Magnepan Magneplanar MG3.6/R loudspeaker

Brian Damkroger, August, 2000

Bonnie and I decided to avoid the crowds last weekend, and instead settled in at home to watch the recent remake of Great Expectations, with Ethan Hawke and Gwyneth Paltrow.  It seemed like a pretty good movie, but before long I found my thoughts drifting to the review I had in progress: my audition and analysis of the Magnepan Magneplanar MG3.6/R.  True, Great Expectations is a little slow, and a few explosions or car chases might have better held my attention, but if ever there was an audio product to which the phrase ďgreat expectationsĒ applied, itís the Magnepan 3.6/R.

mag36.jpg The MG3.6/Rís immediate predecessor, the MG3.5/R, was a breakthrough product for Magnepan.  It was a huge commercial success, and established a spectacular new level of performance for Magnepan in terms of dynamics and transparency.  As if that werenít enough, the 3.6/R comes right on the heels of the MG1.6/QR, another huge success, and an industry-wide benchmark for performance in a $1500 loudspeaker.  I reviewed the 1.6/QR in January 1999; it is the least expensive speaker in Class B of Stereophileís ďRecommended Components,Ē and one of the least expensive to ever appear there.

Months before the MG3.6/R was even introduced at the 1999 WCES, a buzz permeated the Internet about ďthe new Magnepan,Ē and I received a steady stream of e-mail messages asking about it.  ďIs the 3.6 as good as Iíve heard?  Is it really all of the updates developed for the 1.6, now applied to the 3.5?Ē

Nowhere were expectations greater than at Casa McKenzie-Damkroger.  Iíve been listening to Magnepans evolve for two decades.  Iíve admired their coherence and loved the uncanny way they could capture the sense of real instruments playing in a real space.  Conversely, their lack of dynamics and slight opacity were always barriers between the music and me, barriers diminished in each succeeding generation, and nearly eliminated in the MG3.5/R and 1.6/R.  Now comes the MG3.6/R, so maybe...?

Great expectations, indeed.

Basic Technology: What is an MG3.6/R?
Several Magnepan loudspeakers have been covered in these pages, including two of the MG3.6/Rís predecessors, the III and IIIA.  The 3.6/R carries forward the same configuration, layout, and driver technology.  Itís a three-way design with crossover points of 200Hz and 1700Hz.  The planar-magnetic driver is a 0.5-mil-thick Mylar diaphragm, onto different areas of which have been fastened separate, current-carrying wire grids for the bass and midrange.  The top end is handled by Magnepanís unique, 55"-long ribbon—a true, free-standing ribbon in which the current-carrying aluminum ribbon is also the driving element.

The 3.6/R is cosmetically identical to the 3.5/R: a slim, elegant tower approximately 6' tall by 2' wide by 1½" deep.  My pair was covered with an oatmeal-colored, open-weave fabric, with dark cherry strips flanking the panels and separating the tweeter and midrange-bass sections.  The panels are mirror-imaged, with the planar-magnetic driver located to the inside in the recommended setup, and the ribbon tweeter to the outside (footnote 1).  Connections (single or biwire) are made via banana plugs to an external crossover box that plugs into the panelís rear.  Magnepan also makes an optional crossover for bi-amping, but I did all of my listening with the standard unit.

Although it retains the 3.5/Rís basic configuration, appearance, and driver technology, the 3.6/R differs slightly in some system parameters.  The changes reflect both a response to perceived shortfalls in the 3.5/R and lessons learned in the successful transformation of the 1.5/QR into the giant-killer 1.6.  The goals for the 3.6/R were to improve low bass power and articulation, smooth the in-room midbass smoothness, and better integrate the drivers.  The first was accomplished by increasing the midrange panelís area from 170in≤ to 199in≤, allowing the bass/midrange crossover point to be lowered, and the bass panelís tuning to be optimized for a narrower frequency range.  Better integration and smoother in-room response were achieved primarily by careful optimization of the tensioning, damping, and partitioning of the diaphragm—the ďblack artĒ responsible for much of the transformation of the 1.5 into the 1.6.

Footnote 1: The tweeters should be slightly farther from the listener than the bass-midrange panel, so will be placed inboard or outboard, depending on distance and toe-in.



System and Setup
I did all of my listening in my main 17' by 23' listening room, with the Maggies firing across rather than down the roomís length.  The setup put them approximately 3'6" out from the front wall, and the speakersí outer edges approximately 7'10" from the left wall and 4'11" from the right.  The speakersí inside edges were about 5'8" apart, their centers each about 13' from my listening position.  I settled on a slightly toed-in configuration, with the speaker axes crossed at a point approximately 6' behind the listening position.

My past experience with Magnepans led me to expect a fairly easy setup and optimization process, and that proved to be the case.  A few things are worth noting, however.  The MG3.6/Rís radiation patterns—dipole for the bass, a line source for the midrange and tweeter—reduced bass problems with room boundaries, but made sidewall interactions a bit more of a concern.  Positioning too close to a side wall could cause the image to come forward along the side walls, distorting stage placement and image size.  In my room, with a 23'-long wall behind the speakers, it wasnít an issue.  Itís also been my experience that Maggies in general work best when backed by a solid but irregular wall.  Hard plaster and adobe are good, brick and stone are better.  None was an option for me, so I had to make do with drywall and lath over concrete block.

Another consideration is that although the 3.6/R is a benign load—mainly resistive and a fairly flat 4 ohms—at 86dB/2.83V/m theyíre not terribly sensitive.  The VAC Renaissance 70/70 is an unusually strong 70W amp, but wasnít really enough to make the Maggies sing.  The Mark Levinson No.20.6s, VTL Ichibans, and Classé CAM-350s all did better jobs of resolving low-level dynamics and detail, and opened up the soundstage noticeably.  I spent time with all three, but ended up preferring and doing most of my listening with the Classé monoblocks, which are rated as delivering 700Wpc into the Maggiesí 4 ohm load.

The rest of the system remained constant throughout the review period: my VPI TNT IV/JMW Memorial turntable/tonearm combo with Grado Reference cartridge, SimAudioís new Moon Eclipse CD player, and a VAC CPA1 Mk.III preamplifier at the center of it all.  Nirvanaís new S-X interconnects arrived mid-review and immediately claimed their territory.  I biwired the MG3.6/Rs with Synergistic Research Designerís Reference when the Classés were in use, and used Kimberís Bi-Focal XL with the VTL and Levinson amps.

Bright Starís Rack of Gibraltar and Air Mass, Big Rock, and Little Rock isolation products kept everything stable and quiet, and AC was fed through an MIT Z Stabilizer (amps) and Z System (front end), with a Nirvana isolation transformer providing an extra measure of isolation for the Moon Eclipse.

I ended up using only a minimum of room treatment—a single 14" ASC Tube Trap in one front corner (reflective side out), an EchoBuster diffuser panel in the other, and a combination of EchoBuster BassBuster columns and homemade panel resonators in the rear corners.  EchoBuster absorbers were mounted to the rear wall, behind the listening position.

Use and Listening: Can Great Expectations be Met?
Great Expectation No.1: A huge, open, holographic soundstage. Magnepans have always gotten ďthe space thingĒ right.  Whatever their other pluses or minuses, theyíve been able to create a more realistic soundstage than most speakers, and better capture the sense of real instruments playing in a single, coherent acoustic environment.  The 1.6/QRs were very good in this regard; the MG3.5/Rs were outstanding.

The MG3.6/Rs didnít disappoint me in the least.  Their soundstage was huge—extending well outside the speakers, and the deepest of any speaker Iíve used.  Front-to-back layering was superb; in fact, the 3.6s set a new standard in this regard.  They didnít just clearly define the position of the instruments on the stage and the surrounding hall boundaries, or even do so with a greater degree of precision and specificity than other speakers—they also quite clearly described the spaces between the performers, and between the instruments and an adjacent hall boundary.  A lot of speakers can do this in the lateral plane, but none—in my experience—can do it so well with respect to the front-to-back distances.

The effect is particularly riveting on naturally recorded works, where the hall ambience is discernibly woven between the instruments.  For a dramatic example, try John Eliot Gardinerís recording of Henry Purcellís The Tempest, with the Monteverdi Choir and Orchestra (Musical Heritage Society 4479).  Most speakers can assign the correct depth cues to the orchestra and various singers, and correctly place the images on the stage.  Good speakers clearly track the singers as they move forward and backward on the stage.



With the MG3.6/Rs, there was also a continuous ambience field that stretched from the side walls down into the front-to-back spaces between singers, who were clearly and obviously moving around within a single, defined acoustic envelope.  I often felt as if I could actually enter the recordingís acoustic environment and wander around among the performers.  Even on good studio recordings, where thereís no real ďstageĒ per se, the soundstage and images were so tangible that it seemed as if I was almost able to get between and behind the performers.

Great Expectation No.2: Pinpoint precision and extraordinary detail. While Magnepans have always done a good job of soundstaging and their images have always been wonderfully coherent with the surrounding space, theyíve never had quite the precision of the best cone-type speakers.  Each succeeding generation of Maggies has improved on their performance in this regard, and both the MG3.5/R and the 1.6/QR were dramatic improvements over their predecessors.  But the picture was still a little diffuse—certainly not a Monet, but not quite a laser photograph either.

The MG3.6/Rs didnít noticeably improve on the 3.5sí performance in this area.  The performersí images were natural, and there was sufficient detail to resolve, in a general sense: individual instruments within an orchestral section, even within dense, complex passages.  Similarly, the imagesí edges interacted naturally with the surrounding space, the notes blooming and expanding, the overtones dissolving into the background ambience.  However, there werenít the layer upon layer of fine detail, the complexity, or the density with which speakers like the best Thiels and Avalons can imbue an image.

The situation wasnít perfectly black-and-white, however.  I typically sit somewhere mid-hall at local symphony and chamber orchestra performances, and the perspective there isnít terribly dissimilar to the Maggiesí slightly diffuse portrayal.  Conversely, the added detail that the Thiel CS7.2s provided (se February Ď00, pp.119-127) unquestionably made voices and instruments more vibrant and alive.

A great example was ďChuck E.ís In Love,Ē from Rickie Lee Jonesí live acoustic album, Naked Songs (Reprise 45950-2).  Through the MG3.6/Rs, her guitar and vocals, even the audience sounds, sounded very natural, nicely detailed, and dimensional.  With the big Thiels, however, the extra detail and complexity seemed to supercharge the images and make them breathe, and gave the performance a presence and life that had me turning out the lights and sitting spellbound in my chair.

Great Expectation No.3: Seamless top-to-bottom consistency. This is another traditional Magnepan strength, and an area in which the MG3.6/R proved a solid improvement on its predecessor.  The 3.5/R is wonderfully consistent across the frequency range, but if you listen closely, it loses a bit of articulation in two areas: from the midbass on down, and in the upper midrange to lower treble, just before it transitioned to the ribbon tweeter.

The 3.6/R was every bit as seamless and consistent as the 3.5.  There was a slight warmth to its tonal balance in my room, probably reflecting a boost in the upper-bass region, but no overt discontinuities in character or distortions—nothing to draw attention to the speaker.  Both instruments and soundstage remained consistent—cut from a single cloth, if you will—across the entire range of frequencies and levels.

The 3.6/Rís bottom end was an improvement over the 3.5ís, remaining powerful, clean, and articulate all the way down to about 35Hz in my room.  The fast electric bass runs on Fourplayís ďBali RunĒ (from Fourplay, Warner Bros. 26656-2) are a true torture test.  The 3.5/R got muddy and confused during these passages, but the 3.6/R sailed right through them.  There wasnít the absolute power or last bit of detail at the very bottom that I hear from the Thiel CS3.6 and CS7.2, but the Maggie had a goodly amount of slam, with crisp, fast transients and excellent pitch definition.

The 3.6/Rís upper-midrange performance was excellent as well, with no perceptible loss of detail or obvious transition to the ribbon tweeter.  Piano recordings showed this off well, and Dick Hymanís In Recital (Reference Recordings RR-84CD) is a particularly good example.  This very natural-sounding recording has a slightly distant perspective and a very well-defined portrayal of both the instrument and its interaction with the surrounding space.  With some speakers, the piano will sound slightly different as its pitch moves up and down, or its size and placement within the recording space will seem to change.  With the Maggie, the pianoís tonal balance and the combination of the notesí attack, bloom, resonance, and decay were entirely consistent across the instrumentís range, as were its size and placement.

Great Expectation No.4: Pure, articulate upper bass and midrange; airy, detailed highs: The MG3.5/R is superb in these areas, but the MG3.6/R was probably just a bit better.  Vocals were treated well, with a natural mix of chest, throat, and mouth tones, but strings really showed off the Maggieís upper bass and midrange best.  One of my favorite albums is Franz Helmersonís performance of solo cello works by Bach, Hindemith, and Crumb (BIS BIS LP-65).  Listen carefully to some of the slower passages in Bachís Suite No.2, in particular.  When Helmerson draws his bow across the string, I could hear the combination of sounds that were layered on each other to build each note.  The bowís initial contact, the resinous draw across the string, the stringís vibration, and, finally, the resonance building within and expanding out from the celloís body—all were exactly right in their balance and timing.  The result was a beautiful, almost heartbreakingly pure cello sound.



The MG3.6/Rís highs were nothing short of superb.  Piccolos were pure and clear, and maintained all their detail and sharp metallic cut all the way to the top of their range—and without getting hard or steely.  Solo violins were delicate and sweet, and high, massed violin crescendos had tremendous power and presence, but never crossed over into a hard, unnatural screech.  Cymbals are perhaps the best example, and the Maggie unfailingly had exactly the right balance: a rich, bell-like tone at the center, a palpable sense of waves of overtones emanating from the cymbalsí vibration, and, surrounding it all, a cloud of shimmer that seemed to permeate the entire space.

Great Expectation No.5: Dynamics!  From the subtlest micro-shading to the most explosive crescendo: Another longtime Magnepan bugaboo has been the need to play them loud to get a sense of realism.  The MG3.5/R and 1.6 were dramatic improvements over the previous models in their ability to reproduce large dynamic transients, but they still lacked the nth degree of resolution at the pppp end of the scale.  With the MG3.6/R, Magnepan seems to have eradicated this shortcoming.  Big crescendos were startling in their power, as were drum sets, particularly rimshots and toms.

At the other end of the scale, when the 3.6/Rs were paired with a muscle amp like the Classé monoblocks, they did a first-rate job of capturing microdynamic shadings.  On ďWhat a Difírence A Day Made.Ē from her Never Make Your Move Too Soon (Concord Jazz CCD-4147), Ernestine Anderson often floats the faintest, subtlest traces of vibrato on the very last breath of notes.  A lot of speakers, even some excellent dynamic models, canít capture that vibrato, but the 3.6/R did it beautifully.  Iíd often find myself holding my breath, just to make sure I didnít miss these delicate whispers.

Great Expectation No.6: Transparency: no opacity, no texture: For all their great strengths, Magnepan speakers have always suffered from a slight opacity.  The MG3.5/R and 1.6/QR were spectacular advancements in this regard, retaining only faint vestiges of a slightly filmy texture.  The 3.6/R is another big step in this direction, its transparency rivaling that of the best cone-type speakers Iíve heard.  This showed up in added purity through the midrange and upper midrange, slightly more complex harmonic mixes, and improved dimensionality.  The improved transparency was most apparent, perhaps, in how it helped expand and remove congestion in the back half of the soundstage.  The MG3.6/R was the best Iíve heard at opening up the spaces between trumpets, for example, and maintaining their size and detail.

The flip side of the 3.6/Rís transparency, however, was that it wasnít nearly as forgiving as earlier Magnepans.  Even the 3.5 wouldnít penalize a listener too much for their choice of upstream components, as long as they included a clean, powerful amplifier.  With the 3.6/R, I had to be a lot more careful.  My Ultech and Parasound CD players just didnít cut it, for example, and until the SimAudio and Oracle players showed up, I listened almost exclusively to vinyl—and had to scrupulously level, adjust, tweak, and warm up my TNT.  Selecting cables became an agonizing series of trials and tradeoffs.  Even my beloved VTL Ichibans became a limiting factor, ironically contributing a touch of haze of their own.  Ditto the Mark Levinson No.20.6s, which had a slightly dark, liquid presence.  It was only when I installed the Classé CAM-350 monos and optimized the setup around them that I truly appreciated the MG3.6/Rís transparency.

Okay, Iím a Magnepan guy.  Iíve owned several pairs over the years, and I absolutely flipped over the MG3.5/R.  In these pages, I pronounced the 1.6/QR ďone of the great audio bargainsĒ.  Nowhere were expectations for the MG3.6/R higher than in my listening room.  And, point by point, the 3.6/R delivered.

The 3.6/R builds on the great strengths of the 3.5/R, and successfully incorporates some of the magical touches that transformed the 1.6/QR into such a small wonder.  Its re-creation of the original soundstage and recording environment are incredible, and with the latest improvements, its dynamics, resolution, and transparency approach those of the very best speakers Iíve heard.

The 3.6/R does need to be driven by a good, powerful amplifier to sound its best, and will clearly reveal the weaknesses of upstream components.  But when all the pieces are in place, itís magic.

The 3.6/R is unquestionably better than the 3.5/R—stronger, more articulate, and better integrated.  Itís not a quantum step, though, so 3.5/R owners neednít feel the need to immediately dump their speakers in the ďgarage saleĒ pile and upgrade.  Similarly, the 3.6/R is a substantially better speaker than the 1.6/QR, in every way.  Itís flatter, more refined, much better at the frequency extremes—the list goes on.  However, if bucks are really, really tight, I suggest you opt for the 1.6/QR, invest the difference in upgrades elsewhere in the system, and not lose any sleep about it.

Taken on its own, however, the Magnepan Magneplanar MG3.6/R is a sensational speaker, and, at $3750/pair, very reasonably priced.  In some respects itís the best speaker Iíve heard, period.  Even in the areas where itís perhaps not the very best, itís awfully close—even when the very best is several times more expensive.  Some speakers I admire, some I like...the Magnepan MG3.6/R, I think Iíll keep.  Very highly recommended!



Sidebar 1: Specifications

Magneplanar MG3.6/R



Sidebar 2: Associated Equipment

Analog source: VPI TNT Mk.IV turntable, JMW Memorial tonearm, Grado Reference cartridge.
Digital source: SimAudio Moon Eclipse, Oracle CD players.
Preamplifier: VAC CPA1 Mk.III.
Power amplifiers: VAC Renaissance 70/70, VTL Ichiban, Mark Levinson No.20.6, Classé CAM-350.
Cables: Nirvana S-X, Kimber Bi-Focal XL, Synergistic Research Designerís Reference.
Accessories: Bright Star Rack of Gibraltar and isolation systems; Tiptoes; PAC Super IDOS; MIT Z System and Z Center, Nirvana AC systems; Synergistic Research A/C, Reference Master Couplers; VPI 16.5 record cleaner, Decca/Hunt record brush, Sumiko Fluxbuster, Dennessen Soundtractor, Shure stylus-pressure gauge, Immedia Needle Nektar stylus-cleaning fluid; Nordost ECO3 and Music Fidelity DiskSolution CD treatments; Sheffield/XLO Test & Burn-in CD, StereophileTest CD 1 and 2; EchoBuster and ASC room-treatment products.—Brian Damkroger



Sidebar 3: Measurements

The Magnepanís estimated voltage sensitivity was on the low side, at 83.5dB(B)/2.83V/m.  However, in a typical room the speakerís quasi-line-source vertical dispersion should make it sound a little louder than might otherwise be expected.  (The in-room loudness of a true line source falls off in a linear manner with distance, rather than as the square of the distance, as is the case with a point source.) But it should be noted that BD did need a good beefy amplifier to drive the Maggies to useful levels.

The speakerís impedance (fig.1) approximates a resistive load of around 4 ohms over much of the audioband.  However, there is a slight magnitude peak centered at 1.6kHz, due to the crossover between the ribbon and the midrange diaphragm.  The minimum value is 3.3 ohms at 10kHz, which is not going to be problem for any good amplifier to drive, while the increasingly positive electrical phase angle at the top of the audioband is, I assume, due to the residual inductance of the ribbon driver.  There is a small wrinkle in the trace between 50Hz and 60Hz, which is probably due to the tuning of the woofer diaphragm.

Fig.1 Magnepan MG3.6/R, electrical impedance (solid) and phase (dashed)(2 ohms/vertical div.)

This can be seen to the left of fig.2, as the big peak in the mid-bass.  This is a nearfield measurement, which will exaggerate the behavior of the bass panel [see the letters at the end of this reprint—Ed.].  This does not necessarily mean the speaker will ďboomĒ in an out-of-control manner—BD commented that the MG3.6/Rís low frequencies sounded ďpowerful, clean, and articulateĒ down to about 35Hz in his room—but such measured bass behavior does appear to be characteristic of panel speakers.  Yet the midrange diaphragm does not have a response peak apparent.  It neatly covers the 200Hz to 1.2kHz region, with relatively steep rolloffs above and below that bandpass.  From this graph, the ribbon tweeter seems both to be set a little low in level, and comes in rather high in frequency.  I imagine that the narrow peaks and dips in its response, are due to local interference effects.  They should therefore not have any subjective consequences.

Fig.2 Magnepan MG3.6/R, acoustic crossover on tweeter axis 36" from the floor at 50", corrected for microphone response, with the nearfield woofer and midrange responses each plotted below 300Hz.

The Magnepanís overall response, measured on the ribbon axis 36" from the floor (ie, halfway up the ribbon) and averaged across a 30 degrees lateral window is shown in fig.3.  The microphone was at a 50" distance, which results in a significant proximity effect with such a physically large speaker.  This accounts for much of the downward response tilt evidenced between 200Hz and 2kHz in this graph.  The level mismatch between the midrange diaphragm and the ribbon tweeter is still evident, but I wonder how this will manifest itself at a normal listening distance.  (Circumstances dictate that I use a 50" microphone distance for my acoustic measurements.) The logistics of the magazineís relocation to New York meant that I could not perform in-room measurements in BDís listening environment, but I suspect that the MG3.6/Rís behavior will be better behaved in a room.

Fig.3 Magnepan MG3.6/R, anechoic response on-axis at 50", averaged across 30 degrees horizontal window and corrected for microphone response, with the complex sum of the nearfield woofer and midrange responses plotted below 300Hz.



Magnepan recommend that the speakers be used with the tweeters on the outside edges.  There is more treble energy apparent laterally off-axis on the woofer side of the panel (fig.4).  Vertically (fig.5), as expected from a speaker that behaves to some extent as a line source, the balance doesnít change over a significant listening-height range.

Fig.4 Magnepan MG3.6/R, lateral response family at 50", from back to front: response 90 degrees-5 degrees off-axis on woofer side, reference response, response 5 degrees-90 degrees off-axis on tweeter side.

Fig.5 Magnepan MG3.6/R, vertical response family at 50", normalized to response on middle of tweeter axis, from back to front: differences in response 10 degrees-5 degrees above axis, reference response, differences in response 5 degrees-15 degrees below axis.

In the time domain, the Magnepansí step response (fig.6) indicates that the ribbon tweeter and woofer diaphragm are connected in positive acoustic polarity, the midrange diaphragm in negative polarity.  Because the drive-units are mounted side-by-side, this will swing the main lobes to the woofer side of the panel, hence Magnepanís placement instruction.  The cumulative spectral-decay plot (fig.7) shows an initially clean decay in the treble, but then some hashy behavior.  This, I believe, is not due to the presence of resonances but to early reflections from the physically large radiating areas.

Fig.6 Magnepan MG3.6/R, on-axis step response at 50" (5ms time window, 30kHz bandwidth).

Fig.7 Magnepan MG3.6/R, cumulative spectral-decay plot at 50" (0.15ms risetime).

As I have written before in these pages, measuring physically large speakers with in-room quasi-anechoic techniques is in some ways a fruitless task.  The usual assumption, that the measuring microphone is very much farther away than the largest dimension of the speaker being measured, is clearly wrong.  Yet without access to a large anechoic chamber costing many hundreds of thousands of dollars, in-room measurement techniques are all we have to rely on.—John Atkinson



A letter in response appeared in November 2000:

Stiffen those Maggies

Editor: I read the August 2000 review of the Magnepan MG3.6/R with great interest.  It was a very well-written, informative article.  Brian Damkroger is among my favorite writers at Stereophile and has similar tastes in speakers to boot.  However, I do have a couple of questions.

It was stated that Magnepan recommends setting the speakers up with the tweeters to their outside edges (or at least so that the tweeters are farther from the listener than the woofers).  It was never stated if Brian ended up using them this way, or if he experimented with having the tweeters inboard.

Also—my main reason for writing—doesnít anyone make aftermarket stands for Magnepan speakers?  I know how much importance Stereophile generally places on speaker stands, yet there was not even any mention of putting cones under the MG3.6/Rsí feet to stabilize these mammoth speakers.  I canít see how the Maggies would not improve greatly from the use of some good-quality stands or, at the very minimum, a set of cones.

I have an old set of MGIIIAs for which I am fabricating a set of custom stands that will fasten to the standard bolt-hole pattern at the bottom, and will also fasten about 16" from the top.  (The speakers can be easily moved 4-5" front to back at their tops; I would imagine the MG3.6/Rs are not much different.) The base will have 95-100 lbs of sand in it.—Grant VanderMye

Thanks for your kind words, Mr. VanderMye.  To answer your questions:

I tried the Maggies in a wide range of configurations, including many with the tweeters to the inside.  In my room, I could space the speakers farther apart, toe them in more, and still end up with the tweeters farther from my head.  My preferred setup was the one I described, however.  I generally got a more even, more expansive image with the tweeters outboard.  Having them to the inside gave me a more triangular image, with the rear corners becoming murky and shrunken.

If, on the other hand, youíre dealing with a long, narrow room, and have the speakers firing down the long axis, my experience with other Maggies suggests that you have to be concerned about proximity to the side wall.  In those setups, having the tweeter outside might cause the image to move forward along the wall, ďwrapping aroundĒ the listener a bit.  I found (in a shoebox room) that I could sometimes get a better balance of image width and depth with the tweeters inboard.  I would have included more on setup in the review, but it was already longer than it should have been.

I understand and share your concern about speaker stands.  (Iím not positive, but I think that Sound Anchors may make stands for the Maggies.) Itís certainly disconcerting to have this huge, flimsy panel waving in the breeze, after weíve all convinced ourselves that super-rigid coupling is the way to go.

The Maggiesí size may work a bit in their favor, because the panels donít have as much displacement as a cone driver.  I did play around with Tiptoes with the 3.6/R and several other Maggies, but it never seemed to make that much difference in the sound.  The floor/stand coupling isnít the weak link, in my opinion, but rather the stand/panel coupling, the rigidity of the stand itself, and the rigidity of the panel itself.  Just putting the stands on cones—even bolting the stands onto cones—doesnít do much.

I didnít want to do anything beyond that, because my practice (and Stereophileís policy) is to test products in a completely unmodified state.  Back when I had Infinity RS1bs, I did make modified bases for the midrange-tweeter panels that had a triangulated aluminum brace coming about two-thirds of the way up the speakerís back.  But the Infinityís structure was far more rigid than the Maggieís, so the stand and coupling were easy to fix.  I plan to keep the 3.6/Rs; maybe Iíll play around with something like your attachment 16" from the top.Brian Damkroger



Comments on Nearfield Measurements of Panel Speakers: (The following letters were received but were not published.)

Siegfied Linkwitz comments

Editor: The review of the Magneplanar MG3.6/R in the August Stereophile caught my attention.  I am a proponent of open-baffle speakers because of their room acoustic advantages and the absence of sound coloring boxes.  So I looked with great interest at figs.2 & 3 on page 89 showing individual driver frequency responses and their summation.

The nearfield measurements of woofer and midrange in fig.2, presumably taken only an inch or so from the driver surface, are a valid set of data.  You also could have measured the tweeter at such close range and obtained useful information.  Where things fall apart is in fig.3 when you form the complex sum of nearfield measurements and the 50" tweeter ďfarfieldĒ measurement.  This curve does not represent the frequency response a listener might experience at any distance and is therefore extremely misleading.

The nearfield frequency response of an acoustic source is only proportional to its farfield response if the source is small, ie, omnidirectional, and if it is in free-space.  Summing a driver diameter corrected woofer nearfield response to a farfield midrange response works for a small monitor on a stand, but already has errors when the speaker is larger and the woofer is close to the floor—when the conditions move away from free-space or anechoic.

The Magneplanar is clearly not a point source and, being open-baffle, it has an acoustic short circuit between front and back.  This causes a 6 dB/octave low-frequency roll-off in the farfield response.  So from all open baffle nearfield measurements you have to subtract first a 6dB/octave (= 20dB/decade) slope before you can sum the data with other farfield measurements.  When you apply this correction to the MG3.6 woofer response you see that it flattens from 400Hz to 60Hz and shows a peak at 47Hz.  Similarly the midrange has to be corrected before you can use it for the composite response.  The actual room response is still different from this composite, though, primarily due to the effect of the floor on woofer radiation.

You might consider to add a measurement taken with a 50ms time window at your listening position, spatially averaged and half-octave smoothed to include the room.  I think as a measurement that allows true comparison between speakers, this would be more useful than the composite data that are correct only in a few special cases.

I hope this letter helps your readers to understand the difficulties in describing a loudspeaker by measurements.—Siegfried Linkwitz, Corte Madera, CA,

Mike Gough comments

Editor: I read with interest the measurement section of the Magnepan Magneplanar MG3.6/R review in the August 2000 Stereophile, specifically John Atkinsonís comments on the midbass peak of the nearfield response.  The following may be of interest.

Some years ago, while designing the SCM8 dipole surround speaker (the triangular one) for B&Wís original THX Home Theatre System, I was discussing with Quadís Peter Walker the problems of coping with the bass roll-off imposed by front-to-back cancellation of dipole designs.  I was having a problem meeting the (then) THX bass extension with such a small enclosure, but did not want to revert to monopole in the bass (as so many do).

Peter told me of a technique he used on the Quad electrostatics, which I was ashamed I hadnít also thought of, which was to engineer an underdamped bass alignment.  That gave a basically rising response with decreasing frequency down to the nominal cut-off frequency, which compensates the roll-off due to dipole cancellation.  This underdamped characteristic, of course, shows up in a nearfield measurement, but not in the far field.  It is not apparent in the midrange panel because it is not needed.  The dipole cancellation starts at a frequency defined by the smallest dimension of the panel and this is the same for all sections in a common panel size.  The midrange panel operates above this frequency.

So such a nearfield peak is often a deliberate part of the design of dipoles (of which panel speakers are an example).  Mind you, both Peter and I went for much more modest peaks.  The dipole imposes an extra roll-off rate of 6dB/octave.  You can add a second-order Q=1 to a first-order at the same frequency to get close to a third-order Butterworth or, for a more extended ďflatĒ response; a second-order Q=2 added to a first-order at twice the frequency gives something akin to a Tshebychev with a 1dB ripple.  The Magnepan peak does seem a little excessive, but it all depends how it interacts with the modes of the listening room.

This technique does open the debate as to what the ear actually hears.  A Q of 2 has a pretty abysmal transient response and the question is whether the dipole ďequalisationĒ ameliorates that effect in the total response.  As both mechanisms are minimum-phase, I suspect and believe that that indeed happens.  As it is ultimately third-order, though, the response will have an inferior low-level transient behaviour to a well-adjusted second-order.  It should have some similarity to the series C (capacitor) closed-box alignments we used while I was at KEF with Laurie Fincham.  There the -3dB point was lowered by putting a capacitor in series with an acoustic alignment with Q of 1.  In those days we wanted to protect speakers from turntable rumble.—Mike Gough, Senior Product Manager, B&W Loudspeakers Ltd.