Yes makes sense. I never needed to do anything for the netgear switch that was in place so same principle I think.
1 Like
The S100 switch was tried by 3 members here, with a 500 series system. Was found a bit dull and uninvolving.
Etheregen and PhoenixNet were preferred.
3 Likes
Forgot about that, good point. The is a Melco switch for sale on ebay that has been modded to include a clock and upgraded wiring. Sounds interesting, but for the money safer to go with the PhoenixNet I am guessing.
4 Likes
I was fully intending to compare sfp cable with WiFi and Ethernet once the ngkdsm arrives.
Which did you settle on in the end? I guess sfp+ runs the risk of decoupling but adding back in some noise. As does activating the wifi modules inside the dsm. I am hoping I will have done enough to manage things so the noise level is too low to register any significant distortion or at least levels where I can’t hear a difference.
I have the option of sfp+ cable direct from my ubiquiti 8 port pro switch. Already have cable to test it out as well.
3 Likes
That’s a good starting point. Note, the KDSM supports SFP (1G) rather than SFP+ so if your Ubiquiti will support SFP then you are good to go.
1 Like
I will check that.
Exit. My understanding is that sfp is a slower version of sfp+ otherwise the connectors are identical. If I get sfp cable not sfp+ cable I will be fine (not the other way round) but I will double check with ubiquiti. Thanks.
Looks like its a parameter change in the unifi app for the port - auto negotiation, 1gbs or 10gbs FYI.
1 Like
I tried an S100 recently. I wouldn’t say that I found it uninvolving - more that I couldn’t perceive any difference from my existing Ubiquiti network. I tried running it with the standard (switched) power supply, and fed from a Plixir LPS. I also tried aDot fibre in the path, again with no perceivable difference in sound.
Which I have to say pretty much aligns with my expectations (as a network engineer in a previous life).
Of course, if you’ve got a fibre port on your source, it might be beneficial
2 Likes
Thanks. Ultimately thats my goal - to set up the ubiquiti gear so that the last stage into my kds is as close to noise free as possible.
Now I better understand the decoupling principles described by members above, I would hope for a similar result to yourself. I will aim to test the melco switch once my ngkdsm is in place and stabilised. Currently I am dedicating two switches to the hifi - the 8 port ubiquiti switch and a ubiquiti flex mini which is poe so has no power supply.
I did the am radio test earlier in the week which was very helpful to better understand things. I am getting almost no rf noise that the radio is picking up from my ubiquiti flex mini. I also discovered a lot of rf noise coming from the floor - power cables running under the floorboards. So I moved my last leg utp ethernet cable off the floor and that helped some more. I will experiment with some cable trunking in the next week or two.
I have some shorter lengths of quality utp cable arriving to test as well as a 2m run of sfp twinax cable. Also 4 ferrite core donuts to play with. Just for fun. I should be well positioned to experiment with different options for the last leg. Will report all my findings. Thanks for your thoughts.
I’ve put a PhoenixNet switch in on my last leg to the ND555, but to be honest it was mostly to scratch an itch that had been nagging me for a while.
I run multiple (3) Unifi switches to get connectivity to all the places in the house, the garage and the workshop that I want wired devices. The one in the lounge is fed via PoE from my central distribution switch (which is next to the fibre feed into the house). I do wonder if the 48V PoE, fed from a switch a good 30m distant, has been a component in not having much in the way of network noise.
Given the good WiFi coverage, I should really try that to compare one day…
1 Like
Likewise. I feed a cinema room, my study and the hifi room with ethernet. I have most devices on wifi anyway now enabled by 3 unifi APs. Did the pheonixnet make an audible difference or did you not bother to test?
My view is to do the minimum testing to confirm things are distortion free and there is no major difference in sound between setups and then get on with enjoying listening to music!
If you can set things up so the access points are wired backhaul you’ll see notable improvements in WiFi speed. I have a second router in my kitchen connected to the main one via backhaul. I have a 1gb connection and get anything from 800-940 Mbps on WiFi this way, but if I switch it to a wireless access point it consumes much of the bandwidth and my speeds drop to 200-300. It’s really quite unbelievable the difference it makes.
Yep all 3 ubiquiti APs are wired back to my switches.
2 Likes
With the 2 Melcos I’ve had before I tested the main claim that the manufacturer makes and aims to achieve - that it behaves just like a hifi component and not like a computer (if you read up on their vision for these products). To me the Melcos sounded excellent, not only as a NAS but as streamers too, and especially with some better cabling to suppress noise further.
However, at the point I gave up, this was not for SQ but because for me Melco failed on the main one and always felt like a computer environment, annoying user interface, complicated updates, bugs and too much fiddling and restarts almost reminding me of Windows… has this changed? Or is it still just a good sounding NAS?
1 Like
Nice. It’s definitely the way to do it. I’ve been so impressed by the WiFi bandwidth I get, I use the antenna on my NDX2 rather than bother with expensive switches and cables. Sounds pretty good to me.
1 Like
I love the ubiquiti gear. It was installed a few months before covid hit and it was a godsend. The network and wifi setup with 4 people in the house was rock solid.
1 Like
It’s good when setup properly, a nightmare when not! I invested money in my expensive Asus router but still got a dead zone in my wife’s office. Second router in the kitchen linked via backhaul and she now gets up to 800Mbps!
1 Like
‘So one question remains. How about a much better clock than those present in common consumer switches ?’ @frenchrooster
Could we pick up this discussion here? Remember I use a Linn DS streamer. It has a buffer, cpu and clock inside it. The linn ds grabs the data from my media server (roon or asset) asynchronously using its own clock and its buffer and feeds the audio data synchronously into its DAC. The cpu inside the DS deals with the jitter. Not my media server or network.
So I am struggling to understand how a better clock in my network switch will improve my sound? Am I missing something obvious?
1 Like
I can’t say, but I know that the best audiophile switches has ocxo clocks and better power supplies.
From this forum or elsewhere, many guys with top digital sources ( dcs, linn KDSM, Naim Nd555/ twin ps have expensive clocks on their switch.
Pink Faun for instance modifies the Melco S100 switch by upgrading the clock.
@MidnightRambler here, with Nd555 , full 500 dr, has a Mutec clock connected to his 2X Etheregen switches.
Another one has a top Korean clock on his switch.
All say it improves very nicely the sound.
But why, I don’t know.
3 Likes