AQVOX audio grade switch

New to this forum, had to register so I could respond to this, hoping the moderator is unbiased.

What the hell?

Ok first off. Doraymon, I assume you mean optical as in SPDIF or similar? because if not, then I think you must have misunderstood something. If you mean that your laptop is taking in pulses of light, like a fibre-optic connection would, and on it's own converting it into data, and that is improving your DAC's audio output, then, really, no offense meant, but all hope is lost.

Speaking with experience amounting to 10 years of music production life coupled with 5, soon 6 years of electronics engineering, the AQVox ethernet switch does nothing, and is a scam. Absolutely nothing about it is improved, except the physical sturdyness which I assume is an accidental byproduct of all the superglue, hot glue and resin compound they added all over the thing.

No components have been swapped out for improved ones in the board, solder jobs have clearly been re-done on the connectors (likely to accompany their own branded overpriced cables while excluding everything else, and the joints are brittle and not anywhere close to professional. That's it. That's all they did, except add in an extra IC that in it's current configuration literally doesn't do anything, and swapping out the xtal for a cheaper one. Under analysis, practically every component introduces more (not less) noise than the ones on the base model, and I consistently encountered bit-loss and checksum errors when analyzing the chain, more so than on the base model they've used.

Not to mention the fact that all of this should have been obvious to anyone who thought twice about it. This is TCP we're talking about. It goes through error-correction, jittercompensation, delay compensation, buffers, and it's even goddamn encrypted. No matter how hard you try, you literally can't make improvements to it that you couldn't by just getting a decent ethernet switch from a reputable brand, which AQvox absolutely isn't. They're hated by practically everyone of note in both the music industry, and in all engineering communities I've managed to find it talked about in. 

And that's not even getting into their marketing, which is filled with outright plain lies, and now there's a pending class action lawsuit coming their way because of it. (which is good news though because you guys can get your money back with interest when, not if, it succeeds.)

But seriously. I can't stress this enough. The components are WORSE. they made the base model WORSE, and slapped a 3000% price increase on it while lying in their marketing. It's more outrageous than if they had just straight up not shipped it to anyone who ordered it.

Buy the base model, if you have the AQvox one, open both up and see for yourself. If you have an oscilloscope you can really see for yourself. This is a scam, and the fact that placebo has worked so well on you guys is a testament to how low this company has fallen in order to extort money out of people who don't know better. I don't blame you guys for falling for it, but I do blame you for defending it when challenged about it.

The topic mainly concerned itself about which linear supplies 'sound' the best with the AQVOX,  rather than the device itself. Some have the Oem model, others the AQVOX. 

Many IT gear has improved using linear power supplies, since the hash from SMPS is noticeable, discussed at length here.

That article is several years old now, things change, maybe the AQVOX is now in a drawer superseded by another switch design.

Thank you for the insight of what's inside,  a few manufacturers are caught out by changing a few innards to switches and are exposed here.

What we don't appreciate is a holier than thou approach, that could have been left off at the end. Anyway the post can't be edited now unless you approach the owner.

@LiVeen

Thank you for your rant 😉

I agree with your expertise on the technical subject 100%, but as far as I understand the rules in this forum, there is an agreement to give people the head space to exchange their beliefs in a friendly manner, as long as they don't agress/attack people with superior/different expertise, who explain reality to them. For pure reality there is an objective subforum, and from my experience I mayy cofirm there are many topics in a grey area, which may have a low key moderation.

Having the forum's history in the rearview, I may express my feeling of a great loss when a couple of well respected members left this place, as the presence of unregulated subjectivity had been too much to bear for them. 

The thread you're referring is dormant for the last 3 years (thus dead & buried in internet terms), and if I am not mistaken, I might connect it to your extremly miserable experience - objectivly and subjectively  - with that sh**** snake oil device , that you feel the need to criticize former owner/posters now? Imho, not helpful on a social level, even your arguments prove to be right.

Hi DuckToller,

Let’s take my network as an example. The broadband connection arrives at my house via a copper coax cable into a DocSys 3.0 modem. The stream then travels to a router via an ethernet cable. The stream is then converted  into a radio signal which is transmitted to a wi-fi - ethernet bridge, where it is again converted into ethernet format and transmitted to a cascade of switches prior to entering a server, where it is converted to a USB signal and sent to the DAC, where it is again converted, this time into an analog music signal prior to being amplified and used to drive the loudspeakers. Along the way the signal is held in a series of buffers, which again require an electro-magnetic format change. 

All the time the signal is travelling along a wire, its doing so as voltage current.  The voltage originates from AC, which is converted to DC, then encoded into various formats, with various timing, depending on the protocol.  At each conversion, the timing and the voltages are recreated i.e resynthesized. For example, a radio signal is received, converted and re-transmitted down a wire as an ethernet signal. 

In digital we talk about 1s and 0s, but such a thing doesn’t actually exist in reality. Instead the 1s and 0s are ‘represented’ by various electromagnetic states, depending on where they are in the network and in which form. 

Take a bit traveling down a wire. In essence a bit is simply a very small portion (bit) of time with an associated binary voltage level, low = 0, high =1 .  Given that the bit exists physically as time and voltage, it has quality parameters. How accurate and reproducible is the time vs the standard? How well are voltage transitions (between low and high) handled?  How much additional electromagnetic noise is included? 

Each of the conversions along the network has a sonic footprint….and the quality of the input directly affects the conversion process and thus the quality of the output.

So getting back to my original point, if you take a very high quality, low noise, well timed input and pass it through a noisy, badly timed process, the output of that process will be, by definition, noisy and ill timed.   It may well still meet the minimum quality criteria of whatever network protocol is operating or be subject to the protocols’ error correction in order to reach its destination as ‘bit perfect’, but the quality of the DACs input will be reduced, requiring that the DACs electronics and logic processors get very busy cleaning up and refining the input. When a processor has a lot to do, individual tasks typically gets a lot of interrupts, latency increases and the general noise level produced by the processor and its associated electronics increases.   

8 minutes ago, Blackmorec said:

Hi DuckToller,

Let’s take my network as an example. The broadband connection arrives at my house via a copper coax cable into a DocSys 3.0 modem. The stream then travels to a router via an ethernet cable. The stream is then converted  into a radio signal which is transmitted to a wi-fi - ethernet bridge, where it is again converted into ethernet format and transmitted to a cascade of switches prior to entering a server, where it is converted to a USB signal and sent to the DAC, where it is again converted, this time into an analog music signal prior to being amplified and used to drive the loudspeakers. Along the way the signal is held in a series of buffers, which again require an electro-magnetic format change. 

 

All the time the signal is travelling along a wire, its doing so as voltage current.  The voltage originates from AC, which is converted to DC, then encoded into various formats, with various timing, depending on the protocol.  At each conversion, the timing and the voltages are recreated i.e resynthesized. For example, a radio signal is received, converted and re-transmitted as an ethernet signal. 

In digital we talk about 1s and 0s, but such a thing doesn’t actually exist in reality. Instead the 1s and 0s are ‘represented’ by various electromagnetic states, depending on where they are in the network and in which form. 

Take a bit traveling down a wire. In essence a bit is simply a very small portion (bit) of time with an associated binary voltage level, low = 0, high =1 .  Given that the bit exists physically as time and voltage, it has quality parameters. How accurate and reproducible is the time vs the standard? How well are voltage transitions (between low and high) handled?  How much additional electromagnetic noise is included? 

Each of the conversions along the network has a sonic footprint….and the quality of the input directly affects the conversion process and thus the quality of the output.

 

So getting back to my original point, if you take a very high quality, low noise, well timed input and pass it through a noisy, badly timed process, the output of that process will be, by definition, noisy and ill timed.   It may well still meet the minimum quality criteria of whatever network protocol is operating or be subject to the protocols’ error correction in order to reach its destination as ‘bit perfect’, but the quality of the DACs input will be reduced, requiring that the DACs electronics and logic processors get very busy cleaning up and refining the input. When a processor has a lot to do, individual tasks typically gets a lot of interrupts, latency increases and the general noise level produced by the processor and its associated electronics increases.   

 

 

It is quite amazing that we get audio out of our speakers, when looking at the circuitous path it all takes. Seriously. 

On a related note, and keeping in mind that this is the objective sub-forum, I have a lot of experience with audio over Ethernet in my systems over the years. I also come from the IT world. I firmly believe nothing beats personal experience, and nothing is more misleading than folks trumpeting how they believe things should work, based on textbook knowledge. 

In my system I can do things on my computers that change the sound of the audio (running over Ethernet) in ways that many objective leaning folks wouldn’t believe because they have no experience with it. Changes from barely audible to shifts in pitch that sound like auto tune to straight up distortion sounds. The common objective belief that digital either works or it doesn’t, or that Ethernet is error corrected and resent, is good on paper. But, in my experience, I found there’s much more to it when listening to audio versus pulling up a spreadsheet. 

Hi CA, 

Agreed!  

My background is in complex analytical systems so I try to apply what I learned in over 45 years along with 50 years around hi-fi.

Around 4 years ago I started to build an entirely new hi-fi system for local and remote streaming. To get started I first bought a networked speaker and attended a series of networking tutorials ( Thanks to my career I was already familiar with networking protocols but knew very little about setting up local area networks for audio). I started out with the (common) belief that all I needed was a bit perfect stream to a high quality server, but I soon found that the quality criteria applied to the server brought equally good results when applied to the rest of the network.  Those criteria include:

EMI control (internal and external)

Vibration and resonance control (internal and external)

Timing accuracy and stability

Power supply quality 

Cable losses and screening

Network traffic management (Minimizing latency, interrupts and the resulting noise generation)

A key learning was that bit perfect was only one of several important quality criteria 

What I realised very early on is that unlike analog, where the signal is generated only once and must be maintained in as pristine condition as possible, in digital the ‘data stream’ is generated multiple times and can be heavily improved or degraded along the way.  These improvements and degradations have a major impact on final sound quality in that the cleaner and better quality the  date stream, the easier a job the brain has in interpreting sound waves as music with intricate timing, phase, amplitude and spatial attributes.

For IT folks, the network has a single role. To move files. For audiophiles it has 2 roles a. To move files and b. To improve the quality of those files in order to increase the resulting sound quality.

The flaw with the above, nicely written and evocative view of the signal path is that it does not reflect reality.

There is no correlation between the "quality" and "timing" of digital transmission over a data network with the resultant wiggly voltage that comes out of a dac - as long as the basic function of the network components allow delivery of viable data (I won't say error-free because there a plenty of robust mechanisms built into various elements of even basic networks to correct for poor quality data transmission).

This contradicts Mr Blackmorec's description because it there is no link between the voltage representing the bits or the timing between bits. Even in the simplest scenario connecting a music server with digital audio files stored on it with a network-enabled dac via a simple switch this is true. When the ethernet receiver receives some bits (one bit is useless) those bits are passed not to the dac but into a storage buffer, so that packets can be stripped of routing headers, sequencing can be checked and error detection processes implemented. At this point there is no "timing" information because the data is stored in ram in a non-temporal charge bucket and the electrical charge that enables that is generated within the circuitry of the receiver/processor. No "time" passes for the data whilst its stored.

Beyond this point there are a significant number of software processes which must transfer the data from ram, perform a process and write the data back into other ram buffers before the data is passed to the render process which is the interface to the dac processor itself and the timing of the delivery of "bits" across this interface is the first stage in the entire process where timing becomes even remotely important. There is simply no way that this transfer of information can carry the electrical or timing signature of the network over which the data was transmitted. There may be EMI, there may be noise at any stage (and lets face it there's a SOC processor or an FPGA (often both) sat between the ethernet port and the "dac" with all the opportunities to pollute the voltage rails and even the leading/trailing edge of the signal voltage transitions on the IIS interface to the dac chip (or FPGA or ladder) but those are local sources of pollution which the hardware at each of these stage is designed to mitigate.

But let me finish by saying this: If you hear a difference and you want to buy one or another of these doodads, good for you. 

I have never understood the "objectivist stance that seek to reduce other people's experience to a supposedly superior objective reality.  If you spend a few hundred bucks you can afford on something that someone else can "prove" is snake oil - if you perceive the change you expected and you're happy, where's the harm? Everyone gets to make up their own mind, surely? 

10 hours ago, LowOrbit said:

I'm not trying to deny your experience - that was my last point. I'm perfectly comfortable with you hearing an improvement. 

 

But engineering is not "blah blah blah"  - modern civilisation is entirely built on those solid principles. Audio engineering is really not that difficult or built on special cases. 

Thanks for the polite replies 😊 Just to point out that optimized network traffic,  improved, low noise and impedance, high stabilty power supplies, vibration control, EMI reduction, improved clock accuracy, lower clock noise, mains borne noise reduction, ethernet cable borne noise reduction, optimised cable screening amongst many other things are all based on solid engineering.  Also bear in mind that audio engineering and network engineering for audio are not exactly the same thing, at least they weren’t 

Also worth considering is that there is R&D going on in the areas I’m discussing in order to isolate areas in the computer/network where improvement brings sonic rewards. That R&D is funded by certain commercial audio companies. When they find the affects that certain network related conditions have on the eventual sound quality, that information isn’t necessarily splashed around, rather its used to create competitive advantage. You’ll typically see these areas in designs from companies like Taiko, Aurender, Innuos, etc. These companies have theories and ideas, make extensive measurements and confirm subjectively.  Their products take the mitigation of these problems to new levels, giving rise to audio optimised networked products with improvements to all areas mentioned above. Next level products currently in development include new servers, software and OSs, switches, routers etc. 

When i started developing my latest streaming system i shared your belief that all I needed were bit-perfect files arriving at my server.  My first step was to get additional training in networking via a series of podcasts and my beliefs remained unaltered.  Realising that i still lacked certain high level technical expertise and experience on the relationship between audio and IT I set out to locate consultants and experts for advice. Those experts I found in 2 places; the owners of companies like Taiko and Innuos, who generously share some of their findings and developments and a small group of individuals on a Computer Audio DIY forum. These individuals are highly educated and experienced IT engineering specialists who also share a long time passion for high end audio. They have spent years building SoTA servers, modifying and improving various networking components and evaluating all the very latest commercial developments in audio computing and networking. Ask them about how different RAM, network cards, disc drives, oscillators, cables, CPUs, SW, power supplies, capacitors, chokes, input and output sockets, BIOS settings etc impact sound quality and they’ll give you answers based on both their engineering and IT knowledge and on their subjective experience. That’s the kind of advice you need if you are trying to achieve great sound quality. Over the past 4 years i have carefully followed their generous advice to the point i now have suppliers of equipment who help me implement their latest discoveries and improvements to achieve some absolutely wonderful SQ improvements. For example, i have just upgraded all the power supplies serving my modem, router, bridge, switches and server with the latest ‘ARC6’ Mundorf capacitors and choke, mounted on a specially developed panzerholz and resin, constrained layer anti-vibration circuit board.  Another example are the highly specialised DC cables designed by a renowned IT audio engineer and built by one of the experts I mentioned.  I even delved into the DIY side myself, taking a commercial network bridge and extensively modifying its vibration control and power supply 

All i can say is that audiophiles who work to achieve the very best SQ deserve to know about these developments and what it is possible to achieve by improving their networks. ‘Advice’ such as some of the above simply cuts audiophiles off from one of the most effective areas for improving sound quality. It may be well intentioned, but its also very outdated. 

Hi Low Orbit,

I get the feeling that if we keep on discussing we’ll eventually get a point of complete agreement. 

Let me try to explain the ‘improvement or upgrade’ logic that I’m seeing in networks intended for audio.

A network comprises many different components and the data therein can exist in several different states, formats, forms so you are correct, there is no direct connection. But what there is is influence. These devices are all connected together, all have a sonic character and all influence final SQ. The quality of a feed into a component influences its output. The better it’s input feed, the better it’s output. This is pretty easy to clearly observe (I dont use the word ‘prove’ in deference to the objective nature or the forum). My network comprises DC4 ARC6 power supplies, extensive vibration and resonance mitigation, a cascade of ultra low noise 3ppb clock-accuracy switches and a loom of SR Atmosphere cables with all screens star earthed. A short while ago my ISR sent me a new replacement modem in order to handle their new 1gHz stream.  Unfortunately its power inlet did not match my own Mundorf Silver/Gold JSSG360 DC cable, so I had to use the OEM supplied switch mode power supply. The unit needed some time to run in and sounded a lot less dramatic, vibrant and enjoyable. It was the kind of sound that would soon find you surfing the forums. Within a couple of days however it was sounding much better…not a lot of difference from how it was previously I would judge.  In the meantime I’d arranged with one of my kind suppliers to modify my router by remaking the circuit board to DC connector path. He used Mundorf Ag/Au cable and a high quality connector, to give a nice low impedance connection. I returned home, plugged it into my DC4, rebooted the network then left it for a few hours. That evening I sat down, switched on the amps, and I immediately heard that the -40dB, 128kbps, MP3 based Swiss Radio Classic transmission sounded better. It instantly grabbed my attention with fuller sound and greater vibrancy.  Cranked up to -13.5 dB and streaming full fat high resolution files the system has the ability to project the illusion of a 360 degree venue atmosphere with musicians performing within. Some recordings, especially electronic music can be like sonic fireworks display while other music is felt as a  most warm, gentle, caress.  In a few words the system took a quantum leap forward. The system became both physical and more cerebral at the same time. Big hit!  So was the DC4 that much better? Well obviously yes, it was, but in addition the device it was powering had to have the ability to use that superb power without polluting it with an identifiable own identity or loads of noise in order to produce an output that ultimately produced a far more ‘tactile’ and ‘vibrant’ musical experience. But in order for me to hear those improvements made at the modem level, they have to pass through every level of the network. And the only way to explain that ability, given the variation and differences in forms and formats along the way is better in = better out, but this process only works when the device receiving the “better in” has the ability to produce “better out”. If its noise and timing specs are worse than the spec of the incoming signal, the output will be of a lower spec. 

So if you arrange your network as a series of points designed to serially upgrade the output, then you have a series of improved inputs and correspondingly improved outputs and your improvements start to compound, the modem making the router input better, making its output to the bridge better, which made its output to the first switch better and so on. What I’m saying is, regardless of what part of the network process you are talking about, a better input results in a better output, sound-quality-wise when the following stage is superior. 

Essentially what I  am typically refining in my system are:

Network topology to get as close to an ‘Audio-only-network’ as possible. Unrelated network traffic and activity does not help the audio cause

Division of labour…..making sure there are no bottlenecks in the network, exclusivity for audio and video streaming and adequate signal strength and bandwidth for the rest 

Power supplies, best when all of a similar designs throughout to minimize interactions and formation of noise harmonics. Taiko posted some very interesting noise spectra in this regard. 

DC cabling plays an extremely important role and for reasons I don’t fully understand, makes a VERY significant difference to the final sound. 

Vibration and resonance control is important and has an impact at every stage.  

For any input that needs a clock signal, the more accurate and precise the input, the better the quality of the output, when the following clock is superior. 

What I am saying is that a digital audio network rewards when each stage is optimized and the handshake between modules is provided with an optimum signal ie one lacking timing errors, noise  and having an optimum voltage profile.  The more noise that gets eradicated or avoided, the better, because even when there’s no connections to the DAC that noise is still having a negative impact by reducing the quality of interim processes and their outputs.