Jim Austin | Apr 24, 2024
Original link:
https://www.stereophile.com/content/silent-angel-bonn-nx-network-switch
The following content is obtained through ChatGPT/Google Translate, and due to different locales, there may be some differences from the original expression!
Magazine Description: Stereophile is a magazine dedicated to high-fidelity audio equipment reviews and audio technology coverage. It was founded in 1962 by J. J. Founded by Gordon Holt, one of America's most respected audio magazines. Stereophile focuses on evaluating a variety of audio equipment, including speakers, amplifiers, headphones, turntables, digital audio players, etc., and provides professional reviews and opinions to help consumers understand the performance and characteristics of various audio devices. The magazine also regularly reports on the latest developments, technological developments and trade fair events in the audio industry. Stereophile is known for its professional review team and authoritative reporting, and is favored by a wide range of audio enthusiasts and professionals.
As each season changes, new audio-grade network switches are introduced. These products can cost hundreds or thousands of dollars. While they all perform the same basic functions as the network switches that cost about $30 at Best Buy (except for slower speeds in some cases), manufacturers all claim that they build switches with higher standards for better sound quality.
For all the signal conditioning devices that operate entirely in the digital realm, and especially those that work at the packet level (which is more correctly called the "frame level" at the local end of the router, but even few experts are able to distinguish between them), there is a controversy over the sound effects of a speaker-grade network switch, which is usually between those who insist on hearing the difference and those who theoretically insist that there is no possibility of a difference.
Years ago, I would almost certainly be on the side of those who refused to listen for theoretical reasons. Since then, several things have happened, and also as I have grown older and gained experience, I have discovered that I am not as smart as I thought I was when I was younger. This is accompanied by a greater acceptance of uncertainty, a willingness to forego the certainty of scientific proof, and to accept things with less evidence, such as rigorous double-blind testing. I'm not alone in the industry in this regard. I often ask and find that people who work in high-fidelity R&D rarely need such rigorous evidence, they develop their listening skills, and then believe and accept what they hear.
The next thing that made me change my mind was a series of demonstrations that proved the effectiveness of signal conditioning on the digital side, and the results were very clear. The first was to replace the power supply of the purely digital audio server with an improved, lower-noise version, which resulted in a noticeable change in the sound delivered to the speakers through the DAC. The second demonstration involves reclocking the USB signal. USB, like Ethernet, is a packet-based technology; A reclocked USB signal is a reclocked packet, not the audio data within it. When using an asynchronous USB transfer, the data is reclocked at the receiving device – however, I heard a noticeable change in musical performance again that I never expected. These experiences, especially the last one, made me realize that signal conditioning at the packet (more precisely, frame) level could bring about changes (perhaps improvements) in sound performance; Footnote 3).
I chose the Bonn NX as my first network switch test device because it was the most ambitious product I had come across. If you're going to judge the sound effects of a new technology, it's best to start with the one that seems most likely to make a difference.
The Bonn NX ($3,999) from a Chinese company that also makes the much less expensive Bonn N8 Pro ($599) caught the attention of audiophiles a few years ago. Xianlai also manufactures servers, streaming players, linear power supplies, and other products.
To quote Anthony Chiarella, the U.S. agent of Xianlai, "The cheaper N8 Pro was designed as an entry-level product and built for the price, while the NX was conceived and built into a cutting-edge product with no cost limitations." Both devices use the same core electronics — the same switching board (more on that later) — but in NX, both the power and clock circuitry have been redesigned, with the former using what designer Chorus Chang calls "direct clocking" technology.
What is "Direct Clocking" Technology? The N8 Pro uses a 10MHz clock source, which must be converted to 25MHz to drive a 10/100/1000Base-T network switch. In contrast, the NX uses a 25MHz clock, so there is no need for conversion. This 25MHz clock is of the TXCO type and promises to provide "extra low phase noise output". There's also a 25MHz word clock input (Xianlai also made an external word clock, running at 25MHz, called the Genesis GX), probably for when you want to coordinate the operation of a network switch with other devices on the network, as is sometimes done in the studio.
In my opinion, the NX far surpasses the N8 Pro in terms of mechanical design. The N8 Pro is basically a circuit board in a box with rubber feet. The NX's enclosure is a precision design designed to reduce vibration. The chassis is a two-layer structure, with an inner layer made of unspecified "metal", an outer layer made of aluminum, and a shock-absorbing spacer pad in the middle. The NX's bracket is very wide – 2.36 inches – made of stainless steel with shock-absorbing rubber O-rings embedded in it.
But how has that core switching circuit been improved? It is "specifically designed for use in high-resolution streaming networks" and "delivers high-purity signals to connected components, a performance that cannot be achieved with a typical network switch," Xian said. The clock has its own power supply, which is isolated from the power supply of the switching circuit. There is also an inductor to remove high-frequency noise from the AC line. (The inductor can also operate in both directions, and it also prevents NX noise from entering the AC line.) ）
And what about the swap board itself? Chorus Chang told me in an email, "The switch board design is optimized to reduce crosstalk between ports and reduce noise from the digital section to the analog section. (Ethernet signals are analog.) As a result, the quality of the Ethernet signal is better. "These improvements are shared by both the Bonn NX and the N8 Pro.
The NX has an external clock input, which may be superior to an internal clock. The use of an external clock raises the eternal question of which is more important, clock accuracy or clock proximity. NX can also use an external power supply. (The built-in power supply is said to be a "radar-grade" switching power supply; The radar level seems to mean that the noise is very low. There is also a ground box that further reduces noise by providing a low-impedance path to ground for the enclosure.
This network switch costs $4000, but it's packed with features inside. The key question is whether or not these features have an impact on audio.
The sound of the network switch
I have a relatively simple network setup, one that supports my audio system and other network activities, although when I'm listening intently, there's little non-music network activity other than the occasional email (and then my attention is distracted). When I'm listening carefully, I'm tempted to turn off the Wi-Fi, but usually I keep it on because it supports network audio "remotes" like Roon. I'm not a tablet user; My Mac notebook is my go-to "remote control". I did turn off the Wi-Fi briefly for comparison, but most of the time I kept it on.
During the audition of the Bonn NX, a standard, high-quality CAT6 Ethernet cable – not a speaker-grade cable – was connected to the Bonn NX from my router in the back room; From there, the CAT6 cable connects to my Roon Nucleus+, the Innuos Statement server-streamer, and one of the two streaming DACs: the CH Precision C1.2, followed by the Bel Canto Black (under review). Another CAT6 cable connected to a separate, inexpensive network switch 6 feet away for my non-music network needs, including a Wi-Fi access point.
In the late 90's of the 20th century, I did some basic listening training by comparing MP3 samples at various bitrates from ABX, trying to choose between a lossy compressed sample and a full-resolution sample. I remember those samples, including the "Fast Car" and the infamous jingle bell. At first, I did pretty badly - then I started noticing specific exceptions that were present in lossy compressed samples that weren't present in uncompressed samples. (I don't remember what they were.) After solving the puzzle, I got it right every time, although it got harder as the bitrate increased and the compression ratio decreased.
To feel confident about any new hearing test, you need to find a specific clue to listen to. Some differences are suitable for this, while others are not, yet both types of variations are real. As I wrote at the beginning of this post, I no longer need absolute certainty to believe in something, but it's still something I'm striving for. The more certainty, the better. When listening to music with the Bonn NX, I immediately (Note 6) heard, or thought I was hearing an intensification. Most people know what I mean - everything seems more vivid, even silent ("darker background" and so on). This is not a qualitative change, but a change in degree. Such changes are unreliable and difficult to catch. It's hard to keep that sound in your head, and this change may just be due to a greater focus on the music. This is a common review risk that requires experience to overcome. I've overcome it to a certain extent. Although my confidence is not 100%, it is much better than 50:50.
Then I found a sign – a clue – even though it wasn't ideal. It has to do with the structure of the soundstage, especially in large music, such as symphonies: there is a better separation between them. By "better separation", I mean that the gaps between the individual sections (first violin, viola, double bass) become darker (or we can say darker gray) and darker. This gave me a concrete auditory cue. I was able to hear the same effect, though not as dramatically as in the big music – for example, group jazz. The soundstage is more layered in front and back, which is a small improvement in audio (Note 7). Its main value is not in the musicality, but in my confidence that the changes I hear are real. I had reason to believe then that this network switch was influencing the sound. I would even say that this effect is an improvement. I wouldn't say it's dramatic.
What's going on? How can I improve how the network signal affects the audio? The honest answer is: I don't know. But I've asked a lot of people who work in hi-fi audio, including pure digital audio product designers, to focus on the same thing in their explanations: noise. Sometimes it's noise inherent in the signal, and sometimes it's extra noise due to increased computation, such as error correction. This explanation is somewhat vague and not particularly satisfying – we still lack a concrete, verifiable mechanism – but that doesn't mean it's wrong.
Perhaps the most important question, assuming you accept my conclusions (of course, you are free to hear for yourself), is how much is this change worth? I suspect that Tianli doesn't expect to sell a $4,000 Bonn NX to someone who owns a $5,000 sound system; That's exactly what the cheaper N8 Pro is for. If your system is worth 10 times that number, maybe it starts to make some sense.
A few months ago, I put a cable isolator underneath the speaker cables. Are there any sound improvements? Maybe there is, maybe not, but that's not why I do it. I just like the cables that are carefully placed, neatly worn, rather than being placed on the floor to collect dust. I decided that the routing of the cables deserves careful attention, just like everything else in my system. Seeing it as a matter of hygiene, a basic hi-fi routine, a respect for the system and the music, it brings me so much joy. Attention to detail always pays off, if not in terms of the sound that directly affects the system, but also in terms of calm, which is probably one of the best hi-fi improvements.
One final question to consider: In the age of streaming, where does your hi-fi end up and where does your computer network start? Is the network switch part of the system, or is it located outside the system? What about routers?
I mean, sound aside, maybe a $30 network switch from Linksys or Netgear isn't a good fit for a perfectionist's sound system. A $4000 network switch may not be suitable, but it may be. How much you're willing to invest depends of course on your inclinations, your resources, and the value of your system. I'm fairly sure that a network switch can bring a small improvement, although I don't know how it works. Another listener might think that the difference is significant, as people's sense of proportion can vary greatly.
What if you think it's an overpriced useless product? Then don't buy it.
Xianlai NX switch JD shopping link: https://item.jd.com/10093483150919.html
Footnote 1: I am happy to say that there is a third group: those who listened carefully and did not hear any changes, but with the humility of heart, and would not consider this to be clear evidence that these devices are a panacea. These few admirable souls admit that just because they can't hear it, or haven't heard it yet, doesn't mean it necessarily can't work.
Footnote 2: I'm more than open to any proposed improvements that require occult or occult "mechanics" that go beyond physics, and there have been many such proposals over the past few years. I'm just saying that I don't need to have a clear, concrete physical mechanism in my mind to accept evidence of my senses. Specific explanations can come back later. JA and I have both written about this before; See, for example
stereophile.com/asweseeit/787/index.html
stereophile.com/content/nothing-too-wonderful-be-true
stereophile.com/content/audio-meet-science
Footnote 3: I am careful not to refer to any change as an improvement. The first key question is whether there have been any changes – this is a matter of fact, although not easy to verify. Improvement or not is subjective: some people may prefer separation, while others prefer continuity. In fact, in the first demo described above, one audience member insisted that the supposedly poor power turned out to be the best sound. I didn't hear it that way, but I don't want to argue either. You should look for the sound you like.
Footnote 4: I'm not a networking expert, but it's not clear to me why the N8 Pro uses a 10MHz clock and how common 10MHz usage is in speaker-grade networking switches. 10MHz seems to be a common frequency for external word clocks in digital audio sources, and is often used internally. Xianlai's literature makes obscure references to satellite navigation systems known as GPS; Perhaps this 10MHz clock was derived from the universal "atomic clock" used by GPS. All this does not fully explain why the N8 Pro uses 10MHz, while by default the universal network switch requires 25MHz, regardless of whether they are implemented directly or by converting 10MHz.
Footnote 5: Wi-Fi is a notorious source of high-frequency noise, and in my apartment building in New York, turning it off will only remove the most intense and localized sources of electromagnetic radiation. At the moment my MacBook Pro has detected nine other networks in this building and surrounding buildings.
Footnote 6: The idea that a network switch will sound better after a run-in is crazy even for me.
Footnote 7: At least one other reviewer claims to have heard an improvement in transient performance. I also tried to hear, but didn't hear.