To switch or not to switch? Melco S-100 or Innuos Phoenix NET switch?
- Thread starter shakti
- Start date
Hi there,
Its worth noting that any component with Mundorf capacitors, which includes many of Sean Jacob’s power supply designs seem to be subject to long burn in times. Unlike many power supplies, its really not sufficient to just switch on the PS and leave it in a steady state…..the caps need to run dynamically with many charge/discharge cycles and preferably a few switch offs and cool downs. The only time those caps are truly running in is when they are properly working, not just warm with a stable current flow. My current set-up is an AQVox SE with SJDC4 hybrid LPS , a Melco S100/PF Ultra OCXO with SJ dual rail DC4 and an Innuos PhoenixNET, into an Innuos Statement. My ethernet bridge, router and modem also use SJ DC4 hybrids. When I auditioned the PhoenixNET i was lucky to obtain a fully run in unit for a couple of weeks trial. I tried both Melco and PN alone, with the AQVox SE and I tried changing the order….AQVox, Melco, PN, or AQVox, PN, Melco .
Taking out the AQVox altogether wasn‘t as good as when it was in the chain. Both the Melco and the PN preferred receiving the cleaner, retimed output of the AQVox. The Melco/PF alone delivered outstanding PR&T, fantastic purity and a fully formed 3 dimensional soundstage and was brighter (in a good way) and more radiant than the INNUOS akin to bright sunshine on a crisp winter’s day. The Innuos alone really wasn’t that far behind but it tended to be a little softer, a little more gentle but somewhat more natural and organic, with greater texture, like the light quality on a sunny, late afternoon Autumn day, where colours are deep and intense But it was when the two were together that the true beauty emerged. The sound took on a wholeness and a wholesomeness, of real musicians playing instruments in a 3 dimensional venue. It didnt really seem to matter whether the Innuos or the Melco came last (I only need to swap a couple of 1m ethernet cables to change the order) as in either order I could hear both the Melco’s amazing timing, clarity, energy, purity and drive and the Innuos‘s superb naturalness, textures, colours and musical beauty.
But here’s the thing. After the evaluation I ordered an Innuos and it arrived within 2 weeks. I thought it sounded great from the get go but soon a few anomalies started appearing and every listening session I kept reminding myself that what I was hearing was probably due to running in. The Innuos takes about 5 days to warm up, fully settle in and sound its best, but this is nothing to do with running in….its like that every time you switch off and let it get cold. Running in takes far longer….you see from my system description that I have a lot of Mundorf capacitors so I’m pretty used to following their run in patterns and they take ages…..and its a bit of a hurdygurdy ride. Great one day, odd the next few days…you’ll get short periods of sibilance, short periods of bass emphasis, and occasional emphasis of detail, followed by commensurate improvements in those areas…..always the down before the up…,the bigger the down, the bigger the up . I thought the unit was run in at least 5 or 6 times, sounding really good for quite a few days before another off day, followed by a new pinnacle of SQ. But you know when the unit finally reaches stability, because the sound will be sending shivers down your spine and bringing tears of joy…..there’s almost a transformation where the whole suddenly feels like its directly connected to your musical soul. Flowery language I know but its the best way to describe how this combination makes you feel. This level of performance is not really hinted at during the running in period. There’s no doubt that the unit sounds really good from the get go, but the true performance takes quite some time to emerge.
Its worth noting that any component with Mundorf capacitors, which includes many of Sean Jacob’s power supply designs seem to be subject to long burn in times. Unlike many power supplies, its really not sufficient to just switch on the PS and leave it in a steady state…..the caps need to run dynamically with many charge/discharge cycles and preferably a few switch offs and cool downs. The only time those caps are truly running in is when they are properly working, not just warm with a stable current flow. My current set-up is an AQVox SE with SJDC4 hybrid LPS , a Melco S100/PF Ultra OCXO with SJ dual rail DC4 and an Innuos PhoenixNET, into an Innuos Statement. My ethernet bridge, router and modem also use SJ DC4 hybrids. When I auditioned the PhoenixNET i was lucky to obtain a fully run in unit for a couple of weeks trial. I tried both Melco and PN alone, with the AQVox SE and I tried changing the order….AQVox, Melco, PN, or AQVox, PN, Melco .
Taking out the AQVox altogether wasn‘t as good as when it was in the chain. Both the Melco and the PN preferred receiving the cleaner, retimed output of the AQVox. The Melco/PF alone delivered outstanding PR&T, fantastic purity and a fully formed 3 dimensional soundstage and was brighter (in a good way) and more radiant than the INNUOS akin to bright sunshine on a crisp winter’s day. The Innuos alone really wasn’t that far behind but it tended to be a little softer, a little more gentle but somewhat more natural and organic, with greater texture, like the light quality on a sunny, late afternoon Autumn day, where colours are deep and intense But it was when the two were together that the true beauty emerged. The sound took on a wholeness and a wholesomeness, of real musicians playing instruments in a 3 dimensional venue. It didnt really seem to matter whether the Innuos or the Melco came last (I only need to swap a couple of 1m ethernet cables to change the order) as in either order I could hear both the Melco’s amazing timing, clarity, energy, purity and drive and the Innuos‘s superb naturalness, textures, colours and musical beauty.
But here’s the thing. After the evaluation I ordered an Innuos and it arrived within 2 weeks. I thought it sounded great from the get go but soon a few anomalies started appearing and every listening session I kept reminding myself that what I was hearing was probably due to running in. The Innuos takes about 5 days to warm up, fully settle in and sound its best, but this is nothing to do with running in….its like that every time you switch off and let it get cold. Running in takes far longer….you see from my system description that I have a lot of Mundorf capacitors so I’m pretty used to following their run in patterns and they take ages…..and its a bit of a hurdygurdy ride. Great one day, odd the next few days…you’ll get short periods of sibilance, short periods of bass emphasis, and occasional emphasis of detail, followed by commensurate improvements in those areas…..always the down before the up…,the bigger the down, the bigger the up . I thought the unit was run in at least 5 or 6 times, sounding really good for quite a few days before another off day, followed by a new pinnacle of SQ. But you know when the unit finally reaches stability, because the sound will be sending shivers down your spine and bringing tears of joy…..there’s almost a transformation where the whole suddenly feels like its directly connected to your musical soul. Flowery language I know but its the best way to describe how this combination makes you feel. This level of performance is not really hinted at during the running in period. There’s no doubt that the unit sounds really good from the get go, but the true performance takes quite some time to emerge.
If you look into that a bit more closely it reveals just what good value the PhoenixNET represents. (I’m a user with nothing to do with Innuos btw)
The unit includes 2 Statement-class linear power supplies designed by Sean Jacobs. This brings some really innovative engineering in terms of how the regulators are designed and connected directly to their consumer clients….with no DC cables required. Super clean, stable DC right where its needed. The linear power supply regulator is connected directly to the network switch chip, so as stated, no cable and no switching power regulators….just clean uncontaminated DC.
The other LPS is tied to the oscillator, a 3ppb OCXO, which itself is tied to the network switch chip, so again no clock cables or loss of precision
If you look at the switch itself, you find high quality RJ45 amphenol sockets isolated with large, high quality transformers and directly soldered to the main board and no LEDs Targeted EMI shielding is provided at emission points and the unit sits on anti-resonance footers specially made and tuned for the unit.
Considering all the above, I don‘t really see anywhere in particular where the Melco/PF/DC7 combination is superior. You could argue that the Innuos uses a 100Mbps network switch chip, but given that it was selected for its much lower noise floor vs a gigabit chip its not straightforward to assume that the faster chip is superior when judged on the basis of SQ. So, different architecture sure but the PhoenixNET is designed from the ground up for audio and the power supply is custom built with a lot of genuine refinements in terms of its ‘intimate’ relationship with the switch section. In terms of performance I find the units different but if I had to pick one or the other I‘d find it very difficult, given that both are exceptional performers and their differences are simply different combinations of positive attributes. The happy thing is that when used together those differences are preserved in the final sound and meld together beautifully, really complementing one another.
Of course there is one major difference and that is in my system all power supplies are by SJ, whereas you have the SJ and PH combination. As all switches do is essentially modulate the PS’s DC into a digital stream, the power supplies do contribute a large amount of their own identity. Also when different power supplies are used together, you get different noise spectra which can sometimes intermodulate to form a more complex spectrum of noise and harmonics.
I actually have a 3 PS combination: Sean Jacobs (Innuos USB Reclocking) PH for Melco S-100 and Farad Super 3 for ISP Modem Router, but all three are isolated by the CCI modules of my Shunyata Everest so I don't think that such a combination influences the sound. If today I had to choose between Melco S-100 PF and Innuos Phoenix Next and considered the Price-SQ question, I would think about it carefully, but if I only considered SQ, and just what the Innuos sounded like with a month, then my choice would be Melco.
Melco S100 audiophile switch
https://www.underwoodhifi.com/ is a Melco dealer
www.whatsbestforum.com
I demoed an Innuos Statement last year and used its Direct LAN connection into my previous Linn KDSM, I've since upgraded to the new Organik KDSM. As you are probably aware, the PhoenixNET is essentially this part of the Statement.
I had the stock Melco S100 at the time and, frankly, the Statement didn't make the improvement in my system using its LAN output that the stock S100 did.
The gap was even bigger when I used a Farad Super3 on the Melco vs the Statement, so I didn't buy the Statement and went for the PF upgrade instead.
Also, please be aware that not all streamers like the slower 100Mbps connection - my previous Linn KDSM had no problem with the slower speeds from either the Statement or the S100, but the new model wouldn't play music constantly using this connection, I got dropouts. Fortunately the Melco gives you a choice of connection speeds.
The long and short of it is, unless you've heard both in your system, you can't determine which sounds better.
You’re absolutely correct when you say that its absolutely essential to ensure that you have a good match between components, indeed you’re looking for synergy, beyond just compatibility.
I‘ve really no experience with Ethernet out of the Statement as I use USB to my Devialet DAC, which works great with never any dropout and frankly incredible fidelity. I tried USB vs Ethernet into the Devialet at the very beginning and I preferred USB, being the more expressive of the two, but that’s just as likely a function of the Devialet USB input. My Melco with PF Ultra, powered by a dual rail DC4 feeds its output into an Innuos PhoenixNET. The Melco is fed from the output of an AQVox SE, so the Statement is fed by cascading switches, all working at 100Mbps, which seems to be more than enough bandwidth for the music. I tried both the PhoenixNET and the Melco in isolation and I found them about equally capable, but different and in terms of pure musicality, the pair together were significantly better than either alone.
The one thing I was not quite clear about is your comparison of a server and a network switch, as functionally they are quite different and I’ve only ever encountered them working together rather than in an either/or situation. I’m wholly unfamiliar with Linn products so perhaps Im missing something?
Last edited:
As I understand it, using Direct LAN out from the Statement is basically using the Statement as a switch - maybe an upgraded switch, but still a switch.
I was using the Statement for playback versus the Small Green Computer that I had at the time, and, as far as I am aware, unless you are using a player on your server and are connected to your DAC with USB or with another digital output, a Statement, Antipodes K50, or any PC/server with a LAN card that has an output to a streamer is just acting as a server when you use the LAN outputs directly into a streamer.
Both my SGC and the Statement were running Roon.
In order to try determining what either device (S100 or Statement) was doing to sound quality, I tried the Statement without the S100 in the chain and with the S100 in the mix. SQ improved with the S100 in the chain, but unless I had the S100 in the chain, there was no improvement in SQ running the Statement directly into the Linn. Again, this was with a LAN connection only, I did not try USB, so YMMV.
I just picked up a used K50 for a server with my new Linn KDSM and I do hear an improvement using its LAN direct output connected to the Linn and I didn't hear an improvement when I demoed one about a year ago with my old KDSM.
The new KDSM is a substantial upgrade over the model it replaced and, among other improvements, the noise floor has been lowered significantly which is probably why I'm hearing a difference. Is it a large improvement? Not night & day, but it does seem to be improving on top of what the S100 is already doing. Again, using the K50 this way, it is acting as both a switch & server and the Linn is the Roon endpoint.
I am also trying using the K50 as an endpoint into the Linn using SPDIF (just trying this for the past couple of days). Antipodes is very proud of their clock which works through the traditional digital outputs. Using it this way, I can use Roon as a server and Squeezelite or HQPlayer as the player. I'm breaking in a Pink Faun BNC cable and the K50's clock (long story, but the previous owner never used these connections and when I tried, it didn't work - bad internal connection - and I just got it back from Antipodes service which is great btw).
I've only been running the cable & player in since Thursday, but it sounded pretty damn good yesterday. I have a couple of weeks of more burn-in (cable needs at least 300 hours) before I can compare Direct LAN into the Linn vs SPDIF.
Are you sure about this? My understanding is that it's being used as a renderer, but I could be wrong as it does in fact get confusing. How do you like the k50 compared to the statement?
I am also wrestling with the decision on getting a switch. I want the Melco with PF upgrades just waiting to hear back from Jorg. Of course I hate the price of it, exploring the price on upgrading just a plain old Buffalo with PF clock.
Last edited:
My Linn streamer is a renderer and my K50 is running Roon Core, so it is a server. Here's some info from Antipodes about how their system architecture works https://antipodes.audio/solution-architecture/
So far so good on the K50.
So far so good on the K50.
Congratulation and looking forward to your feedback.
Any information from where you bought it.
HNE Labs, small shop young engineer/designer in Vietnam. I engaged him when I saw the design, he was friendly helpful (English was good enough) and answered all my technical questions. Build quality looks good, all Mundorf caps, nice toroid, and the PS is based on the DC4 form Sean Jacobs, quality clocks used (2) one for Switch other for the fiber converter (built in) which is an option I like since I want to run fiber and try it out. He's on FB, doesn't have an online presence like website etc. He does alot of business in Europe and Asian markets.
Last edited:
Mr. Blackmorec, may I ask: did you consider the Synergistic Research switch before settling on the Phoenix to feed into your Statement? I also have a Statement and am trying to determine my best option to upgrade the quality of the digital stream into the system. Thank you.
Hi CDC77,
No I didn‘t consider the Synergistic Research for a couple of reasons…the major one being it wasn’t available at the time I chose my switches. The other reason would have been to do with power supplies and network architecture. I use Sean Jacobs DC4 regulated supplies with Mundorf Silver/Gold DC cabling throughout my entire network chain and I wouldn‘t want to deviate from that as power supply noise can interact to give more complex spectra including harmonics. The third reason is I have a particular architecture that I follow for my network in that every stage of the network sees an improvement in power supply noise spec, clock specification for jitter and phase noise, vibration control and cable screening. It was easy to maintain this architecture with the Melco and Innuos switches and that architecture is extremely important to the final sound in that it generates compounding benefits that improves the quality of the data stream as it travels over the network.
Similar threads
