Mark Levinson No.53 Reference monoblock power amplifier

Mark Levinson founded Mark Levinson Audio Systems in 1972, but sold it, and the right to market audio gear under his own name, to Madrigal Audio Laboratories, then owned by the late Sandy Berlin, in 1984. Harman International bought Madrigal in 1995. As well as Mark Levinson, Harman's Luxury Audio Group now also includes digital processing pioneer Lexicon, speaker manufacturer Revel, and JBL Synthesis. The Mark Levinson brand is now headquartered in Elkhart, Indiana, at the Crown Audio facility, another Harman-owned brand. The No.53 ($25,000 each; $50,000/pair) is Mark Levinson's first new Reference series monoblock since the No.33, way back in 1993, when Madrigal owned the company. Like other Mark Levinson products, it is manufactured at an independent facility in Massachusetts.

Although it has a switching output stage, the No.53 is not a class-D amplifier. While the original rationale for the class-D technology was its efficiency—lightweight, cool-running amps that could produce a great deal of power—the No.53 is big and heavy, resembling the No.33—and it runs warm even at idle. It weighs 135 lbs, with claimed outputs of 500W RMS into 8 ohms or 1000W into 4 ohms, 20Hz–20kHz, at no more than 0.1% total harmonic distortion.

What's wrong with class-D
Without going into too much tech detail, class-D, aka "switch mode," amplifiers are more efficient than class-A or -B designs because their output transistors are either full on or full off. Rather than produce a higher-voltage version of the input signal, a class-D amp produces a high-frequency pulsed rectangular waveform: its transistors alternately connect the output to the positive and negative supply rails, with no in-between state. If the width of each pulse can be made proportional to the input signal's instantaneous level, the power delivered to the speaker over time from this "pulse width modulated" (PWM) signal will be equal to that of a conventional amplifier. When this PWM signal is low-pass filtered, the theoretical result is the original signal at a much higher voltage, just as you get from a class-A/B amplifier.

Of course, producing those proportional pulses in the first place and then perfectly filtering the squarewave are no mean feats. What's more, the high-frequency pulse signal will "broadcast" RF energy at upward of 300kHz.

Mark Levinson's innovations
The No.53 uses what Mark Levinson calls a patented, multistage "very high speed switching amplifier technology," Interleaved Power Technology (IPT), which is claimed to offer significant advantages over prior switching-amp topologies.

In a conventional class-D amplifier, one output transistor connects the output to the positive voltage rail, another connects it to the negative voltage rail. It is very important that when one transistor is turned on, the other must be turned off and vice versa, otherwise the positive and negative voltage rails will be short-circuited and the amplifier will self-destruct. However, because one transistor will be turned off before the other is turned on, there will be a short period of "dead time" when neither is turned on. This is analogous to crossover distortion in a conventional class-B amplifier.

ML's engineers set out to develop a new PWM output stage designed to overcome these limitations. The transistors switching the positive and negative voltage rails to the load in Levinson's "class-I" topology are each connected to the load via a large air-cored inductor and to the opposite voltage rail via a diode. The result is a stable amplifier with an effectively doubled switching frequency. Modulating the duty cycle of the 500kHz switching frequency of each transistor results in a 1MHz PWM output signal but without producing anything like the usual amount of ultrasonic noise. So minimal are the ultrasonic artifacts in the No.53's output, it is claimed, that, rather than requiring a sonically degrading brick-wall filter, all that's needed to remove them is a simple notch filter.

The No.53 uses four interleaved class-I/IPT stages in a balanced bridge configuration, to produce an effective PWM switching frequency of 4MHz. This allows the No.53's signal bandwidth to be extended to 100kHz.

Four Subsections
The No.53's interior is divided into four sections: the single-ended and balanced inputs (SE signals are converted to balanced and remain so throughout the No.53), a modulation section, the amplifier itself, and the power supply. A six-layer printed-circuit board containing about 1500 parts incorporates the four isolated IPT modulator circuits, as well as the proprietary Link2 and MLNet system-control functions and the No.53's protection circuits. The output stage includes multiple high-voltage, high-current, high-frequency vertical MOSFETs and eight air-core (nonferrite) inductors said to be virtually immune from saturation at high current levels.

The power supply is at the base of the tower, and includes a large toroidal transformer, 188,000µF of capacitance in the main supply, and an additional 105,600µF of "local" capacitance. While one of the supposed disadvantages of amplifiers with switching output stages is a low damping factor, due to the output low-pass filter, ML states that the No.53 has a very high damping factor, which allows the speaker to see a "virtual short circuit path back to the amplifier" that limits the effects of back EMF and results in tight, deep bass.

The result is a powerful amplifier capable of enormous peak current delivery, "tremendous" headroom, operational stability, improved electrical and thermal efficiency, and substantially reduced mass.

Setup and Use
The No.53 is about as tall (20.9") as it is deep (20.4"), but only 8.4" wide, with carrying handles cleverly integrated into the curved front and rear faáades. Nonetheless, carrying one was not easy. Once they were in place, it was easy to make connections on their open, uncluttered rear panels. In addition to the RCA and XLR inputs, each panel has two sets of sturdy speaker terminals fitted with wide, flanged, easy-to-grip plastic nuts. There are also low-voltage mini-plug trigger inputs and outputs for signal-sensing turn-on, Link2 inputs and outputs, and an Ethernet port to allow you to connect the No.53s to your computer network. Doing so allows you to connect to the amplifier's internal Web page with Internet Explorer, where you can modify the network setup, pull down a window to select one of four levels of intensity for the front-panel display, get basic status information, and track system-related error messages. In the event of a problem, this page will also be used by Mark Levinson Customer Service.

The ML Net protocol allows you to control two or more network-capable Mark Levinson products simultaneously from their Ethernet ports by setting up master and slave units. If the system doesn't work right, make sure the master and slaves are "properly daisy-chained." A front-panel status LED blinks differently, depending on what it's trying to tell you.

Harman Luxury Audio Group
1718 W. Mishawaka Road
Elkhart, IN 46517
(888) 691-4171

tmsorosk's picture

Many may find Mike's less than stellar review a surprize , not me , I found his sonic discription about on par with mine , and thats coming from a guy thats on his third generation of Levinson amps.

Mark Levinson needs to go back to the drawing board on this one.

Will Levein's picture

I was looking forward to this review because my understanding is the No.53s have pretty much split-opinions between those that 'get-them' and those that don't. Obviously Mr Harley is in the latter camp & he hasn't exactly held back has he!

I myself am the very satisfied owner of the stereo No.532. Unlike the commentor above this was the first ML purchase I've ever made, and believe me, I compared it at length (both in terms of time & geography!) with all the expected contendors. It emerged far and away as my favourite: the best power amp I've heard in 20 years of listening.

I accept that the consensus is emerging that the No.53s are indeed disappointing, but what I would say is don't write off the 'lesser' amps in the range, because they perform superbly and offer what I think is good value for money too.

John Atkinson's picture

Larry Greenhill favorably reviewed the No.532H in the August 2011 issue of Stereophile: see

John Atkinson
Editor, Stereophile

Will Levein's picture

Of course I meant Mr Fremer, not Harley. I read too many Hi-Fi magazines!

Axiom05's picture

The technology used in the 532 and No. 53 amps is very different. They are very different amplifiers and one should have no expectations about one based upon the other.

tmsorosk's picture

Good point { Axiom05 } , the 53 's were Levenson's first attempt at class D , but Levinson will send you a bunch of imfo stating it's not class D if you question them on the technology . From there written description it sound's allot like class D to me , but I'm know expert .


It was Larry's review that put the nails in the coffin for the #532H for me . After a little over a year listening to it , I've found it to be as he described. Thanks again Larry.

JohnnyR's picture waited a whole year after LISTENING to the amp yourself THEN read a review and decided it was a crap product but only after letting someone else tell you it was? xD

"It was Larry's review that put the nails in the coffin for the #532H for me . After a little over a year listening to it , I've found it to be as he described. Thanks again Larry."

So much for trusting your own ears and judgement. LMAO

UpAnA's picture

It is not Class D. It is based on the idea of "interleaved Class I". They are fundamentally different, although both switching amplifier technology.

As they say, when all you have is a hammer, everything sort of looks like a nail. In audio, some folks will call switching amplifiers "Class D" and even assign the false moniker like "Digital Amps" to them.

tmsorosk's picture

You may want to read the whole post JohnnyR , and get off drugs.

JohnnyR's picture

Where did I misquote you? Get off drugs yourself. Don't blame me if you kept a "faulty" amp for over a year.

tmsorosk's picture

 " Putting the nails in the coffin "  Would mean it was the last step in a very long process of reading , researching , listening in many systems as well as a long term loan of the #532H in two of my own systems , listening to different music at different levels and trying many cords , cables etc over a period of months to make things just right , before purchasing . Larry's timely review simple confirmed what I felt .  

 I also feel the " faulty amp " as you described it , is one of the best I've ever heard  and will likely keep it for years .

                 Clear ?

UpAnA's picture


Nothing this terrible should be foisted upon an unsuspecting listener. Please allow me to send you a forwarding address and I will properly dismantle it and melt down it's parts for scrap.

I will even refund all of the money recovered from the scrap process, and do the entire procedure for free.

Shall we split the shipping cost?

gn77b's picture

I can't help but wonder about few things and simply notice others. first, are the measurements and subjective reviews completely independent here at Stereophile and the results kept secret until publication? there seems to be a correlation between good measuring gear and subjectively perceived "dry", "fatiguing" or "uninvolving" sound. that, to me, seems suspect. I can imagine a scenario along the lines of "oh, these are the measurements? they look very good, I expect this to sound bad". maybe even unconsciously so. which, amusingly, proves that there is indeed a correlation between sound and measurements.

I have wondered if what we read here are not indeed slightly biased opinions and the result of idiosyncrasies. would anyone reviewing the No. 53 judge it the same?

also, the Soulution is an exceptionally good measuring amp but not class D and not based on large amounts of NFB. but again we read about that "clinical" sound. I can't think of many good measuring amps here at Stereophile that weren't characterized as sounding "dry", "clinical" etc. the same way I can't think of many horribly measuring SET amps that weren't sbjectively praised.

it gets even weirder with speakers. the also recently reviewed Alexandria XLFs have an impedance dropping to 2 ohms and a "schizophrenically" looking frequency response. I can think of many much cheaper speakers reviewed here at Stereophile that have much better measurements. and yet Mr. Atkinson ends his comments by saying "But overall, this is an impressively well-engineered design."

note that I'm not saying I disagree, I have myself perceived some sonic traits that seem to correlate with certain types of amplifiers but no topology/technology seems to be free from its inherent flaws. yet, many times there seems to be a double standard in some reviews, certain equipment receiving harsher critique, just because.

I'm at a loss on how to interpret all this.

Michael Fremer's picture

Actually many great sounding amps measure well too and I never did write that the Soulution wasn't a fine sounding amp. I just described how it sounded, something measurements really don't fully tell you. As for the XLFs, if you look at but one measurement, rather than at all of the ones that really provide a window on how a speaker actually sounds, especially of a complex speaker design like the XLF, you'll often come to a wrong conclusion. A pair of Russian engineers insisted based on the measurements that I have "boomy bass" at home. I told them I don't. They insisted the measurements don't lie so I invited them over. They came with their test CDs and I let them do as they wished with volume etc. Their conclusion: "no boomy bass"! Then I said let me play you some records. "Oh we don't like vinyl" they replied. I said "as long as you're here and since I was kind enough to invite you over, indulge me please." So I played them a few records---not "audiophile" records---that left them literally dumbfounded. They'd never heard anything like that (in a good way). One record in particular moved them---a 1975 recording of "Porgy and Bess" with Maazel conducting the Cleveland Orchestra. "Is that on CD"? they asked. I said "I'm sure, but it won't sound like that." "Well it should," they responded. "Buy it and let me know" I said. So they did and a few weeks later they got back to me and said "You are right! It didn't sound nearly as good". But I'm sure gn77b that it measured better.

kevon27's picture

Ah EMOTIVA XPA 1 mono blocks - made in China and I can get thee on eBay for $1600 and still pay my mortgage..
I love being frugal and can still get great quality gear.

georgehifi's picture

@John Atkinson, can you clear something up.(no AP low-pass filter?)

Is this a miss-print in square wave shot of figure 2 (10kHz squarewave into 8 ohms no AP low-pass filter). If so, that's very impressive for Class-D!

Cheers George

John Atkinson's picture
georgehifi wrote:
Is this a [misprint] in squarewave shot of figure 2 (10kHz squarewave into 8 ohms no AP low-pass filter).

That's what I wrote. The workbook for the No.53 is in my storage unit, so next time I have to go there I will check the notes.

John Atkinson
Editor, Stereophile

georgehifi's picture

If it is, there's something very special going on with this amps output filtering for this kind of switching noise attenuation and to be able to handle this kind of power.

Cheers George