Seit meiner ersten Begegnung mit einem Melco vor sieben Jahren höre ich in den eigenen vier Wänden so gut wie nie mehr Musik-Files vom Computer. Zwischenzeitlich verwendete ich gar einen Melco N1ZH/2 und einen N1A/2 in verschiedenen Anlagen. Seit dem Frühjahr ist nun der N10/2-S38 eines der beiden Topmodelle der japanischen Digital-Spezialisten.

Schon der erste Melco war ein klanglich rundum überzeugender Datenlieferant: Er stellte die auf seinen Festplatten gespeicherten Musikdateien über einen speziellen, für Audioanwendungen gefilterten LAN-Anschluss im Netzwerk zur Verfügung und agierte so als netzgebundener Speicher oder NAS. Zudem konnte er auch Streamer-Funktionen übernehmen und dank eines integrierten Renderers Digitaldaten per USB direkt an einen D/A-Wandler schicken. Anfangs war die Kombination eines Melco mit einer Streaming-Bridge, die dann per USB einen DAC speiste, dessen direkter Ansteuerung in Sachen Wohlklang deutlich überlegen. Damals lief auf dem Melco ab Werk ausschließlich ein Twonky-Server. Alan Ainslie, der den Vertrieb von Melco in Europa aufbaute und jetzt noch in beratender Funktion begleitet, empfahl, den Minim-Server zu installierten und gab dazu auch praktische Tipps. Inzwischen gibt es die zweite, am Zusatz „/2“ erkennbare Version aller „Music Libraries“, wie Melco seine Geräte nennt. Schön, dass beim Vertrieb gegen eine geringe Gebühr auch auf ältere Geräte die neue Firmware-Generation aufgespielt werden kann. Die lässt einem dann unter anderem die Wahl zwischen dem Twonky- und – wie ich finde klanglich überlegenen – Minim-Server.

Zudem hat die neue Firmware ein deutliche Verbesserung in Sachen USB-Ausgang gebracht: Die Klangunterschiede zwischen der Daten-Ausgabe per LAN an eine Streaming-Bridge und von dort per USB zum Wandler und der direkt vom USB-Ausgang zum Wandler sind deutlich kleiner geworden. Darüber hinaus ist nun dank der neuen Firmware auch das Programm Song Kong auf dem Melco verfügbar, das die Organisation der Musik-Dateien besonders für Klassik-Freunde stark vereinfacht, wie Roland Dietl in seinem Artikel berichtete.

Anfangs gab es zwei Gerätelinien, N1A und N1Z, dazu kamen noch Festplatten verschiedener Größen und Bauart. Aber egal, ob Speicher und Elektronik im einfachen Gehäuse in Normbreite oder in einem solchen aus recht massiven Aluminiumplatten mit einer Breite von lediglich 350 Millimetern untergebracht sind, sie werden immer von ein oder zwei Schaltnetzteilen gespeist. Vor einer Weile kam dann die N10/N100-Serie heraus: In deren nur 215 Millimeter breitem Gehäusen ist kein Raum für ein integriertes Netzteil. Je nach Qualität des Models kommt ein externes Zwölf-Volt-Netzteil oder ein in einem zweiten 215-Millimeter-Gehäuse untergebrachtes Linearnetzteil für die unterschiedlichen Spannungen zum Einsatz. Auch für die N10/100-Musik-Bibliotheken gibt es wieder Festplatten verschiedener Größe und Bauart: Das Angebot wurde aufgrund der raschen Entwicklung bei den Speichern zwischenzeitlich ein wenig unübersichtlich.

In dieser Phase haben sich die Kollegen und ich mit zwei Produkten beschäftigt, die auf der Melco-Website unter dem Begriff „Accessories“ geführt werden, aber weitaus mehr zu bieten haben, als ich mit Zubehör assoziierte: Wolfgang Kemper war vom Laufwerk D100 zum Rippen und zur Wiedergeben von CDs so begeistert, dass wir ein Exemplar als sogenannten „Redaktion-Ripper“ erworben haben. Melcos LAN-Switch S100 bietet nicht nur getrennte RJ45-Schnittstellen für Datenraten von 100 respektive 1000 Mb, sondern auch noch zwei Schächte für SFP-Module zur Datenübertragung per Lichtwellenleiter. Aber vorrangig für sein hervorragendes Preis/Klang-Verhältnis wurde es mit dem STATEMENT in High Fidelity ausgezeichnet. Das neue größere Switch S10 mit zusätzlichem Linearnetzteil im zweiten Gehäuse werde ich Ihnen vorstellen, sobald es in ausreichender Stückzahl verfügbar ist.

Ich hörte bis vor kurzem rundum zufrieden mit dem Melco N1ZH/2 meine Files. Doch nachdem ich meinen Chord Electronics DAVE mit einem Dreifach-Linearnetzteil auf ein deutlich höheres Niveau gebracht hatte, musste auch der Melco dran glauben: Er bezieht nun ebenfalls drei Spannungen aus einem Linearnetzteil, das AudiaZ respektive Chiemsee Hifi für diese Anwendung gebaut hat. Da Melco inzwischen eine Music Library anbietet, die von einem recht großen Ringkerntransformator samt einer Reihe von Siebelkos in einem eigenen Gehäuse gespeist wird, habe ich den N10/2 zum Test bestellt – und wenn schon, denn schon gleich mit der 3,84 Terabyte großen SSD-Festplatte. Dass nicht wie bisher oft zwei Platten verbaut werden, sondern deren nur eine, mag zwar auch an den Platzverhältnissen im schmalen Gehäuse, in dem die Elektronik der Music Library untergebracht ist, liegen. Hauptgrund dafür ist aber, wie Alan Ainslie erläuterte, dass man in diesem Fall auf einen RAID-Controler verzichten kann und dass dieser Verzicht recht klare klangliche Vorteile habe. Das Argument, dass zwei Platten Vorteile in puncto Datensicherheit böten, lässt er nicht gelten. Auch diejenigen Melco-Nutzer, denen es in erster Linie um den Klang geht, hätten deshalb bisher RAID 0 gewählt und lieber regelmäßig Backups auf einer externen Festplatte gemacht. Weitere kostenlose Tuning-Tipps gab's übrigens schon im Artikel über den Melco N1ZH/2.

Wer schon mal einen Melco benutzt hat, dem gibt auch die Bedienung des N10/2 keine Rätsel auf: Er besitzt auch die beiden LAN-Anschlüsse, einen zur Verbindung mit einem Netzwerk, über das die Steuerbefehle kommen, und einen zweiten für einen Streamer oder Wandler mit entsprechender Schnittstelle. Auf der Front gibt es einen, auf der Rückseite zwei USB-Abschlüsse. Während es bei der N1A- und N1Z-Serie noch drei Buchsen waren, die durch eine feste Zuordnung zu Funktionen – Backup, Wandler oder Erweiterungsfestplatte – auch unerfahrenen Benutzern die Bedienung erleichterten, glaubt Melco ihnen heutzutage zumuten zu können, den Buchsen auf Vorder- und Rückseite beliebig zu nutzen. Nur einer der beiden hinteren Anschlüsse wird für die Verbindung zum Wandler empfohlen. Den nutze ich erst einmal nicht: Der N10/2 übernimmt die Funktion meines N1ZH/2, und deshalb verbinde ich ihn per Ethernet mit dem Ansuz-PowerSwitch, denn im Test des Melco Switch S100 hatte sich gezeigt, dass die Verwendung eines externen Switch' gegenüber der Nutzung der beiden LAN-Buchsen im N1ZH/2 klangliche Vorteile bringt: In digitalen Ketten ist der kürzeste Signalweg nicht der beste, sondern der, auf dem das Signal am effektivsten von HF- und sonstigen Störungen befreit wird.

Der N10/2 war funkelnagelneu als er bei mir eintraf. Deswegen habe ich ihn direkt nach seinem Eintreffen mit dem Netz verbunden und von einer Festplatte, auf der all meine Audio-Files gesichert sind, mit Musik bestückt. Die Auslastung seiner Festplatte lag dann ähnlich wie beim N1ZH/2 über 40 Prozent, so dass hier annähernd gleiche Bedingungen herrschen. Die recht langwierige Bestückung seiner Festplatte mitgerechnet, hatte der N10/2 eine knappe Woche Zeit, um Betriebsstunden zu sammeln. Er lieferte auch für die Auralics in Wohn- und Schlafzimmer sowie die Küche Daten. Da ich nur drei Ansuz Digitalz D-TC Supreme zur Verfügung habe, übernahm ein Göbel Lacorde Statement Ethernet die Verbindung von Switch und Medienkonverter, beide Melcos hingen an Ansuz Digitalz. Um die beiden Melcos in ihrer Funktion als reine Datenspeicher und -Lieferanten zu vergleichen, brauche ich in Auralics Lightning-DS-App lediglich per Fingertipp zwischen zwei Musikbibliotheken hin- und herzuschalten – ideale Testbedingungen, die schnelle Wechsel ermöglichen.

Deswegen fange ich auch mit einem Album an, das ich eher zufällig in meiner Musikbibliothek entdeckt habe: Yours And Mine von Akira Tana und Rufus Reid – moderner, melodiöser Jazz in Quintettbesetzung mit zwei Saxophonisten. Schon bei den ersten Takten fällt der sehr natürlich wirkende Raum auf, in dem das Schlagzeug erklingt. Besonders angetan hat es mir aber Eddie Harris' „Freedom Jazz Dance“, den Bassist und Schlagzeuger im Duo interpretieren. Trotz des auch hier schönen Raums würde ich das Stück nicht zu den audiophilen Schmankerln zählen, allein schon deshalb, weil bei der Aufnahme des Kontrabasses auch Tonabnehmer mit im Spiel waren. Reine Mikrofonaufnahmen halte ich für realistischer. Dafür entschädigen aber die satte Bass Drum und das bunt und variantenreich strahlende Blech der Becken. Überhaupt: Das Duo agiert so spannend und voller Spielfreude, dass es der reine Genuss ist, sich den „Freedom Jazz Dance“ jetzt vom N1ZH/2 ein paarmal zur Gänze anzuhören. Als ich dann wieder zum N10/2 zurückkehre, erlebe ich eine Überraschung: Melcos Topmodell sorgt für ein noch feiner durchgezeichnetes Klangbild – und das hat nicht das Geringste mit langweiliger Analyse zu tun. Im Gegenteil, der Titel spricht mich auch rhythmisch mehr an, wenn er von der einen Festplatte des N10/2 kommt. Ich hätte nicht gedacht, dass das neue Melco-Duo das Hörvergnügen noch einmal so steigern kann.

Da muss ich unbedingt den ein oder anderen vertrauteren Track hören, wie zum Beispiel Ravi Shankars „West Eats Meat“: Das Tabla-Intro scheint beim N10/2 in einem größeren Raum aufgenommen worden zu sein. Man hört deutlich die Reflexionen an der Rückwand. Gerade in den schnellen Passagen macht sich die präzisere Abbildung des Zehners positiv bemerkbar. Im Vergleich wirkt die Wiedergabe vom N1ZH/2 nervöser und einen Hauch weniger strukturiert. Die Pauken im zweiten Teil kommen über beide Melcos ähnlich druckvoll rüber, aber in Sachen Klangfarben hat der N10/2 wieder ein bisschen mehr zu bieten. Ich probiere es noch mit Schostakowitschs Symphonie Nr. 15 aus der Living Concert Series. Aber auch dabei kann ich nicht den geringsten klanglichen Vorteil entdecken, wenn die Daten von meinem gepimpten N1ZH/2 kommen. Das Signal, das vom N10/2 kommt, und auf seinem Weg durch das Ansuz-Switch, das Waversa-Filter, die Auralic-Streaming-Bridge und den Chord-Electronics-M-Scaler ja noch vielfach behandelt wird, bevor es der DAVE mit dem Rest der Kette in Schall wandelt, muss einfach mehr Feininformationen enthalten. Wie wäre es sonst zu erklären, dass die Instrumente nun noch besser fokussiert und von mehr Luft umgeben scheinen? Hier gilt wohl wieder die Erkenntnis aus analogen Zeiten: Die Qualität der Quelle ist (mit)entscheidend. Informationen, die sie nicht liefert, kann auch die beste nachfolgende Kette nicht mehr hinzufügen.

Jetzt wäre es natürlich schön, auch mal eben die USB-Ausgänge der beiden Melcos miteinander zu vergleichen. Aber das geht nicht, da der USB-Ausgang für den DAC bei meinem Melco einer Buchse für das externe Analog-Netzteil weichen musste. Schließlich wollte ich ja keine irreversiblen mechanischen Veränderungen am Gerät vornehmen. Da bleibt nur, den N10/2 statt über Switch und Streaming Bridge direkt mit dem USB-Eingang des M-Scaler zu hören. Dazu installiere ich die kostenlose Melco-Music-HD-App auf meinem iPad. Sie ist eine Variante der mconnect-HD-App, daher ziemlich vertraut und auch leicht zu bedienen. Da ich nicht einschätzen kann, wie schnell der Wechsel von der Ethernet- zur USB-Datenausgabe dauert, verlasse ich mich wieder einmal auf Keith Jarretts wohlvertrautes „God Bless The Child“. Und das ist dank des N10/2 und seiner hohen Feinzeichnung und luftigen Wiedergabe ein Hochgenuss, wenn der Weg über den Auralic Aries geht. Beim direkteren Weg vom Melco per USB zum M-Scaler ändert sich daran erfreulicherweise nicht viel. Am auffälligsten ist, dass die Bass Drum und Gary Peacocks Viersaiter bei der direkten Verbindung noch einen Tick mehr Energie besitzen. Der Aries hingegen sorgt in diesem Frequenzbereich für mehr Kontur und Klarheit. Zumindest solange die Børresen 05 SSE noch zu Gast in meinem Hörraum ist, ziehe ich die Bass-Präzision der größeren Tieftonfülle vor. Aber hier bewegen wir uns schon lange im Bereich persönlicher Klangvorlieben.

Viel bemerkenswerter ist, wie weit sich die USB-Schnittstelle der vor Jahren noch deutlich überlegenen Datenausgabe per Ethernet angenähert hat. Momentan könnte es durchaus die bessere Wahl sein, in den N10/2 statt in ein kleineres Melco-Modell zu investieren und – wenn's wirklich um die letzten ein, zwei Prozent geht – später auf den Aries G2.1 zurückzukommen. Ich vergleiche die Schnittstellen noch einmal mit „Raijin“ von Arild Andersens Album Mira: Der Dialog zwischen Schlagzeuger Paolo Vinaccia und Tommy Smiths Shakuhachi-Flöte macht klar, dass der Auralic Aries neben mehr Präzision im Bass auch noch für einen minimal größeren imaginären Raum sorgt. Wenn man sie gehört hat, will man eben auf die vielleicht zwei Prozent mehr Wohlklang durch die Streaming Bridge nicht verzichten. Aber um zum Thema zurückzukommen: Die klanglichen Vorteile des N10/2 gegenüber meinem N1ZH/2 mit analogem Netzteil sind größer die zwischen den Schnittstellen von Melcos Topmodell.

In the first part I described how the power cables and power conditioners designed by Garth Powell can steer a system towards its maximum sound quality, even when the power supply topic had not been treated with neglect before. In this interview you can read what the developer has to say about his solutions and learn about my experiences with Niagara and Co. 

In the meantime, I've got used to the Thunder and Tornado cables as well as the Niagara 1000 and 7000: My system sounds more open, spoils me with larger – partly virtual – rooms, and above all it fascinates me with dynamic capabilities that are simply closer to the live experience. I am not alone with this almost enthusiastic verdict, by the way; I had invited Roland Dietl to listen to my newly acquired Chord Electronics M-Scaler. This colleague had already experienced the advantages of upsampling to 705.6 and 768 Kilohertz respectively during the BLU MkII test. Therefore, after one or two songs, it was clear to him that the quite significant improvement could not be attributed solely to the extrapolation of the data with the one-million-tap filter. He suspected that there must have been further changes in my system and discovered the Niagaras and the intertwined cables on closer inspection. It’s always reassuring when one’s own judgement is confirmed by a connoisseur of the sound system at hand.

However, my enthusiasm was not completely unclouded after I had enjoyed an LP in peace again: I was carried away by the gripping dynamics and increased the volume a little bit every now and then. When the disc was finished and I had raised the tonearm, a slight noise was audible – not the usual hum of the phono branch, but slightly higher in frequency. Until now, Einstein's balanced phono preamplifier had neither bothered me with such a thing nor with the slightest hint of hum. I then checked once again whether the power cables of the two were also plugged into a group of the Niagara together with that of the preamplifier. That was the case. Even the attempt to use a separate group for the bearing heating and tube power supply of the record player and the two mono preamps could not solve the problem. That’s why I left everything as it was, as firstly, the sound was only audible when the volume control was turned way up and secondly, Niagara developer Garth Powell had announced that he would visit me. I could then clarify the problem with the expert. One or two weeks later I also connected the Studer: As soon as I selected the appropriate input, a slight hum would be audible - even though the machine wasn't even switched on yet. The problem disappeared, however, as soon as the Studer took its power from the same group of sockets and filters as the preamplifier; another task for Garth Powell. Until his visit, I kept my hands off the tape and the vinyl, but that wasn't difficult considering that playing music files in previously unknown quality compensates for that.

Garth Powell came, saw, and agreed: Even if the individual groups for constant current consuming devices in his Niagara are so strictly separated that digital sources connected to one group do not have the slightest influence on the analog sources connected to the other group and on the preamplifier, it could make sense to feed them via the Niagara 1000, just because of the large number of digital devices in my system. With regard to the background noise, he noted that he knew the reason as well as the solution, but would like to try again if the phono stages could not be connected in such a way that they performed their task silently. When he heard the annoying noise, he immediately ruled out that it could be caused by a ground fault of the drive, tonearm or phono amplifier. It was clearly due to the power supply – including grounding, of course. Garth Powell managed to minimize the interference by cleverly distributing the power cables of the preamplifier, the phono preamplifiers, the bearing heater and the tube power supply of the turntable to the two groups of the Niagara 7000. But that didn't satisfy him, of course.

He then had the influence of the Studer A 80 on the power supply demonstrated to him. Afterwards he decided that the Niagara 5000 was the better choice for my system: In hifi and studio technology, different grounding concepts are used, which is only a problem if the power supply works with a "floating" ground, like the Niagara 7000 does with the balancing of the mains voltage. The 5000, on the other hand, offers all three groups a solid ground. In addition, only half the financial investment would be necessary. In terms of filter effect – and ultimately sound improvement – the two differed by just eight percent. Frankly, I don't have the slightest idea how to quantify something like that. But I tend to believe someone whose power correction improves the dynamic performance of power amplifiers as much as the Niagara 7000 does. Did I mention that the 5000's power correction is the same as the 7000's?

Even if the idea of being able to get almost the same sound advantages at half the price may seem tempting, I have to be patient first. Of course, even the developer doesn't carry a net conditioner in his hand luggage. So, Garth Powell and I will stop experimenting with the hardware and prepare the planned interview. But first I'd like to tell you how I got to know Garth Powell. In recent years it had almost become a tradition that Audioquest's brand ambassador Joe Harley met with my wife and me at an Indian restaurant on the last day of the High End show in Munich. The year before last, Joe Harley brought his colleague Garth with him. Nevertheless, it was not primarily about hifi. Joe Harley had not only made the recordings for Audioquest's own label, but had also recorded for ECM, Enja and GrooveNote. He was also significantly involved in the production of the Reissues of Blue Note classics on two Music Matters 45 discs. In the course of the conversation he then asked what "sommelier du son", my wife's and my record label, had recorded recently, and we reported about a previously unreleased six-channel recording of a concert by Günter Baby Sommer – which completely inspired Garth Powell: He had invited the drummer and co-founder of Free Jazz in the GDR to the USA, was well acquainted with him and even played together. With this Garth Powell not only revealed himself as an active percussionist, but also – which is even rarer in the USA – as a connoisseur of European jazz. And that's why my first question was about his musical career:

The fact that Audioquest's network technology is so extensively appreciated here is not primarily due to the very impressive sonic effect in my system, but rather to the fact that Garth Powell explains very convincingly why his mains conditioners should bring advantages in every system. Up to now, the transferability of experiences made during a test was quite low for both cables and measures to improve power quality, which is why only testing in one's own system could ultimately provide information about the improvements to be achieved. Garth Powell cites the strongly changing effectiveness of most line filters over a wide frequency range as the reason for the varying results in different systems. But listen for yourself:

What makes the Niagara 7000 – and the Niagara 5000 that it has been replaced with – seem almost indispensable to me at the moment is the Power Correction System, thanks to which Einstein's The Poweramp now plays as if it were on speed and knew no dynamic restrictions. And that with a nominal power of just 90 watts per channel! Garth Powell had already developed a circuit that can provide an enormous amount of additional power for the duration of transients for the Furman mains conditioners. For the Niagaras, he has refined it once again: Here he reveals why energy supply is so important to him:

Garth Powell has long been aware that some of the interference enters the equipment via grounding and has also tried to develop countermeasures. However, these were only effective temporarily or only in special device configurations, i.e. anything but consistent – if they complied with all safety regulations. Inspired by Audioquest CEO Bill Low, he then looked at the previously neglected direction of cable and found a solution:

The aim in developing the power cables was to achieve a very low characteristic impedance. The intensive collaboration with Bill Low also led Garth Powell to develop a special bi-wiring cable with extremely low characteristic impedance for the low frequency range. He had brought one of the first examples to Gröbenzell, where we could hear it the morning after the interview in comparison to the Goebel High End Lacorde Statement: The two cables played at the same, extremely high level, but each had minimal advantages and disadvantages in the different disciplines. To choose one of them, you just need more time. The waiting time for the standard models can be shortened by Garth Powell's comments on impedance and cables:

Meanwhile the Niagara 5000 has already spent a considerable time in one of my two Artesania audio racks. Unfortunately, I had to send back the 7000 before the "smaller" – which is only valid for the equipment, but not the dimensions – model arrived. So, I couldn't compare the two directly. But if memory doesn't deceive me, the differences between the two are extremely small – at least in comparison to the sound improvements that Audioquest's cables, the Niagara 1000 and 5000 bring in absolute terms. There are always recordings whose dynamics – and, not quite so often, their huge (imaginary) space – I practically rediscover. The PS Audio conditioner and a mix of high-quality power cables had also inspired the sound of my audio components before, but they don't come close to the holistic effects of Audioquest's power technology.

Peter Banholzer felt the same way, by the way, when he recently enjoyed one of my and also his test tracks in the listening room after a long hiatus. He hadn't heard Abdullah Ibrahim's "Calypso Minor" from the album “Sotho Blue” so well before in my room. The track lives from the sharp riffs of the wind instruments, the gnarly double bass and the fat kick drum. Or to put it another way: It’s full of transients. By the way the M-Scaler, who helps the DAVE in this discipline, was employed for the Chord cable test in the living room and not part of my system at the time.

More than a year ago, Robert Hay, AudioQuest‘s Marketing Director for Europe, suggested that I should do a review of one of their Niagara power conditioners. Then it took a little while until some updates were incorporated into general production. Finally, we planned a really comprehensive exploration of AudioQuest's power supply products: it was worth the wait!

As I have so far strictly separated the power conditioning for digital and analogue sources in my system – whereby the preamplifier is added to the latter – and supply the power amplifier(s) directly from the mains, it was obvious not only to replace a mains filter or regenerator with an AudioQuest solution, but to include all components in the test and to evaluate an overall concept. Of course, Robert Hay and I gradually replaced each individual component from the previous power supply with the new one and also evaluated each individual step in terms of sound. A colleague, whom I had told about this exercise, would also have liked to know how AudioQuest's power conditioners behaved in comparison with my old ones if my power cables were not replaced. But unfortunately, a Niagara 7000 or 5000 does not replace my PS Audio P5 with its downstream multiple-socket-strip in one go, nor does the Niagara 1000 replace the old Sonic-Line filter for the digital sources. By the way, Audioquest doesn't have simple multiple-socket-strips in its range: The AudioQuest method is that all devices should be connected directly to the power conditioners.

In order to supply the power amplifier directly from one of the three wall sockets (the other two host the PS Audio and the Sonic Line filters), I use a SunLeiste with a long supply cable, and from there, very high-quality power cables lead to my power amp or two mono amplifiers. Robert Hay, however, had a six-meter-long "Tornado High Current" cable made for the connection of Einstein's The Poweramp, so that the power amplifier could use one of the four "High Current Low-Z Power Correction" outputs of the Niagara 7000. The cable for the Studer was available in the required length only in the Thunder version, but not in the Tornado version. In short: due to the different configuration of the existing and the AudioQuest power supply, and due to the lack of all cable qualities in the different lengths required, we would not have been able to hear all conceivable combinations – not even taking the time expenditure into account. Since Garth Powell, the developer of the new power cables and conditioners, will also be speaking in detail, I will have to create this article as a two-part.

When AudioQuest Founder/CEO Bill Low decided that his company should look more closely at power supplies, he chose the same approach as he had used to enter the headphone and the mobile D/A converter markets. For these product groups, he had secured the collaboration of proven specialists: Steve Silberman and Gordon Rankin for DACs and Skylar Gray for the headphones. Good decisions, as the success of the Dragonflys and NightOwls show. Garth Powell was Bill Low's dream candidate for the development of power technology, because he had made a name for himself with Furman, a manufacturer that is well established on the professional market. My colleague Bert Seidenstücker tried out a Furman power conditioner in his system years ago and was so convinced that he bought it spontaneously.

But back to Garth Powell and AudioQuest: at first was reluctant to move from Furman in northern California down to AudioQuest in the southern part of the state. During the time that Bill Low, put the project on hold. Only after Garth Powell's decision in favour of Audioquest and the move did the topic of power technology regain high priority with the cable specialists. Since joining AudioQuest, Garth Powell has brought three power conditioners and several power cables to production readiness: above all, he wanted his solutions to work in all systems. In my experience to date, measures to improve the power supply were more or less effective depending on the system they were used in, or local grid conditions. Garth Powell himself will explain the technical details of his equipment in the second part of this article. Therefore, I will limit myself to a short technical description of the Niagaras and the cables and focus on the practical experiences with them.

Let's start with the NRG-Z3 cable: even in this relatively inexpensive model, the conductors consist of AudioQuest's Perfect Surface Copper and silver-plated drain wires, via which interference picked up by the shielding is conducted to ground. The patented "Ground Noise Dissipation System" protects the conductors from high-frequency interference. It consists of several interleaved layers of synthetic carbon, which largely convert the radiated energy into heat. The running direction of the current conductors and the shield is important and, are determined by the asymmetrical surface structure created during the production process. This surface structure effects the impedance of the conductor, it is fractionally higher in one direction. This can be used effectively to help drain away high-frequency RF/EMI noise that is picked up by the system. In this way, high-frequency interference is kept away from the equipment. In addition, common mode interference is also linearly filtered out by the use of a “Common Mode Phase Cancelling” array. Here, identical conductors are paired, but critically, based on the surface structure, they are laid in opposite directions. The idea is that they both pick up identical high frequency noise but as they are arranged in opposing directions they phase cancel each other. The final part of the story is that all of the cables are only connected at the source, and half at the load, so that residual noise is sent back to the electrical earth.

Beim Thunder bestehen die Leiter dann aus Long Grain Copper, also Kupfer mit langer Struktur. Die Oberfläche ist hier natürlich ebenfalls so behandelt, dass sie Audioquest Anforderungen für Perfect Surface Copper entspricht. Beim Thunder werden die stromführenden Leiter und die Erde getrennt geführt und geschirmt. Die drei Adern werden dann miteinander verflochten. Bei der Erde kommt selbstverständlich wieder das patentierte „Ground Noise Dissipation System‟ zum Einsatz. Das Dialectric-Bias-System oder kurz DBS sorgt mit seinen 72 Volt für ein stabiles dielektrisches Feld und soll so eine gleichmäßiger gefilterte Rauschableitung garantieren. Für noch bessere klangliche Ergebnisse steht das Tornado bereit, das sich vom Thunder vor allem durch einen 34-prozentigen Anteil von hochreinem Kupfer – laut Audioquest-Terminologie: Perfect Surface Copper+ – unterscheidet. Thunder und Tornado werden in Versionen für konstante Stromverbraucher und solche mit hohem Leistungsbedarf angeboten: Source und High Current.

The top model of the three mains conditioners is the Niagara 7000, which offers three groups of four sockets each. The first is intended for high current devices and features the so-called "Power Correction" system: this provides a power reservoir for peaks of up to 80 amps (80 for 220v 90 for 110v) with reduced mains impedance for amplifiers that is available at short notice – in milliseconds. This should benefit all power amplifiers, regardless of whether they operate with linear or switched-mode power supplies. The other two groups have isolating transformers with pre-stressed dielectric: the patented AudioQuest technology is designed to significantly reduce non-linear distortion and increase the bandwidth and efficiency of the transformer's noise dissipation. The connected loads are thus supplied with a symmetrical mains voltage of two times 115 volts related to the floating earth. The patented "Ground Noise Dissipation System" is used in all groups: it provides directional ground noise dissipation in six stages. This patented technology is designed to significantly reduce ground noise without compromising safety or creating ground loops at low levels. Garth Powell placed particular emphasis on the fact that the noise dissipation technology is linear over 21 octaves for all three conductors. This guarantees that his power conditioners are effective in all configurations. With conventional conditioners that have many resonance peaks, this is only the case if they coincide randomly with the interference. The mains plug and the sockets of the Niagara 7000 are very low-impedance specimens from the NRG series. The Niagara also provides protection against overload and short-circuit to ground.

The Niagara 5000, which is almost identical in appearance but only half as expensive, is also equipped with the "Power Correction" system for power amplifiers or integrated amplifiers, but has to get by without the balancing isolating transformers and has a solid earth instead of a floating one. Nevertheless, common mode interference from 20 kilohertz to one megahertz is reduced by 30 decibels. The voltage applied to the eight sockets for constant current consumers is freed from interference in four groups by linear noise dissipation systems. Of course, there are low-impedance inputs and outputs as well as protective circuits for overload and short circuits. The surge protection and the high-quality sockets can also be found on the Niagara 1000, which offers one high-current output and two groups with a total of five outputs. Of course the patented "Ground Noise Dissipation System" and the linear "Noise Dissipation System" are also used here.

Before Robert Hay arrived, some boxes with various cables as well as a Niagara 7000 and a 1000 had arrived in my place in Gröbenzell. Robert Hay had received a detailed list with all the components of my system and then selected the cables he deemed to be suitable. We had previously agreed that the cables should meet a high standard, but that they should not be AudioQuest's top products. A one-meter long Tornado costs 1100 Euro, the Dragon costs 4400 Euro. But this would not be in any reasonable proportion to the value of the components for the phono stages, the Melco or the DAC. For the router, the switch, the Aries Femto, the M-Scaler and the heater and the tube power supply for the turntable, Robert Hay had chosen cables of the NRG series.

On a Wednesday in the second half of October the time had come. Robert Hay had arrived the evening before and appeared the next morning in my listening room, full of energy. First, we freed the cables and power conditioners from their boxes in the living room. Otherwise the amount of packaging material would have blocked half of the listening room. Einstein's The Poweramp was the first to benefit from a Tornado. Since the three sockets used to supply the system were occupied, we supplied the Einstein with energy from a socket not previously employed for my system, using the six-meter-long Tornado already mentioned: the power amp now sounded minimally more dynamic and also a little more open than before. In a second step, the Niagara 7000 was used in front of the long mains supply line and made the playback even more lively and dynamic: I was persuaded of the beneficial effect of the "Power Correction" after just a few minutes.

From then on we used the mains socket to which the now unused SunLeiste was attached to power the Niagara. The 7000 continued to supply the power amplifier and we began to gradually re-connect the preamplifier, the two (mono) phono preamps, the heater and the tube power supply for the turntable from the PS Audio and the Audioplan distribution box to the Niagara: the improvements were clearly noticeable in the preamplifier and also with the equalizers – but not as strong as in the preamplifier. On the other hand, I was surprised that the tube power supply benefited so clearly from NRG Z and the cleaning effect of Niagara. The bearing heater, however, did not show any sound improvements, so I reconnected its power cord to the PS Audio for further experiments: in this way I could connect the phono preamplifier, the tube power supply and the preamplifier to one group of the Niagara and find out whether the analog playback would be affected by the use of the second group.

Although AudioQuest’s delivery also included the Niagara 1000 with a further six sockets, Robert Hay had only ordered it because the turntable with heater and power supply, the two phono stages and router, switch, NAS, streaming bridge, up-scaler and converter simply required a large number of sockets. I almost forgot about the Studer A80, which also demanded a socket. The AudioQuest specialist considered the separation of analog and digital sources on two network conditioners unnecessary: "The Niagara has two completely separate output groups.” After my previous experiences with the effects of the location of a router on the playback of records, I was extremely skeptical: of course, it was not clear at that time whether the impairments caused by the Apple Time Machine, which was set up very close to the system, would be induced solely by its RF signals or whether they would be transmitted over the mains, despite two filters. Since the use of a linear power supply for the router had a positive effect, I have preferred to be cautious since then and try to separate digital and analog devices as rigorously as possible. As the list of all components involved in the playback of files has already shown, the second group of Niagara's four sockets cannot supply all remaining devices. We therefore fed the M-Scaler from a Poweradd with 23,000 mAh and connected the router, Melco, Aries Femto and DAVE to the 7000 – without listening to them. That’s because the music continued to come from the vinyl record. To my astonishment – and as Robert Hay had predicted – the sound did not change at all when the digital group was added. I then connected the switch again instead of the Melco, didn't just leave it at a test track, connected and disconnected the digital components again. In no case was the record playback affected by what happened to the third group of the Niagara: Garth Powell has succeeded in completely isolating both groups from each other. Respect!

So, if your system requires four sockets for digital equipment and four for preamplifier and analogue, a Niagara 7000 is the ideal solution. Not to forget: thanks to "Power Correction", your power amplifiers or active speakers will work as dynamically as if they suddenly had twice as powerful power supplies. Unfortunately, my system is a little bigger, so there was still a lot of work to do before Robert Hay and I could enjoy the sound improvements of AudioQuest's power cord and conditioners. We unplugged the turntables and phono stages, and reconnected the digital components to the Sonic-Line filter: everything back to baseline. Then we let the converter and co. move to the 7000 one by one. Thanks to the Niagara, the DAVE sounded a bit more open, and the improved current allows the Melco a slightly wider representation of space. Miraculously, the router took it in the same direction when the power flowed through the AudioQuest. The Auralic Aries Femto, which – like the router – is powered by a Sbooster power supply, showed no improvements, neither did the Switch nor the M-Scaler. After this test marathon, it was time to clean up, because according to Robert Hay, it looked like a spaghetti factory had exploded around the stereo. However, I had never seen such thick spaghetti before. But even if it looked chaotic and we hadn't managed to connect the Studer yet, we called it a day and ended up in a Bavarian pub.

The next morning, we took the PS Audio out of the rack, put the Niagara in it and put some unused, easily accessible cables aside. Then Robert Hay headed to the airport and I went to a baroque festival for recording. A few days later I removed the rack with pro-equipment and the Sonic Line filter from the listening room to make room for the Niagara 1000 and completely rewire the system with Tornado, Thunder and NRG. Even if it is not necessary – as described above – for sonic reasons, the cables of the digital components entered the Niagara 1000, and the cables of the turntable, phono stages and preamplifier were plugged into the Niagara 7000. When, finally, there was some order restored, I could enjoy my system – out of comfort, with digital material only. First I listened to one or two albums that had almost fallen into oblivion, but then I also listened to a few of the usual test files: The sound in my listening had never been so compelling before. AudioQuest's power technology takes my system to a higher level.