barrows

Yes, the main board is new in both the oR and the Signature (Deluxe versions), and it is nearly identical in both. The performance is quite different though due to the power supply in the Signature and the fact that it has a real chassis (think vibration control and lower processor temperature) and the close coupling between the power supply and the main board: digital products perform best with ultra low power supply impedance, and external supplies make this almost impossible to achieve due to the impedance of the connecting cable between devices and the connectors themselves.

I am not sure of the question, but here is the news on Signature Rendu deluxe Completely new main board, with a new, better, switch chip, much improved clock distribution circuit, and various other small improvements. The power supply improvements include new refinements in the transformer secondary damping circuit, and the final decoupling filter before the discrete regulator circuit. Generally speaking, the Signature Rendu is the flagship product of Sonore. The Signature includes a very high quality, made in the US chassis, which results in much better control of vibrations, it includes a no compromise power supply design using the best parts available, including a custom built, made in the US, transformer designed specifically for this product, and a discrete regulator circuit which performs at a lower noise level and lower output impedance than any IC based circuits can. Internally, the AC side of the product (AC input wiring, filtering, and transformer) is shielded for both RF and EM from the signal processing and DC power side of the component. The Signature Rendu is designed to achieve the best possible performance without concern for compromise regarding parts quality costs, etc. Each unit is hand built and tested one at a time here in the US.

@SQFIRST, as you have the X-1, a 1Gb device, I assume when you reference using 10 Gb hardware you are talking about using 10 Gb hardware but running it at 1 Gb, at least to the X-1, right? Are you saying you run other Network links at 10 Gb, and only the final link to the X-1 at 1 Gb? I am considering trying some 10 Gb capable hardware, but running at 1 Gb, as I am interested... also to consider is that the Sonore opticalModule Deluxe, which provides the fiber link to my endpoint uses a very good clock (a better clock than what I have seen in 10 Gb commercial hardware), so I also wonder if jitter is the source of the "problem" you hear with 1 Gb hardware then perhaps that "problem" would not be present with a well designed piece of hardware with a very low phase noise clock, even though it nominally is 1 Gb (oM-D).

Sterling silver is an alloy, usually of copper and silver-why one would want to use such an alloy for a USB signal makes little to no sense. Sterling makes silver stronger, and is used for jewelry and silverware. For USB signal carrying wires, the best conductor is a solid core copper wire, with a heavy plating of the purest silver possible. The reason for this is that the signal is carried on the very surface of the conductor due to its high speed (google skin depth in high speed signals for more info). With a heavy silver plate on a copper wire the signal will be carried entirely via the silver layer, without the additional expense of a full solid core wire of silver.

I also saw that you use Roon, right? Roon can convert to DSD 256 and does a decent job of doing so, I expect as well as any streamer's built in oversampling. If you do find that DSD via Roon oversampling sounds best in your system, at that point you might want to consider HQPlayer, as HQPlayer's oversampling to DSD is advanced well beyond anything else-the HQPlayer modulators are really special. I am convert to using HQPlayer oversampling to DSD, so much so that I built a dedicated DAC optimized for doing exactly that.

Hi Dale, I am not speaking for John here, but I think it is important to clarify right away a couple of things: Ethernet data jitter has absolutely no correlation to audio jitter which happens in the DAC (not that Ethernet jitter might not, in a very roundabout way, contribute to audio jitter in the DAC): I just want to be sure for everyone reading this that we understand that DAC conversion stage jitter is a very different thing from Ethernet jitter. That is, Ethernet data is not sent as a constant time stream like audio data is from say a USB interface to a DAC conversion stage. Ethernet data is just packets of data, they are not even, necessarily, in order in terms of timing, consider that the data representing the first second of a track might be sent over Ethernet after the data corresponding to the second second of a track (please disregard the units I am using here). So Ethernet timing is not so critical as long as the data gets to where it is going in time (not too late) and for audio data integrity this is never a problem-subtle degradations of sound quality caused by Ethernet transmissions are caused by noise factors, not by jitter per se as it relates to data integrity. My understanding of John's view on Ethernet clock phase noise, is that this phase noise (clock inaccuracy) can create other electrical noise which may make its way to the DAC, causing noise perturbations in the DAC(causing audio jitter, perhaps DAC Vcc fluctuating, and perhaps interference with the analog stage of the DAC as well), and hence degrading sound quality there. As to Ethernet clocking-my (somewhat limited) understanding is that all Ethernet devices work in an asynchronous way-that is they receive data packets into a buffer, and then they re-clock them out of the buffer using their local clock. Ethernet is also a two way system-all devices talk back and forth to each other, making sure all data is accurately received in each device. So this would happen in a switch, in a router, in a WAP. Each Ethernet device does this re-clocking step, all along the way. If we consider streaming music (or video, etc) from Internet sources, of course we can see that every device must re-clock, asynchronously, along the way, or else the accumulated timing errors would become so large that data would never make it intact.

This is accurate, but has nothing to do with the topic at hand. Additionally, if the internal DAC clock is really good, great. But if it is not, a DAC may benefit from an external clock which is way better than the internal one. All things being equal though, indeed, for the audio clock, having it being in the DAC as close to the conversion stage as possible, and the dat re-clocked just before the conversion stage via that clock, will be ideal. Hence why async interfaces (async USB, Ethernet) are preferred (where the DAC clock is the master audio clock), again, all things being equal...

Hi Peter, not sure what you are asking. All my playback is via HQPlayer, and my optical Ethernet runs to the system, to the (fiber input) Signature Rendu, using HQPlayer NAA protocols. Then from the Signature Rendu USB cable goes to the DAC, which uses a fully isolated (galvanically) USB interface, which indeed runs a 24 MHz XO for the USB transmission. Audio clock is single 45.xxx MHz SC Pure, and audio data (i2s) is re-clocked on the DAC side just before going to the DAC. I can run only a single (audio) clock because all audio is re-sampled to DSD 512 in HQPlayer, and the DAC is optimized for DSD 512 playback. As far as Ethernet protocols go, there are a few, right... Ravenna, UPNP, NAA, Roon RAAT, etc... I am certainly very, very happy with NAA myself, and have no "issues" there.

No need. For example, in my system I run a Sonore Signature Rendu with optical Ethernet input. My DAC (DIY DSC-2) has an isolated USB input which is floating: the USB input ground does not connect to the DAC ground or chassis of the DAC at all. So the USB input ground only connects to G at the Signature Rendu. With only a single point of ground reference there is no chance of having a ground loop. This is how an isolated USB input should be implemented. My USB interface is powered from the USB, which is good, as it provides a completely isolated power supply from the DAC, and the Signature Rendu's USB power supply is engineered properly to be ultra low noise and ultra low impedance (it uses two dedicated LT 3045s in parallel). Yes, the source quality will matter in this case, as it always does. The only reason I could think of to run optical USB is if one needed to go a long distance-in my case, the long distance connection travels on fiber Ethernet to my lab room, where the server, switch, router, and opticalModule are located. In the system, the low noise area, there is just the Signature Rendu and its external Well Tempered Master Clock, the DAC, and the amplifier. For sure in saying the problem is solved, I was referring to high end two channel audio systems (no video) using well engineered components. Things like multichannel and Atmos introduce many other variables, and so does video, as it uses different clock rates which alone create additional problems. But for two channel audio, with good engineering, the problem is solved.

A problem which does not exist: Fiber Ethernet is ideal for transferring any data content, whether music signals or not. The audio clock should not be considered a "problem" because audio clocking should only be done in the DAC right before conversion. Fiber Ethernet transmits audio data perfectly, without introducing noise artifacts into the DAC, and has no need for an audio clock. The only potential drawback of using fiber Ethernet to bring audio to the DAC, is that the computing power needed to decode it to digital audio is more than that required by, say, Async USB-so the interface itself is a noise source. But this need not be a problem: good engineering can isolate this noise from the DAC's sensitive clock and analog circuitry. The same is nearly true for Async USB: Again, no audio clock in the signal, audio clocking happens only inside the DAC right before conversion. Async USB requires the same isolation from the receiver circuitry to the DAC circuit, but this is well understood now and is easily solved. Of course, by using wires, noise can be introduced into the DAC from the USB connection-again this can be solved via 2 things: a well engineered USB source designed specifically to provide a low noise USB feed, combined with a DAC which properly isolates and re-clocks the digital audio data stream from the USB receiver section. Async USB has the further limitation of only being good for short connections. There is no need for a "new" way to do it, we have "perfect" ways already. What matters is good engineering of the ways we have to eliminate the propagation of noise to the clock, DAC Vc, and analog circuits of the DAC. For example, an Ethernet DAC using optical Ethernet input: The optical Ethernet receiver circuitry needs its own independent power supply, this could be a well engineered low noise SMPS, or if you insist, it could be a well engineered linear power supply with a dedicated transformer. The Ethernet receiver circuitry would then be as well, shielded from the rest of the DAC with a faraday cage, just as the power supplies (generally) should be shielded from the sensitive DAC circuitry. DAC clock is sent to the Ethernet receiver through isolation, and the i2s output comes out of the Ethernet receiver through isolation. The i2s data is then re-clocked by the DAC clock, directly before conversion. This is pretty much the same way well engineered async USB works, albeit without the advantage of fiber isolation from the source. This is a solved "problem" and does not require the development of some new way of doing things. AES and SPDIF are technically flawed because they attempt to provide the clock from the source. HDMI is total mess, and requires outstanding engineering solutions because it attempts to transfer so many differing signals on a single cable.

Exactly, HQPlayer support is a really good thing! And, HQPlayer has taken another step forward with v5, and the client app for iPad. Sonore optical streamer with HQPlayer works fantastic.

Yours is almost built, just waiting on our capacitor order to arrive from Germany, FedEx mis-sorted the package...If all goes well, it will ship on Monday before I take a little break here.

Update: Sonore VP has tried the Synergistic Master Fuse, and reports it does result in an improvement. As always, I am hesitant to recommend it due to the very high price-but if everything else in your system is "perfected", perhaps a Master fuse is what you need to gain that last little bit of sonic bliss. Sonore will offer the Master Fuse as a special order option for the Signature. Please note that it is unlikely that we would stock these fuses due to the cost, but we will special order them for new Signature builds if requested.

1. If one has a poorly designed DAC, with a poorly designed analog stage and output filter, it is possible that a preamp with a good RF filter on its input could improve sound quality. Of course, the better solution would be to not purchase the poorly designed DAC! From what I have heard with the Tambaqui running amp direct, and from what I know of Mr. Putzeys' engineering chops, I would not expect the Tambaqui to have this kind of a problem. 2. Digital attenuators will reduce low level resolution at very low volumes, but consider: the volume level has to be really low for this to happen, unless the gain matching of your amp to source is way off, this should not ever present a problem. Also consider-at these very low listening levels, the very low level information lost is not anywhere near audible anyway. I hope that makes sense... What I am saying is, that what you lose at high levels of attenuation is low level details which you cannot hear at very low playback levels anyway (the low level signals have dropped well below the noise floor). Also consider, that analog volume controls have their own problems, especially at low levels-they are not perfect, the resistors add distortion and noise (thermal noise), and they add more distortion and noise the more one attenuates the signal. Analog volume controls just add a bunch of resistance to the signal path-this is not ideal! Digital volume controls, when well implemented are superior, as long as the system gain matching is reasonable-and by reasonable I mean you are not listening at -100 dB on your volume control, or something crazy like that. Also consider, if you like the sound of the Tambaqui, a bunch of digital processing of the signal does not bother you-the Tambaqui is always doing a bunch of DSP on the signal with its asynchronous SRC, etc. The digital volume control is just part of all that processing, singling out volume control as somehow "evil" is nonsense-like anything what matters is how well that processing is done, again I defer to Mr. Putzeys' expertise here. Let's also consider that nearly all contemporary recordings, and re-masters, etc are mixed in the digital domain, using digital level controls, and have been so for many, many years now.

1. "Galvanic Isolation" does not mean that all high frequency noise cannot pass the isolation-this is important to understand. The same is true for "isolated" USB inputs. While I am in favor of designers using isolated inputs, having them does not eliminate noise, it reduces noise. We are talking about high frequency noise here, which is high enough in frequency to be broadcast into the air as RF. So, internally, it can be floating around inside the DAC, and be picked up by ground planes, analog output stages, etc. Additionally, some noise will pass across a "moat" due to capacitive coupling, or through isolation transformers (in the case of the Ethernet jack) via the same capacitive coupling. So, "isolation" helps, but does fully address the potential problem. I have never experienced an "isolated" DAC input which did not perform better when fed via a well designed low noise source, be it SPDIF/AES (usually transformer isolated), USB (usually isolated by various chip isolators), or Ethernet (usually isolated by tiny transformers in the Jack, and/or a transformer based chip). This is why optical fiber Ethernet can be a big benefit, despite the "isolation" incorporated in a given component. 2. Given the robustness of the Tambaqui's output stage, I very much doubt that there would be any benefit to adding a preamp if you have no need for one (unless you want to add something to the sound, like tube color or something like that). While some DACs do benefit from having a preamp, those DACs which benefit do so usually because their output stage is not "beefy" enough in terms of current headroom- I have tested this pretty extensively. Remember, a distributor, or dealer, is always going to be happy to sell you another component. 3. The Makua is a very fine preamp/DAC, and if you need the input switching, and/or have analog sources, or want to add a turntable set up (with the very fine phono stage onboard the Makua) go for it. But If you just have the single DAC source, I would not bother. There is no significant difference in terms of the output stage of the Makua vs. the output stage of the Tambaqui. If you are still curious... be sure to do a precise level matched A/B comparison in your system-it must be precisely level matched, so make measurements to be sure. First listen sighted and determine if you hear differences, and learn what you think those differences are, and then do a blind test and make sure you can pick out the winner at least 75% right, to be sure there is a real valid difference.

I fully agree with your skepticism! I shared this point of view entirely, as I was very leary of how could, from a technical perspective, the close in phase noise of the Ethernet clock, possibly matter for sound quality. Looking at it technically, it was very hard for me to come to any kind of conclusion as to how these could matter. John Swenson has proposed possible technical reasons for why this is so, but I am still not sure if his proposed theories are correct (or if I even fully understand them!). In any case, I did experiments myself, including adding an ultra low phase noise, Ethernet clock to my streamer-this clock was about as close to SOTA as one can get. My expectation was that there would be no difference, but I went to the effort because so many audiophiles had reported significant improvements via improved Ethernet clocking, and because I am capable of testing this for myself, I did so. The external clock I built and added ("Well Tempered Master Clock") has about 3x lower for close in phase noise than the, pretty good, Crystek 575 in my streamer. The improvement was not subtle in an audiophile sense, and was immediately apparent, to my astonishment. I still have no easily understandable, reliable, technical explanation for this, but the results of my own testing are irrefutable, so I have to accept them. As such, I encourage those looking for better sound to try for themselves Ethernet distribution components which use better clocks, or allow for the application of better external clocking.

Consider using purpose designed (for audio) components to provide your optical isolation, such as the upTone EtherRegen and the Sonore opticalModule Deluxe. Regular consumer FMCs are built to the lowest possible price point and have a lot of comprises in their design when used for audio purposes. Suprisingly, for example, the quality of the clocks in the converters matter for sound quality, as well as the noise performance of the onboard voltage regulators, and the choice of Ethernet integrated circuits. I would expect a well sorted optical Ethernet isolation scheme with high end components and really good power supplies would bring the performance of the Ethernet input of the Tambaqui to another level.

We here at Sonore have not tried the new Master Fuse, given the expense, it is hard to consider it! But if anyone does try on, please report your results here, as we love to know how much better this new fuse might be.

The mentioned thread does not appear to state what the Network situation was... The configuration of the Network is going to have an influence on the sound of using the Tambaqui Ethernet input. If one just connects the Ethernet input to a busy home Network, with no attempt to reduce the Network noise, then the results will not be what they could be. On the other hand, if one uses an approach like running a pair of Sonore opticalModules (one at the upstream switch, one close to the Tambaqui) with a run of optical cable to the audio system, with good power supplies on the opticalModules and a good, short, Ethernet cable to the Tambaqui (short to avoid any noise pick up there), one will have different results. Likely even, probably, a little bit better would be an opticalModule upstream, and an etherRegen downstream-providing a degree more isolation. By running optical fiber to the audio system one avoids most of the Network noise getting to the Tamabaqui, and by using quality, designed for audio (like the Sonore and UpTone products mentioned), optical converters, one also adds in better clock quality for the Ethernet clocks, which, surprisingly, also improves the sound. Just as the quality of the USB source matters (from a noise perspective), so does the quality of the Ethernet source.

Yeah, I think would be an ideal way to use the Tambaqui's Ethernet input, as it would provide a very high degree of isolation from any Network noise.

Honestly, considering that the Tambaqui re-clocks and asynchronously re-samples all incoming rates to its own internal masterclock, I would doubt there could be any advantage to using AES/EBU over the Ethernet or USB inputs, as long as the Ethernet and/or USB forces were well sorted designs with low noise outputs. Using an EtherRegen for Ethernet, via optical fiber and a short Ethernet cable to the DAC should be an ideal way to use the Ethernet input, and using a Sonore opticalRendu or Signature Rendu optical would be good ways to get the best performance out of the Ethernet and USB inputs, as both these sources, connected to the home Network via optical fiber, offer extremely high quality and low noise output.

Unfortunately, this device is not a Roon endpoint at this time.

The Sonore endPoint has an AES/EBU output which performs very well, with complete galvanic isolation and very low phase noise clocking if one is looking for a relatively affordable AES/EBU streamer. See some details here: https://www.sonore.us/endPoint.html We do not make a lot of these, as most filks prefer to use our USB output Renderers, but for those who prefer an AES/EBU output this is an option.

What do you really want to know? Are you curious of your server's entire power consumption? If so a simple way to figure this out os to use a watt meter inline with the AC supply from the wall to your server's power supplies. These meters are easily available from Amazon, etc and are not expensive.

One consideration of the 1612 vs 1656 is that the 1656 has higher output current availability. Theoretically, the 1612 has "enough" output current to properly drive the Purifi modules from a technical perspective. IMO though, one often will experience better sound quality when a driving circuit has a higher amount of current headroom available-I have experienced this a few times with my DIY DAC builds, and I also preferred the 1656 (to the 1612) driving the Purifi modules in my own DIY Purifi amplifier. The 1612 has slightly better specs when it coms to distortion than the 1656, but if one looks closely, both of these ICs have distortion specs which are comfortably below levels anyone would consider audible, so there are no worries there... While technically the 1612 should be powerful enough to drive the Purifi modules, it does appear to be pushing pretty close to its maximum levels for both output voltage (considering +/- 12 VDC rails) and current when driving the Purifi modules at high levels, perhaps the greater headroom of the 1656 is responsible for the (subjectively) better sound quality.