you target. For music mixes the "10 dB tilt" is likely to be most useful. If mixing for film/cinema then switch to X-Curve voicing.

Since you are using a sub-woofer the change to 21/2" drivers seems reasonable, with the array used to 80 Hz or so. Just be careful when selecting an alternate driver and make sure to listen to and measure the candidate drivers before making your final selection.

Best Regards,
John

[ernperkins]

Perhaps the Dayton ND65-8 would be a good fit for 2 1/2 inch drivers?

[Neil_C]

John, thanks for your reply. I thought about going to 20 drivers as it's a nice load but what concerned me was the power handling reduction and reduction in cone area. While I'm using the Decware sub right now, that's actually temporary - we built it for a laugh - and I had wanted to do away with a separate sub because even built in 25 mm MDF and braced, I can "hear" the box. I would really like to use the MCLA full-range so would the driver no. reduction make much of an impact? I don't listen at extreme levels and I'm basically going to be 9' from the array or less under most circumstances.

If the answer to the above is, yes a sub would be advisable, I was looking at either a couple or three Ripoles or open baffles using servo-controlled 10" or 12" woofers.

Regarding a 2 1/2" option, ernperkins is correct; I was thinking of an array of 32 ND65-8's. I was also thinking of purchasing a license for Winspeakerz to model the ND65-8's if that's something you think would be worthwhile. My brother and I have built a lot of subs over the years using BassBox Pro 6 but often the results in "extreme" cases (all of our cases are extreme!) don't work out quite right.

Kind regards
Neil

[Tin_Ears]

I realize this is an old thread but the MCLA has always intrigued me. I'm going to try to gather the funds to build something very similar.

I have a question about adding a center channel. Since the center is a flat wall, what's the best way to closely match the corners? Should I used a double row of center drivers and different EQ?

[Tin_Ears]

^^^ anyone??

[billfitzmaurice]

There's no need to closely match the corners, as the center channel information isn't the same as the L/R. Using the same drivers is all you need to be sure of. I did in my L/R line arrays and curved array center and they work just fine.

[Tin_Ears]

Thanks Bill. What I'm wondering is how to compensate for the lack of gain in the bass region. The corners will surely have a bit more oomph at the low end. I know I could use EQ but I'd rather add more drivers in the center.

[billfitzmaurice]

What I'm wondering is how to compensate for the lack of gain in the bass region.

You don't. That's what subs are for.

[Tin_Ears]

You don't. That's what subs are for.

Of course but I want the center to match the L/R as closely as possible. Given the low-end boost the corner speakers will enjoy, should I use more drivers in the center?