Tuesday, April 29, 2014

Mundorf AMT 164UM2.1 Tweeter Review

We all started listening to sound by hearing the things around us. You learned to recognize the voice of your parents and then your relatives; friends come later and followed by other personal interests. You knew a person by the unique sound of their voice even over distorted mechanical devices like poor cell phone connections or across a crowded room. You can easily describe how your father's voice sounds compared to your mother's (higher pitched, nasal, faster, etc.). You then learned to apply this talent to the way other things sounded like the barks of dogs, tweets of birds, the wind, on and on. And after you have mastered putting into words what all of these sounds are like, you find that one stereo sounds different from another.

You are able to intuitively know that the stereo system you are listening to now sounds better or worse than another you have heard in the past. We typically struggle to find the right words and describe what it is we have just subjectively observed. But as your experience with the way stereos characteristically sound, so does your vocabulary.

Much like acquiring a preference for chocolate over vanilla or coffee over tea, with time we all develop certain biases toward certain types of sonic attributes. Some people need more bass, others need it louder, and still others clarity or uniformity. All are noble attributes and some are difficult to achieve within the same stereo system. For example, if someone leans toward neutrality, they generally shy away from horn speakers because most (not all) tend to screech. And if someone else leans towards a wide dynamic range, then horns are (traditionally) the only way to achieve such intense dynamics.

It is difficult to find one thing, especially with loudspeakers, that does everything right or at least right-enough to satisfy all of the words you have developed in your ever-growing high-end vocabulary. Some image well, others have really good phase coherency, others extended high frequency response, and still others uncanny uniform sound pressure across a wide frequency range. You gradually develop a sort of “must have” list that evaluates a piece of audio gear before you to even entertain its purchase. You eventually determine what you like and you know what systems will and will not deliver. For example, after enough time auditioning gear you realize that a 4” woofer cannot produce the low-distortion deep bass you desire. While an array of 8 or 16 of them can go pretty deep, their characteristic distortion due to long-cone excursions tell you just by looking at them that it will not sound the way you desire.

These “common sense” prerequisites is therefore a list of minimum attributes any piece of audio gear must be capable of delivering in order for you to know that it sounds better or worse than what you prefer. If any speaker does not deliver on this list of prerequisites, then your opinion of them is one of – well let’s say that you wouldn’t bring them home with you even if the price was right.

I personally prefer the sound of planar loudspeakers over that of most others (dynamic, horn, etc.). In this category there are two basic types: direct radiators and bipolar systems. Direct radiators are box-styled speakers where the sound emanates from the front of the driver (the sound from the rear is damped within the box). Bipolar (aka dipole) speakers use both the front and rear of the driver to create sound. Which is best has a lot to do with the room in which they are played along with your personal must-have list, but I digress here. I want to stay focused on driver technologies and your personal must-have list rather than the influence rooms have on this list.

I listen to a lot of esoteric toys, most of which sound good and some even sound great. But new ideas and revived technologies always intrigue me, especially when someone finds a new slant on an old idea. Such is the case in this review. Dr. Oskar Heil developed a new type of driver called the Air Motion Transformer (AMT for short) way back in the early 1970s. His idea was instead of pushing the air with a circular diaphragm and voice coil as typical dynamic drivers did to make sound, his driver pinched the air between a zig-zag membrane that looked something akin to the folds of an accordion’s bellows (a pleated diaphragm). A massive magnet structure and flux focusing fingers were required to create a strong-enough field over its pleated-diaphragm for this dipole driver to operate properly. BTW, this type of diaphragm moves air roughly five times more efficiently than does the flat diaphragm of a dynamic driver.

Heil AMT Driver

Recently, there has been resurgence in interest in the lowly AMT driver by several well-known driver manufacturers. Known primarily for its work in state-of-the-art passive products (capacitors, inductors, resistors, etc.), the Mundorf corporation has expanded its prowess into the Pro Monitor loudspeaker realm offering a well-designed monopole variant of the traditional Heil dipole AMT. This series of drivers have sensitivities of 97-98dB/W/m requiring typically ¼th the amplifier power to create the same sound pressure levels of typical high-efficiency dynamic drivers. And their drivers are physically far smaller than their earlier Heil counterparts. How do they do this? Magic? NO, NO NO! Science, and really good science at that.

First, Mudorf uses high-strength neodymium magnets in these drivers. To give you an idea of how strong these magnets are I positioned two of the these drivers near each other on my dining room table. I set the first 2.4 pound (1100g) driver face down on the table and as I set down the second one, the first moved away scooting it toward the center of the table. This surprised me. The closest I could slide the second to the first before it moved was about an inch, and then it slid another two inches across the table top before stopping. These magnets are so strong, I feared the magnetic media in my camera may become ruined if approaching it too closely to them so I immediately backed off my camera and gave these drivers the full respect they deserved.

164UM2.1 Unpacked

The Bozak woofers have a rated sensitivity of about 96dB/W/m so at a rated sensitivity of 98dB/W/m these tweeters seemed to be an excellent match, something incredibly difficult to find with conventional dynamic or planar tweeters. Plus the Bozak B-200Y dynamic tweeters originally used in these speakers crossed over at a very low 2.4KHz and few drivers of any kind regardless of design could operate in this wide of a band with this required sensitivity. Pro audio drivers do have this kind of sensitivity but most pro audio drivers were definitely not high-end audio drivers, until now.

I initially set the drivers up next to the satellites with alligator-clip jumper cables and waited for them to burn in. I disconnected the existing tweeter (a Peavey RD1.6) and allowed the Mundorf to operate from the same 8-ohm designed third-order crossover point of 2.34KHz with no alterations of any kind. The two graphs below show the initial differences measured between these two tweeters. As a reminder, focus only on the region above 200Hz since frequencies below this region are not even close to being accurate from my hand-held cell-phone RTA.

Peavey RD1.6 (L) and Mundorf AMT 164UM2.1 (R), First Measurements

Although the left-hand RTA graph of the Peavey ‘looks’ pretty good, in reality it sounded a little off (looks can often be deceiving as we already know). The Peavey has a lot going for it but to consider it as a high-end audio driver may not be the best choice of words. It IS a really good tweeter with really high sensitivity and it does to a very good job especially below 8KHz. However, it is quite soft in the top octave, something I was hoping to resolve with the introduction of the AMT's design. Time will tell if after the burn-in period if the driver’s overall response will improve.

The impedance compensation for the tweeter’s T-pad attenuator was also designed for low-reflectivity at 8-ohms in/out. Checking with the published specifications, I noticed something unusual: the table claimed an Zmin=8.7 ohms and the graph's scale clearly showed a much lower value (Zmin=6.8 ohms). A quick email to Mundorf corrected the misunderstanding explaining that the impedance chart was incorrect and that it would be corrected straightaway. I was also told that this driver's impedance at 2.4KHz was 8.8 ohms.

Since the impedance of the Mundorf is not matched to the existing crossover network, additional compensation to correct for this variation was implemented later in the week. Mundorf also recommended a 60 hour burn-in before properly evaluating their performance and I complied with their recommendation waiting over a full week before conducting serious listening tests and other measurements. But the character of the drivers changed wildly over this time and I wanted to report my subjective impressions so you can understand what you could expect to hear when you purchase them for yourself.

Mundorf AMT Drivers Temporarily Mounted in my Satellites

Day 1 – Amazing. Everything I did not like about the Peaveys was dashed into oblivion by these superb drivers. Even though I knew they would get better after the burn-in period, I was immediately captivated by their full and neutral sound. The best way to describe it is that they revealed more of what was included in the signal sent to them – or not as the case may be. In other words, these drivers are no longer the weak link in my audio chain. When I turned them on, my wife was in the next room fiddling in the kitchen and she dropped what she was doing and said, “whatever you did, this is good!” The longer I listened to these AMT tweeters in my reference music, the more I am a little disappointed at their performance above 7KHz despite the fact that everything below 7KHz is really clean. Patience…

Days 2-5 – Not much to note. I am getting used to how the sonic contour has changed by playing familiar tracks over and over again in complete disbelief. I hear more now than I have ever heard from any other speaker regardless of cost, size, or connected equipment. While I am reserving my impressions and still a bit disappointed in the top octave, I cannot wait for them to fully break in. However, I am skeptical that they could possibly sound even better than they now do. I have never heard a driver make a significant change to the point where a day-night difference could ever be experienced. BTW, the top octave is still a bit dull lacking the unusual transparency the old Peavey’s were capable of revealing. I adjusted the attenuation levels in both the midrange and tweeter at the end of Day 2 to better match their sensitivities and an adjustment to the impedance of the AMT to a nominal 8 ohms.

Day 6 – Something happened. Here I am sitting down writing this review in the early morning hours and all of a sudden the audio world I was listening to “came to life.” It was as sudden as an earth quake in Los Angeles and equally as moving. The top octave just opened up and the clarity and focus above 7KHz completely changed my opinion of the overall performance of these tweeters. What I once perceived to be a really good midrange driver has now been revised upwards in more ways than one. Truly, nothing has changed in the past three days in my adjustments to the crossover network. I was listening to track 10 of Randy Star’s Dreaming Didgeridoo when the change snapped. This is way cool. I cannot wait for my wife to get out of bed and see if she hears this same difference...she did.

Day 10 - Still breaking in. The best way to describe the sound of this tweeter is "unique" where it possesses the virtues of the best dynamic dome tweeters while approaching characteristics of the finest ribbons. "Balanced" is another word that comes quickly to mind where the vast majority of dynamic drivers do not fall into this category. These drivers lean more toward the sound of a dynamic driver than a ribbon so dynamic fans will perceive this as an asset. But those of you who love the sound of bipolar ribbons may also be swayed, just not as easily. There is a magic that these AMT drivers flaunt that bridges the dynamic-ribbon gap.

For example, the newest 192/24 Chesky release from HD Tracks titled Wake Up Your Ears Sampler is a good way to hear what it is that I am saying. Track 5 "Stank" begins with a drum solo whose echoes reveal the size of the recording studio. As these echoes fade, their reverberations blend in the far corners of both channels. With both dynamic and ribbon drivers, emphasis is unnaturally thrust towards the upper resonances of the cymbals ignoring the texture of the drum skins. With the AMT driver, the emphasis is placed on inner detailing of the drums, skins, sticks, and accompanying percussion. Here, your focus is drawn into the interpretation of the artist rather than the enhancing effects of the speakers.

In track 6 of this same piece, the flute has that sweet hollow sound capturing its harmonics in perfect balance to its fundamental tones. Here, a neutral character swells from the folds of the AMT that does so with an effortless, unstrained quality unlike that of even the best light-weight dynamic drivers. My old ribbons just could not do this as smoothly but would rather favor one band and perform just average in another.

Most drivers sound sweet when listening to them loud but the sign of a superior driver is how good it sounds at many different volume levels. These Mundorf AMT drivers maintain acoustic linearity through all but the quietest levels making them even more listenable than anything else I have ever auditioned. For example track 10 of this same piece features a throaty saxophone with its well-recorded bell resonances. Turning the volume up/down accurately retains the timbre of this instrument with essentially no change to its texture. Even though the various sonic intensities shift in the classic Fletcher-Munson pattern, the character of the instrument is unchanged. In other words, it still sounds exactly like the sax at a high volume level as it did at a lower level, quite an achievement to say the least.

With this much change in the sound of this driver in just over 10 days (about 60 hours of playtime), I was interested in observing what the measurable differences would be so I ran a quick RTA measurement to see. Below are the "out of the box" (left) and the 60-hour break-in (right) RTA measurements. Remember that I matched the midrange and tweeter levels and changed the design of the T-pad on the tweeter to nominalize it for the 8-ohm design of the crossover network.

As you can see, there is a pretty dramatic difference in the system's response after only 60 hours of burn-in time with no other changes to the system. Everything above 300Hz is far smoother and the top octave is starting to show definite signs of life. People have often reported that their speakers sound different within a short period of time after their purchase and these graphs confirm that subjective impression.

The main asset of these drivers is that they encourage you to listen deep into the music. As it is when you hear a fine set of electrostatic loudspeakers, you walk away with a feeling that another layer of grunge has been removed from the audio chain between the source recording and your ears. It is uncanny how these AMTs do this without any of the high voltages, esoteric films, and atypical design constraints otherwise associated with an electrostatic speaker. The fluid response lacking the usual peaks and dips of dynamic drivers is also quite ear catching. Any minor non-uniform peaks and dips for some reason seem far less objectionable, something that I just cannot rationally explain.

And as it is with high-sensitivity drivers, these AMTs produce an exceptional dynamic range that can frankly scare the pants off of you from tracks you only thought you heard at their best. Similar in resolve to pro-audio's extreme sensitivity, if you like your music loud, these drivers will definitely deliver revealing how much headroom your amplifier has in reserve. At the other extreme, low-level reverberations are no longer lost in the noise floor again with a similar uniformity that I have frankly never heard before from any kind of tweeter.

My enthusiasm towards the neutral sound of these drivers cannot be understated. Simply bolting this replacement into your existing system will - without a doubt - point to the shortcomings in the rest of your audio gear. In my own system, I now question the choices for many things, especially the interconnect cables. No longer will you wonder if your speakers are the limiting factor in your system; you can now search for that optimum combination of interconnects and speaker wires to allow them to reveal its full potential.

I swapped my reference interconnects with a new series I am investigating using CAT6e shielded wire. Leveraging off of the tiny plastic insulator that maintains uniform spacing between conductor pairs and this wire's inherent extended bandwidth, I have swapped wiring configurations within the cable finding one that suits my ears. Below are two RTA measurements showing the effect of changing these interconnects on the performance of the system. As you can see, there is clearly another step up in helping these tweeters perform to their full potential.

As found in my other reviews, a second level of break-in occurs typically around 200 hours of play time with gradual improvements along the way and the Mundorf AMTs were no different. After 200 hours, the definition just above the crossover point bloomed. There is even more depth and inner detailing than at the 60-hour mark with more roundness of woodwinds and the hollowness of brass bells and crystal bowls.

I questioned my amplifier’s ability to reproduce the upper octave since the published HF data and the perceived HF response did not match up. So, I replaced the MC2100 with a newly modified Onkyo M-504. The change was a positive one revealing an apparent compromise in transient response of the McIntosh output autoformers.

For example, in the artful 2010 album “Yoga Sessions: Masood” by artist Masood Ali Khan, for track 1 Atman Kama he plays a steel drum. The sound of the drumstick listened trough the McIntosh was clunky with unnatural hammer noises as the stick contacted the drum surface. With the Onkyo, these sounds were far more natural creating a mellow smoothness more realistically representing the sound of this wonderful instrument. The Mundorf AMTs were then permitted to reveal the timbre of the drum surface itself thereby eliminating another level of grunge between your ears and the music. The post-note tap of the drumstick also sounded far more realistic through the Onkyo with much better damping.

With stringed instruments such as violins you hear the sound of the string itself along with the character of the body materials and quality of its design. Cheap violins sound thin and good ones sound robust and full. It is the low-level detail that these AMT drivers do very well that permits you to hear more of this type of information even from mediocre MP3 sources.

Compared to most individuals I am a person who does not enjoy extremely loud music so I rarely listen at moderate listening levels. In other words, even when demonstrating my system to someone will I rarely “crank it up.” However, my wife is an entirely different story and when she watches an action movie (like Avatar or Salt), she appreciates the full dynamics of a good audio system coupled with an equally refined video system. With the Onkyo in place, she cranked up the system to her usual intense level. To her surprise and delight, the Onkyo produced about 50% more power than the McIntosh increasing the system’s overall dynamics. Apparently, this is what the Mundorf AMT’s needed. That is, to help them properly break in they needed to be played loud.

After two such movies, I noticed another level of maturity arise from these already stellar drivers. The positional stability of the instruments, the ambience detailing in the far corners of the sound stage, the three-dimensional sound stage size (height-width-depth), and the ability of the drivers to reproduce the subtle nuances and low-level details of any instrument or vocal were all completely transformed. After four such movies this maturation evolved even more so extending this same enhanced resolution to very low volume levels (peaks at about 0.001 Watts).

For example, track 5 “Stank” of the HD Tracks Wake Up Your Ears Sampler begins with a drum solo whose echoes reveal the size of the recording studio. As stated earlier, while these echoes fade, their reverberations blend in the far corners of both channels. Added to this appreciation is a smoother decay of this echo revealing more of the character of the environment and of the production-enhanced reverberant effect. A previously unnoticed bloom of all of these percussive instruments is clearly enhanced.

As previously stated in track 6 of this same piece, the flute has that sweet hollow sound capturing its harmonics in perfect balance to its fundamental tones. Now, I can hear more of the air blowing into the lip prior to the formation of the notes (somewhat reminiscent of Jethro Tull’s whistling techniques).


The Mundorf AMT 164UM2.1 is a wonderful driver capable of faithfully delivering on everything except the highest band of the top octave. With a transient response better than its dynamic driver counterparts and 1-Watt sensitivity usually found only with fine horns, this driver should be seriously considered for your next high-end speaker project. Well suited for high-quality tube gear, you may find that your output autoformers are not capable of delivering the best possible signal to these superior drivers.

The image stability over a wide range of volume levels puts this driver in a class of its own. No longer do you need to make your ears bleed to hear nuances in your best software sources. You will find yourself listening to familiar pieces hearing more than you ever have before and with greater dynamics than previously experienced.

These drivers are worth their weight in gold since what they do, they do very, very well. I am known for my understatement in comments about things but with these drivers I struggle to find the proper superlatives to praise their character. This is one of the few times that words escape me. Vocals are glorious and complex instruments are reproduced with uncanny accuracy and believability. If there is a shortcoming about their performance, it would be in the upper octave (that is, the lack of uniformity in the top octave and a slight loss in speed compared to a fine ribbon). Everywhere else they are rock solid blending the world of superb dynamic drivers with really fine electrostatics. What a feat to achieve!

With proper crossover network design, these drivers will make your heart beat for joy faster than what you thought possible. In my crossover network I used all Mundorf Supreme capacitors bypassed with a 0.01uF/200V Vishay MKP. (I would expect that the Mundorf Silver-Gold-Oil versions would permit them to sound even better.) Also in this photo is an impedance matching T-pad (five resistors on the LH side) that properly terminates the driver to the network and achieves the desired attenuation (-6.56dB). Use this calculator to determine proper resistor values for your required attenuation.

Tweeter Third-order Network, Fc=2,358Hz

The schematic for the above 8-ohm crossover network is shown next. The optimum values are C1=5.62uF and C2=16.87uF; actual values are C1=5.71uF (1.6% error) and C2=16.84uF (0.18% error). Also included is a HF compensation circuit that boosts the top octave by circumventing most of the attenuation. This compensation circuit does extend the response at the expense of shifting the phase (it measures better but may not sound as good). Values are shown for what I had on hand, not the optimum design (may sound better with a smaller total bypass capacitance). BTW, changing the brand/style of capacitors in this compensation circuit will impact how objectionable the phase shift appears to be. C-SA=Clarity SA. MS=Mundorf Supreme. WW=Wire Wound. NI=Non-inductive.

Tweeter Crossover Network Components

I experimented with phasing of the driver just to see what effect it would have on the sound and RTA measurements. In its temporary mount, the inverted phasing measured a little worse at the crossover point but sounded the best overall. The AMT is temporarily surface-mounted on the face of the baffle board instead of counter-sinking it as is the mid-bass driver. The distance from the edge of the midrange driver is also not optimized (frames presently ¾” apart). This temporary mounting mechanically mismatches the planes and positions of the drivers causing phase differences at the crossover frequency. After finalizing the satellite box design and counter-sinking the AMT, this polarity-reversal will most likely change back to normal. In its current reversed-polarity configuration, the top-to-bottom sonic signature of complex-harmonic instruments became far more believable. Both measurements do not include the HF compensation circuit.

After about 300 hours of play time, the Mundorf AMT 164UM2.1tweeters are still revealing more to my ears with anything that I play. I recently bought the 24/192 version of Beethoven’s Sonata in C Major performed by pianist Hyperion Knight. Already intrigued by the outstanding performance, the piano was never exploited by my old Peavey ribbons. With the AMTs, more of the sound board of the piano can be heard and the resonances it creates. The difference is almost like listening to it with the lid opened and closed. There is so much more life, realism, and clarity revealed in the harmonics than I believed this recording possessed. What a treat!

The level of sophistication these drivers can achieve must be heard to be appreciated. Everything I thought I understood about any previous-favorite performance is now being relearned. Do yourself a favor: buy a pair of these and try them out. See for yourself if what I hear, you hear also.

My thanks go to Madisound and Mundorf for their coorperation in making this review possible.


Yours for higher fidelity,
Philip Rastocny 

I do not use ads in this blog to help support my efforts. If you like what you are reading, please remember to reciprocate, My newest title is called Where, oh Where did the Star of Bethlehem Go? It’s an astronomer’s look at what this celestial object may have been, who the "Wise Men" were, and where they came from. Written in an investigative journalism style, it targets one star that has never been considered before and builds a solid case for its candidacy.


My other titles include:

Copyright © 2015 by Philip Rastocny. All rights reserved.

No comments:

Post a Comment

To comment on this blog, you must first be a member. All comments are moderated.

Note: Only a member of this blog may post a comment.