Innuos PhoenixUSB Reclocker Review: A Meatball to replace the Spaghetti?
Those of us who slave over improving our digital streaming audio chains invariably end up with a plethora of boxes, cables, power supplies, accessories, more boxes, and more cables. In other words, what we — affectionately, and our detractors deridingly — call “spaghetti.”
When it comes to the USB part of the chain, Innuos is offering us a different alternative. A meatball, if you will, to replace the spaghetti. According to Innuos, maker of the highly-regarded ZEN, ZENith, and Statement music servers, their PhoenixUSB Reclocker (abbreviated Phoenix going forward) “takes the USB signal from any source and completely regenerates it to an extremely high-precision signal to feed into your DAC, allowing it to perform at its best.” All the ingredients you might need for high-quality USB playback are covered. Bespoke high-quality linear power supply? Check. High-precision Oven Controlled Crystal (Xtal) Oscillator (OCXO) clock? Check. Complete USB signal regeneration? Check. It’s all there — in one box.
Priced at $3195 (all prices shown in this article are US MSRP, unless otherwise indicated), the Phoenix is clearly aimed at high-end systems. Innuos’ primary motivation for offering the Phoenix is as a standalone upgrade option to their existing ZENith SE Mk.II Std (from $7000; abbreviated SE going forward) customers, to provide an upgrade path to approach the sound quality of their flagship Statement server (from $13,750). The Statement already contains both the USB regeneration capability delivered in the Phoenix, as well as Ethernet regeneration. However, as their product literature makes clear, the Phoenix will work with, and deliver value to, any USB source.
That gets my attention. As readers who’ve followed my digital audio journey know, I have a bad infestation of spaghetti in my chain (see Review Topology section). In fact, the USB portion of my chain costs more than the Phoenix (see next section).
Does the Phoenix provide a significant sound quality upgrade over the already-excellent ZENith SE server? And can the Phoenix provide a winning combination of lower cost, single-box simplicity, and superior sound quality over my spaghetti chain? These are the questions this review will delve into and answer.
Setting Context — Pricing Spaghetti?
Estimating the cost of a spaghetti chain can be misleading if you just add up the cost of the individual pieces. I made the following assumptions when amortizing the cost of shared components:
- Divide a multi-rail PSU’s cost by the number of rails to get per-rail cost.
- Assume a reference clock is clocking 3 devices, so the cost per device is 1⁄3 of total.
Using these assumptions, the USB portion of my spaghetti chain comprises:
- an SOtM tX-USBultra SE ($1470),
- a 12V rail, including DC cable, from a 3-rail Paul Hynes SR-7 DRXL PSU ($1770), and
- a 10MHz reference clock output from a MUTEC REF 10, including a Habst 5N BNC cable ($2300).
This cost, approx. $5500, is significantly greater than the MSRP ($3149) of the Phoenix. This puts the daunting MSRP of the Phoenix in some context. Even DIY optimizations of sufficient quality get very expensive, so the Phoenix needs to be evaluated with that in mind.
From my own experience achieving stepwise sound quality (abbreviated to SQ for the rest of this article) improvements with the spaghetti approach, I know that all of the following factors contribute to SQ on the USB path:
- Rejecting noise from the USB source and regenerating a clean signal to the destination,
- Using a high-precision, low-phase noise clock in the regeneration, and
- Quality of the power supplies to the clock and the USB chipset.
I asked Nuno Vitorino, the technical brains and chief designer of Innuos’ products, to delve into the technical details of the Phoenix.
Rajiv: I know you use power supplies designed by Dr. Sean Jacobs, a highly-respected PSU designer. Can you describe the Phoenix’s power supply design?
Nuno: We use a 120VA audio grade toroidal transformer, custom designed to our specs. There are 2 independent rails, which are used the following way. One of the rails is exclusive to the OCXO clock. The other rail is to feed the remaining USB board. However, the USB chip requires 3 independent voltages as we wanted a chip that does not include internal switching regulators. So from the 5V rail we actually use LT3045s to further regulate the rail in order to get the 3 voltages we need into the USB chip.
Rajiv: Please explain your use of an OCXO clock, and your clock design.
Nuno: When designing something at this level, it’s all about the sum of little details. One example is the OCXO clock. Phase noise is far from being the only thing that matters. There’s a lot of other variables around how the oscillator is built that have an impact on sound quality. Otherwise, it’s like trying to buy amplifiers based on THD figures. The most important factor we found is to ensure it’s as stable as it can be. It comes to details such as preventing the exposure of the clock to temperature variations inside the box, from “fencing” the clock on the board, the power supply and the position in the chassis so that it’s not on the path of the convection of airflow for example. The phase noise figure is rather pointless if it can only be achieved in perfect lab situations and only when the oscillator is new – aging should be a concern. In this sense, after testing several clocks, we have found the ones we use to provide the best results for our application. They’re more expensive than the popular Crystek but well worth it, in our opinion.
Rajiv: What are the advantages of the single-box approach, and having full control of the design?
Nuno: The first advantage is, like the Statement, short paths. Short path to power, short path from OCXO to chip. All these are fundamental. We do the same EMI absorption treatment on the board as we do on the Statement and use the same high-end feet as used on the ZENith SE and Statement so to minimise how vibration affects the components. Even the USB connector has rubber washers to absorb some vibration coming from the chassis and it’s soldered directly on the board, contrary to some designs we have seen that use a good connector only to have it connected by a long cable to the internal board. As with the Statement, the board itself uses quite large tracks with gold-plating throughout. And of course, we use the same grade of power supplies as with the Statement, with Mundorf caps and Sean Jacobs’ newly-designed regulator architecture. We could go on and on – there’s quite a lot of small details that add up quite significantly.
From a usage perspective, the product couldn’t be any simpler. Other than the Power On/Off switch, there are 2 USB ports on the back panel:
- USB ‘B’ port for the input from a USB source
- USB ‘A’ port for the output to the USB DAC.
Once connected, no further action is needed.
Review System Topology
Shown below is a pictorial representation of the audio system I used in my review. I prefer to test gear in at least two separate systems, both to discern how they behave in different systems, and to test with both headphones (my primary system) and loudspeakers.
Since the listening for this review happened during the 2020 Covid-19 pandemic when I, like much of the world, was in lockdown, I was unable to test on a second system. Instead, I configured two distinct baseline topologies in my system, which are described below.
Topology with ZENith SE
To assess the Phoenix for users of the ZENith SE server, I acquired a loaner SE, and replaced my current music computer with the SE. The topology with the SE and the Phoenix is shown below. I compared this topology with and without the Phoenix in the path.
Figure 1: SE Topology
Shown below is my current baseline configuration.
Figure 2: Baseline Topology
For comparison, here is the review topology with the Innuos Phoenix in the path, as shown below.
Figure 3: Phoenix Topology
To enable you to listen to the same tracks that I did, I have created a public playlist on Qobuz USA. This playlist includes the tracks mentioned in this review, as well as some of the others I listened to in the course of this evaluation. Please note that in some cases, the Qobuz track will only stream at 16/44.1, whereas I may have used a local hi-res version. Still, this gives you a sense for the music I listened to for evaluation.
I started out by refamiliarizing myself with the sound of the SE directly feeding my DAC as shown in figure 1. While I have moved on to a DIY music server since I owned the SE, this experience reminded me of what had attracted me to the SE and why I had enjoyed it so much. What it does, it does extremely well. It delivers a solid, dense, and utterly grain-free presentation of music. It’s no wonder these units have held their value, and get snapped up in no time when the rare unit shows up on the used market.
Effect of the Phoenix with the SE
Let’s get this show on the road. What does the Phoenix bring to the party?
The title track of Fleetwood Mac’s album Tusk (Rhino-Warner, 2015 Remaster, 24/96) starts out with a lot of intricate effects and ambient sound fragments, building up to a thudding hypnotic drumbeat. With the Phoenix in the path, there is almost a startling opening up of the soundstage. Fine details are both audible and discernible. The texture of the drumbeats and the strums of the guitar strings are more palpable and realistic. The bass of the drums is both tighter and more palpable.
Given how much I love Brahms’ music, it is surprising I had not heard his Piano Concerto No. 2 until recently. And what a piece it is, especially in this wonderful album, Brahms: Piano Concerto No. 2 & Handel Variations (Ondine, 24/48). In the first Allegro non troppo movement, the presence of the Phoenix in the chain made Lars Vogt’s piano appear to materialize in front of me in all its three-dimensional glory. The sense of space and ambience was much greater, as was the size of the soundstage. And as I know so well with good USB regenerators, individual instruments sounded more real and easier to locate in the soundstage.
This comparison was an easy one. The deceptive thing about systems like this, built around quality components like the ZENith SE, is that until you hear the improvement, you can convince yourself that “it can’t get any better than this.” And then—bam!—you hear how much better it can get. This is something we’re used to in audio, when considering traditional component upgrades like amps or DACs. The fact of the matter is that devices like the Phoenix can deliver just as massive an improvement as a conventional component upgrade.
If you’re an existing SE owner, the Phoenix is a highly recommended upgrade. Unfortunately, the logistics did not work out for me to compare the combination of the SE and Phoenix with the flagship Statement server in my system. Having heard this comparison at a show, my sense is the Phoenix gets you a large portion of the way there.
Comparison of Phoenix with the “spaghetti”
Having established the value of the Phoenix in the Innuos ecosystem, it was now time to evaluate it in the wild against its “spaghetti” competition. Could the single box Phoenix, placed downstream of my custom music computer, equal or outperform the combo of the SOtM tX-USBultra SE, a 12V rail from a Paul Hynes SR-7 DRXL PSU, and a 10MHz reference clock from a MUTEC REF 10? Let’s find out. I compared my baseline topology in figure 2 to the Phoenix topology in figure 3.
First up, I turned to this delightful new release of Mozart: Symphonies Nos. 39, 40 & 41, Ricardo Minasi, Ensemble Resonanz (Harmonia Mundi, 24/96). Listening to the Andante con moto movement of the 39th, on both the Phoenix and the spaghetti, I was first struck by how similarly engaging they sounded. With extended listening came the realization that the Phoenix was bringing more to the table. I’ve long lamented a slight thinness in the signature of the SOtM tX-USBultra SE. Now, it was shown up by the Phoenix’s dense and meaty presentation. This was not at the cost of transparency and air. Despite the tX-USBultra being powered by the best PSU in my stable, an SR-7 DR, and one of the best 10MHz reference clocks, the MUTEC REF 10, the Phoenix seemed to have a lower noise floor and a blacker background, while portraying just as big and accurate a soundstage as the combo. Nuno’s thesis, as articulated in the Technology Overview section above, is that it’s all about the sum of little details. This was being borne out in my observations. The synergy of the Phoenix design was outperforming my spaghetti. This was exciting!
Switching gears, I fired up this recent Mobile Fidelity Sound Labs (MFSL) DSD remaster of the Dire Straits classic, Communiqué (MFSL, DSD64). This is a wonderful remastering of this classic album. On Where Do You Think You’re Going, the Phoenix and the spaghetti combo both did an outstanding job rendering the detail and intricacies of the guitar notes in the spare opening. The Phoenix pulled ahead on density and body. The drum lines were just more visceral, and on the whole the track sounded effortless and less fatiguing with the Innuos box.
And so it went. On track after track, the Phoenix matched my spaghetti combo on resolution and air, while pulling ahead on bass articulation and a smoother, more liquid, tonality.
This was a really impressive showing by the Phoenix. While the tX-USBultra SE combo isn’t the only option for USB optimization, it is one of the highest SQ options out there. The fact that the single-box Phoenix outperformed it, and at a lower cost, means that for those looking to improve their USB chain, there’s no need to mess with spaghetti. Just buy a Phoenix, add power and USB cables, and live happily ever after.
Enter the MUTEC REF 10 SE 120 clock
Late into the review period, I received an evaluation unit of MUTEC’s latest reference clock product, the REF 10 SE 120. The official release date and US MSRP are expected to be announced soon. The European MSRP is €5498 incl. VAT.
The main feature of note is that the SE 120 reduces the already impressive phase noise of the REF 10 even further. Using just one metric, the phase noise @ 1Hz offset from the nominal 10MHz frequency, the SE 120’s claimed phase noise is “under -120 dBC/Hz,” compared to the original REF 10’s “under -116 dBC/Hz.” Indeed, Chris Peters of MUTEC told me the particular OCXO installed in my unit was measuring at -121dBC/Hz.
Of course, the only way to establish whether this lower phase noise had a sonic impact was through careful listening. To determine this, I compared the impact of deploying the REF 10 SE 120 in place of my REF 10 in the baseline topology (figure 2).
For this comparison, I turned to one of my favorite recordings, Sibelius Symphonies 3, 6, & 7 (BIS, 24/96) with Osmo Vänskä conducting the Minnesota Orchestra. Listening to the last Moderato movement of the 3rd, it didn’t take a trained ear to hear just how big an impact the SE 120 was having! The triumphal rushing strings and soaring horns that characterize this movement provide the perfect ingredients for comparison. With the SE 120, the massed strings have far better texture, sounding more like a collection of individual instruments than a homogenous mass. When the horns soar, they seem to soar higher with the SE 120, with a longer overhang, or decay time.
The overarching advantage of the SE 120 over its predecessor was one of tighter articulation and focus, or greater coherence and resolution. It all added up to an uptick in realism — a more palpable sense of being there, of instruments sounding more fleshed out and dimensional.
To be honest, I was not expecting as big an impact as I heard. My reference clock journey has spanned: no clock, Cybershaft OP-14, MUTEC REF 10, and now the REF 10 SE 120. I expected the law of diminishing returns to kick in, but it did not. If anything, the jump from the REF 10 to the SE 120 was bigger than what I remembered going from the OP-14 to the REF 10.
If you’re an existing REF 10 owner, upgrading to the SE 120 will not be cheap, but rest assured that if you derived a big improvement with the REF 10, the SE 120 will give you a substantial uplift. Only you can (and should) determine if this gain is worth the cost.
Comparison of Phoenix with the “spaghetti” with the REF 10 SE 120
Once more into the breach went the Phoenix, this time against stiffer (and more expensive) competition. Factoring in the SE 120, I estimate the spaghetti chain’s cost was now approximately double the Phoenix.
In this comparison, I modified the baseline and Phoenix topologies in figures 2 and 3, by replacing the REF 10 with the REF 10 SE 120.
I was in the mood for something majestic, and what could be more so than Alan Hovhaness’ Mount St. Helens Symphony, Gerard Schwarz, Seattle Symphony (Delos, 16/44.1). Hovhaness’ love of the mountains is evident in his evocative depiction of the 1980 eruption. The Phoenix did as wonderful a job with the delicate, gamelan-inspired Spirit Lake movement, as it did with the thumping bass drums, discordant horns, and clashing cymbals in the climactic Volcano movement. What the SE 120 chain added was an even bigger sense of scale and grandeur. Take the bass drums. The Phoenix still had the advantage on bass depth and sheer physicality, but with the SE 120 chain the drum strokes themselves had more texture and detail. Similarly, the cymbal clashes had a more real, metallic timbre. Finally, the soundstage size was larger with the SE 120 chain, with a greater sense of front-to-back depth.
This comparison illustrates why reviews shouldn’t be about assigning winners and losers, but to describe differences and set the context to help readers make their own decisions. On the one hand, the spaghetti solution with the SE 120 clock “won” this comparison, but did the sonic differences justify the almost-double cost? That depends.
If you already own some of this spaghetti, and don’t mind the plethora of boxes and cables, this approach does have the advantage of upgradability — like when a better reference clock like the SE 120 comes along, allowing you to scale up SQ. In such a situation, would I recommend you reverse course, sell the spaghetti, and buy a Phoenix? Of course not.
On the other hand, if you’re starting with a clean slate, and/or value simplicity and order, I would strongly recommend the single-box Phoenix. Yes, it’s pricey, but it’s still a bargain relative to the comparable spaghetti, and benefits from the synergy a careful designer can bring to the table.
As I said in the introduction, I was very interested to see how well the single-box PhoenixUSB Reclocker could perform in two scenarios. First, as an upgrade for Innuos ZENith SE users, and second, in comparison to my incrementally crafted, but terminally ugly, spaghetti solution.
The reason DIY tinkerers like me end up with spaghetti digital chains is not because we love the clutter, but because manufacturers have not (yet) offered components that deliver the SQ we are able to achieve. I’m happy to report that the Phoenix addresses this gap for USB. It delivers outstanding improvement to ZENith SE owners, allowing them to approach the sound quality of the much-more-expensive Statement server, while preserving their investment. Even more impressive is the fact that the Phoenix outperforms my already-pricier spaghetti chain, providing sound quality that is richer, denser, and with more bass depth. It was only with the REF 10 SE 120 in the chain that I could improve upon the Phoenix, but this was at almost double the cost.
USB reclockers are an important element of a highly optimized digital audio chain. If your budget allows it, forget the spaghetti, and get yourself a Phoenix. Its addition in your USB chain will bring a huge grin on your face!
- Product: Innuos PhoenixUSB Reclocker (378 KB PDF)
- Innuos Main Page
- Price (MSRP): 2249 GBP · 2499 EUR · 3149 USD · 4099 CAD
Community Star Ratings and Reviews
I encourage those who have experience with the Innuos PhoenixUSB Reclocker to leave a star rating and quick review on our new Polestar platform.
Music Computer: Custom computer: Phantom Gaming Z390-ITX/ac, i7-8700K, Apacer 2x8GB Industrial ECC RAM, HDPlex H3 case,
64GB M10 Optane SSD for OS, JCAT Net Card Femto, running Euphony OS with
Stylus or Roon+StylusEP music software
Music Storage: Synology NAS DS916+ 4-bay, attached to router via Ethernet
Headphone Amplifier: Cavalli Liquid Gold
Headphones: Sennheiser HD800 (SD Mod), Meze Empyrean, Abyss AB-1266 CC
DAC: Chord Hugo M-Scaler + DAVE
USB Regenerator: SOtM tX-USBultra SE
Ethernet Switches: SOtM sNH-10G, Uptone EtherREGEN (ER)
Reference Clock: MUTEC Ref 10 10MHz clock driving the tX-USBultra SE and switches
Power supplies: Paul Hynes SR-7 DR (dual regulation) for M-Scaler, ER & tX-USBultra,
Paul Hynes SR-7 SR (singe regulation) for sHN-10G,
Paul Hynes SR-4 for JCAT Net Card Femto, HDPlex 400W ATX LPS for music computer motherboard ATX & EPS
Power Details: Dedicated 30A 6 AWG AC circuit, Sound Application TT-7 Power Conditioner
Power Cables: PS Audio AC-12 (wall to P12), Cardas Clear Beyond (Cavalli Amp, SR-7), Cardas Clear for all other components
USB cables: Sablon Reserva 2020 USB
Clock cables: Habst 5N Cryo Pure Silver 50Ω a nd SOtM dCBL-BNC 75Ω
Ethernet cables: SOtM dCBL-Cat7, Supra Cat 8
DC cables: Ghent Audio custom OCC JSSG360 ATX and EPS cables Paul Hynes fine silver (SR-4, SR-7)
Interconnects: Cardas Clear XLR balanced
Headphone cables: Transparent Ultra cable system, Cardas Clear balanced and SE cables
Accessories: Synergistic Research Tranquility Base XL UEF with Galileo MPC
Many thanks to the following companies for supplying cables and accessories to aid in this evaluation:
- Cardas Audio, for a full loom of Cardas Clear cables.
- Transparent Audio, for the Transparent Ultra headphone cable with a full complement of headphones leads and source terminators.
About the Author
Rajiv Arora — a.k.a. @austinpop — is both a computer geek and a lifelong audiophile. He doesn’t work much, but when he does, it’s as a consultant in the computer industry. Having retired from a corporate career as a researcher, technologist and executive, he now combines his passion for music and audio gear with his computer skills and his love of writing to author reviews and articles about high-end audio.
He has "a special set of skills" that help him bring technical perspective to the audio hobby. No, they do not involve kicking criminal ass in exotic foreign locales! Starting with his Ph.D. research on computer networks, and extending over his professional career, his area of expertise is the performance and scalability of distributed computing systems. Tuning and optimization are in his blood. He is guided by the scientific method and robust experimental design. That said, he trusts his ears, and how a system or component sounds is always the final determinant in his findings. He does not need every audio effect to be measurable, as long as it is consistently audible.
Finally, he believes in integrity, honesty, civility and community, and this is what he strives to bring to every interaction, both as an author and as a forum contributor.