Hence the physical separation in the NDS/ND555 and the galvanic isolation around the networking components.
Not the same - the “Ethernet” is a complex set of standards evolving over many years, originally designed for packet-based communication between computer servers.
The TCP/IP protocol was designed in the 70’s, but ‘Ethernet over twisted pair’ or 10BaseT at 10Mbit/s was from 1984
100BaseTX (Fast Ethernet) provides 100Mbit/s up to 100m, specified in 1995
1000BaseTX provides 1000Mbits up to 100m was specified in 1999
2.5GBaseT and 5GBaseT specification in was 2016.
Then there are similar specifications for Fibre-based deployments
S/PDIF was set by Sony (S) and Philips (P) as part of the 1983 Redbook Standard, and designed to carry 20-bit signals (so perfect for 16/44.1, which was selected as the lowest common factor between the manufacturers).
24-bit signals can be carried by way of 4 extra bits that were just included at the time, but it was not intended for 24-bits.
Limitations are that the receiver does not control the data rate, so it must avoid bit slip by synchronizing with the source clock. However, clock recovery during this process may produce jitter - plus faster the stream, and this process becomes harder to implement.
Ok, well i don’t think there are many people in the audiophile world who would argue that a DAC + streamer combo produces better results than a physically separated DAC and streamer. But if that is your preferred approach then ofcourse you are completely free to follow that.
I think you are wrong. It’s not what I can read elsewhere, for instance in WBF. The best digital streaming sources seem to incorporate the streaming board inside the dac. Because usb is compromised, as spdif.
Ethernet only connection seems to be more and more chosen as the best solution.
Yes, I ran that type of system for many years initially a Meridian CD transport via S/PDIF into a DAC, a dual differential DAC7 with 24-bit filter, then with a 2nd NIC card into the DAC from a Squeezebox 3 with output upto 24/48. I did try a SBT, and 24/96 wasn’t possible and the SQ wasn’t as good as the SB3 on digitial out.
All replaced with a ND5XS with different UPnP servers.
Which all catered for the upgrade through NP5XS Power to NDS/555DR both with Asset UPnP server and Roon & the SonoreUPnP Bridge.
The isolation provided in the NDS/ND555 between networking components and DAC board and then Analogue Output board, all running on separate power feeds is as good as separate chassis arrangement.
I have listened to many systems with separate DACs, also fully active digital systems e.g. Meridian 861, 8000SP.
Ok fair enough! I guess there are different schools of thought on this.
This topic has been discussed many times in the past ofcourse. Here is an older post from @Simon-in-Suffolk describing the benefits of separating the streamer and the DAC:
Or as @jmtennapel describes, two devices physically separated within a single box. Which is also possible ofcourse, as long as the shielding is good enough.
Naim has of course always preferred two-box solutions to separate the PS and not the DAC/transport instead, while others preferred the latter. This was even remarked on by What Hifi in the CDS3 review 17 years ago:
This two-box CD player isn’t split into the conventional transport and digital-to-analogue converter configuration of most rivals. Naim doesn’t design players in this way: it feels the timing corruption suffered by the digital signal is too great to produce the best possible results. Naim’s solution is to put the transport mechanism and DAC in one box, the £4875 CDS3, and the power supply section, the £2369 XPS2, in another.
Not quite, there is no timing corruption with regard to sample timing… and remember this is transport timing… nothing to do with sample timing… so we are talking intermodulation coupling, some seem to get these two things confused with each other.
Naim go to huge extents both internally and externally to minimise actual sample timing modulation … it’s mostly the same method for internal and external signals to create the clocked signal.
From a marketing perspective I once asked Naim the question (around the time of the launch of the ND555) , and the reply was something along the lines of the cost to make a streamer digital only output was pretty close to the cost of making a streamer with digital and analogue signal outputs, so it was more cost effective to give customer choice from a single unit.
I am not saying that WHF’s interpretation is necessarily correct, I am just saying that Naim not wanting to separate DAC and transport is not exactly a surprise in 2021.
I suppose the ND555 whitepaper that I linked earlier in the thread has much to say about the separation topic, but it’s been a while since I read it
If you see DCS, Soulution, CH precision, and MSB, they all choosed to integrate the streaming board inside the dac.
But they go with outboard power supplies.
For XLR, they all go for it. Naim doesn’t, but a lot criticise that lack.
See my post above … it’s economics.
If we ever see a Statement streamer, I’d be surprised if the transport and DAC was not physically decoupled in some way.
Naim have a great presentation about their DNA being based on separation and decoupling… and in their streamers they go to huge extents, especially in the ND555 to decouple as much as they can within the same physical enclosure between transport and DAC.
Not sure what this has to do with what I said, which was what i wrote above. But anyway …
Like literally everything in the world of commerce. It is always trade-offs to meet a price point, but it is still not a surprise that for 2 decades and up to the current ND5555, Naim thinks that overall it is better this way, within the constraints.
Sure the point I was challenging was the statement of so called ‘timing corruption’ say from an external digital source… it simply doesn’t make sense…
I suspect a ND555 feeding a separate ND555 in DAC only mode will be pretty impressive… although very much diminishing returns.
yeah, ok, I am not disagreeing, it’s What Hifi after all, and it’s not as if Naim were generally a stranger to mystical claims. Really, my intent was only to point out how long it has been that way, in the context of the earlier discussion
Possibly but remember the relevance with these products is that they use very different reconstruction, oversampling and conversion methods, so they create their analogue signal in quite a different way from each other. Digital to analogue is ultimately a compromise, so you choose the method that feels and sounds most comfortable to you.
Not sure I follow… FPGA is typically a programmable chip device … in the DAC world FPGAs are often used for high speed DSP functions… Naim however use an Analog Devices SHARC processor for their DSP functions…
Willl Naim carry on using AD SHARC processors in their future products?.. I see no reason to change unless they you use their own reconstruction and filtering algorithms… I doubt Naim will do this as that requires a level of specialism that I suspect Naim don’t have.
Oh what i meant is that since the Chord DACs (and i think the new Linn DACs too?) only use FPGA’s and no off the shelf PWM DAC chips like the Burr-Brown’s, perhaps this would be an interesting option for Naim too to retain their own sound and have a lot of flexibility in controlling it…
Ok, but Naim don’t use PWM (Delta Sigma) type DAC chips or at least not in DS mode, they use multi bit converters … and the FPGAs are used for the reconstruction filters and noise shapers with some other manufacturers such as Chord, not the actual converters…
So perhaps your question is will Naim continue to use off the shelf converters as they do now with the Texas Instruments devices such as the PCM1704K, or will they use custom converters… I suspect Naim will continue to use standard off the shelf converters.
