Harmonic Convergence: the Effect of Component Tweaking DIY Monitoring

Sidebar 2: DIY Monitoring

Inquisitive readers may like to hear what kind of garbage resides in their amplifiers' rails. Using an old AC-coupled power amplifier and speaker as a monitoring system, hook this amplifier's input ground to your hi-fi amplifier's central ground point (usually the 0V busbar between the reservoir capacitors). As you want to hear the AC signal on the rail, make sure that the DC voltage (which can be up to 100V) will be safely blocked. Connect a 100V, 100µF electrolytic in series with the hot input lead of the monitor amplifier, oriented so that the positive side of the capacitor is connected to the test amplifier's positive supply rail; ie, the positive terminal of the reservoir capacitor which isn't grounded. If you cannot reliably identify these parts, do not attempt this test!

Switch on the monitor amp first, but disconnect the monitor speaker. With both amplifiers switched on, reconnect the monitor (ideally placed outside the listening room) and play music through your hi-fi amp, which should be hooked up to your speakers in the normal way. On the monitor, unless you have a perfect amplifier, you will hear a rather unpleasant mixture of hum and buzz, with highly distorted music in the background.

The rails you are listening to are only displaced from the amplifier's output by the output device or devices.—Ben Duncan

Postscript: Harmonic Convergence in 2019

This article is one in a war of words of over 40 years, to show just how wrong—and bovine!—the hardline objectivist "sonic denialists" have been. Since 1992, a great deal more has been shown and proven (footnote 1) where the listeners'—the people's—score is 1, and naysayers 0. Some 20 years ago, Cyril Bateman discovered how to measure the ugly distortions in capacitors, previously scathingly denied by hard-line objectivists like Douglas Self (footnote 2).

Nowadays, one can read accounts of the Rauch amplifiers' sonics in the "Speakerplans" PA-users forum, where the DVT-50 appears under topic titles like "Are PA amps Hi-Fi?" and "Best sounding amp?" Tellingly, Bateman's work showed how using capacitors in a particular, niche way will minimize their distortion. Exactly what's inside the Rauch amplifiers, designed 20 years earlier. Also in my AMP-01 main-frame/back-plane preamplifier design, described in the May, June, July, August, September, October, and November 1984 issues of Hi-Fi News (footnote 3), in which I'd started using back-to-back Elcaps in my designs.

In the years before 1992, deep spectrum analysis was barely available to the poorly resourced audio industry. It was exotic equipment, complex, large, heavy yet insanely delicate, that mainly covered radio frequencies and models with enough dynamic range for hi-fi, was kit that only government and military-industrial contractors could afford.

Frosty January '87, an early Audio Precision test set arrived at the Rauch factory—the second UK user, after the BBC. The AP System One swept, and swept the board for regular THD+N percentage measurement (footnote 4). By 1990, my lab was equipped with the Dual-Domain DSP option, which created a spectrum analyzer, inside a machine already made for serious audio testing, ie, with robust and balanced connections—what even today's "house-priced" RF analysers and also toylike soundcard types both omit. As with these, harmonics could be shown as discrete spikes. But the AP's DSP also allowed plotting of harmonics against frequency and level. As above. Making raw harmonic information easier to grasp. Yet this way of presenting the picture appears to remain rare, years on.

I first read about the importance of harmonic structure in "Amplifier Musicality," an article by Jean Hiraga in the March 1977 issue of Hi-Fi News. In 1978, my console designer neighbor Harry Day showed me work by Russ Hamm, a recording engineer, about varied harmonic structures in recording equipment, and where the musical qualities were identified, and related to instruments.

In my 1996 world-reference hardback book High Performance Audio Power Amplifiers, I cite articles from Wireless World going back more than 70 years, where the need to consider harmonic distortion from a musical perspective is grasped. The Radio Designer's Handbook, a compendium from 60+ years ago, charts the musicality of harmonics up to the 20th. Is knowledge is regressing?

A difficulty with harmonic structures, is that they're harder to judge. But as with music, patterns can resolve. Between 1997 and 2000, I tested over 120 amplifiers using the Audio Precision System One Dual Domain with listening help from the musically adept ears of Norma Lewis, and spotted something about varied equipments' distortion hitherto undocumented, and came up with a single figure of merit. Simplicity with corroboration. This development was related in my then column in Hi-Fi News.

As for power supply noise rejection, the continuing widespread market for power conditioners may speak for itself.—Ben Duncan (footnote 5)

Footnote 1: The professor removes his right sandal and thrashes it violently on the desk, awakening listeners with a pure transient.

Footnote 2: From the 1970s through the 1990s. Douglas Self has been denying in print that capacitors can affect sonic quality. However, in light of Bateman's hard measured proofs of distortion in capacitors in 2001–2003, which backed up nearly all the things audiophiles had heard as to best capacitor types, he nowadays brazenly mentions in his books capacitor choice, for example, and usage, in particular, to avoid distortion—a design aspect, naturally he has always known about all along. (Sarcasm Knob set at 13.) This is a classic strategy—stone-walling; denial; scathing criticism; but then making out he invented it all along.

Footnote 3: I still have the example of the single-chassis version of Ben's AMP-01 that I used as my reference preamplifier before purchasing an Audio Research SP-10 in 1985. I should retrieve it from my storage unit and give it a listen.—John Atkinson

Footnote 4: Stereophile purchased an Audio Precision System One in 1989 and had it upgraded to Dual Domain status a year later. I used it until it was replaced by an Audio Precision SYS2722 at the end of 2007.—John Atkinson

Footnote 5: Ben Duncan's job description ranges though analog electronics engineer, product designer, inventor and research scientist, bringing musical quality to the tribal sonic worlds of recording studios, live stages, festivals and dance venues, thro' to serious hi-fi.

Anton's picture

Thanks for uploading this.

A truly fun read.

Bogolu Haranath's picture

As a side note ....... The Mark Levinson No.20 models mentioned in the 'Footnote 12' have very low output impedance ... between 0.01 and 0.05 Ohms ........ In the current models, Mark Levinson makes one such model, with very low output impedance, the No.534 stereo-amp (Stereophile Class-A) ...... See Hi-Fi News measurements of ML No.534 :-) ........

jeffhenning's picture

Very interesting article.

A few thoughts:

• Not that you had to mention it, but switch mode power supplies seem to have advanced to the point that, when done well, they seem to have some big advantages over the standard types. They're used to tremendous effect by Benchmark and Devialet.

• Both Benchmark's and Devialet's amps, which are radically different designs, outperform any Class A amp I've ever heard of. The Benchmark AHB-2 AAA circuit designed by THX uses some type of crazy Class A, A/B and B voodoo to get the advantages of all those types as well as feed-forward correction.

• Reading a transcript of a town hall with Bruno Putzeys and Peter Lyngdorf of Purifi Audio (between them they have around 60 years experience creating groundbreaking Class D amps), their latest design uses something like 85dB of feedback to eek as much out of the amps as possible. While in classical designs that would be insane, in the world of Class D it's apparently a great thing.

• In the end, I don't care how a piece of audio gear gives me perfect linearity, bullet-proof operation and a total absence of noise & distortion as long as it does that... or something imperceptibly close to it!

Graham Luke's picture

I would imagine that the roll of lavatory paper is the most pertinent item in the photograph above...

Archimago's picture

Or maybe 5W or 10W into that 8-ohm load pre- and post- modification rather than slightly below clipping for a 200W amp!?

Let's see the difference in harmonic structure at actual real useful power output levels!

Ortofan's picture


Ortofan's picture

... performed and analyzed independently, so that it could be determined which modification might have made the most difference, between the resistor replacements, the electrolytic capacitor substitutions, the op-amp rolling or the redressing of the wiring harness.

johnny p.'s picture

...until the last few years. Power-supply regulation and circuit-stage design remained the same, with high noise and distortion to boot.

I think the Sanders amp in 2008 was the first that used Thermal-Trak transistors and an IC-regulated power supply.

Since then, few have done the research as no (designers) seem to care.

But all-A circuits, like Valvet and Pass, have paved an alternate but scientifically-legitimate method of amplification. If used w/ highly-sensitive speakers.

adrianwu's picture

The necessity of class A/B with all the attendent power supply modulation and crossover distortion only serves the purpose of increasing the power output to beyond 20W. But a 300W amplifier only plays 12 dB louder than a 20W one, and all the extra power is often just converted to heat by over complicated crossover networks and inefficient drivers. The high end industry has been focussing on the wrong problem, which is to produce high power amplifiers that can drive complex impedance loads, when what they really should be doing is try to produce high sensitivity drivers and using active crossovers. High sensitivity drivers have distortion levels orders of magnitude lower than low sensitivity drivers, and TDH of drivers are measured in percents or even tens of percents levels, not 0.1% percent level of amplifiers. By reducing power output to 20W or less, power supply design can be simplified, and negative feedback often becomes unnecessary, which avoids instability problems and the generation of unpleasant harmonics.

Don Whittle's picture

Great article by Ben Duncan that shows how better quality components and thoughtful wiring layout produce a clearer sound.

I am a civil engineer with a lifelong interest in audio and
particularly appreciate articles such as Harmonic Convergence that use
observation along with engineering and science to improve audio performance.

Over the years I have owned and listened to many home audio and professional amplifiers including Krell, Kelvin Labs, Electrocompaniet, Pass Labs, C Audio, TOA, EMC, Bryston, Carver, HH, QSC, Turner, Turbosound etc, etc.

But learning to use a soldering iron and scope has led to some amazing aural places just like the article suggests. As my best power amps for everyday listening today are modified Rauch DVT250s & 50s, which I found were designed by Jerry Mead, Ben Duncan and team for recording studios and touring artistes.

Over several years I've bought a number of battered old Rauch amps online. They'd come out of London's clubs and venues, all closing down. Even after just a clean-up, they produced surprisingly realistic sound. Then I came across Ben's upgrade path much as mentioned in the article. This article on component quality and wiring layout is showing graphically the musical improvements that I'm now hearing. It's great when musicality and solid engineering combine.

This article seems to be a continuation of the concept in Ben Duncan's book on power amplifiers and his numerous articles on audio that focus not simply on circuit design but on all aspects of the audio system out to subtle distortions, mains and RF, and looking at how every tiny part can affect the final sound.