Ethernet cable installation...

Huge posted:

This has got perverse!

whats the converse to that...?

sorry, i reverse what i said earlier

enjoy

ken

ken c posted:

Thanks Graham -- i can confirm that the switch is not on the dedicated hifi radial circuit.

i can see you have recovered from Salisbury visit jet-lag very quickly indeed... 

enjoy

ken

In that case my *guess* is that the difference will be tiny to non-existent.  But then who am I to know as I haven't tried it?  Not the first time I've been completely wrong

You could try moving the switch power to the dedicated circuit.  If it sounds worse then it would at least confirm the negative effects of the 108's SMPS.  But it's also dependent on the design.  For instance Statement also has a SMPS in each unit.

Yeah, 200 mile round trip was a bit tiring, even in same time zone!

i might actually try it one of these fine days -- it doesnt seem to be that expensive.

practically, i cant move the switch to the hifi radial as the sockets are in my listening room and the switch is in the conservatory. in any case, i dont really want to mess about with my nice radial circuit -- keep all the SMPS stuff away --else neurosis would kill me :-)

the SMPS's in Statement must be somewhat 'special' i guess... but then the Statement itself is special anyhow...

enjoy

ken

I thought the Statement SMPS is only there to handle standby functionality (to satisfy low power standby legislation). Once powered on it's disabled as the main PSU takes over. 

Graham Clarke posted:

Ken has a dedicated mains spur for his system.  Assuming the switch is NOT on that spur, do you still feel that changing the switch PS will make an audible difference?  My gut feeling would be "no", but then I've never tried this.

Ken and anyone else worried about their home network and audio playback...

I'd have thought one way to test out whether reducing network noise is influencing replay quality on the network streamers is to put a few favourite tracks you'd normally play over the network on a USB pen drive and play them via the NDX / NDS (whichever flavour you have).

Unplug the network connection - does it sound better / worse / no different ?

james n posted:

I thought the Statement SMPS is only there to handle standby functionality (to satisfy low power standby legislation). Once powered on it's disabled as the main PSU takes over. 

Yes, that's right.  It's to meet the EU regulations on low power usage when on standby of being under 1W.  Apparently just running the big transformer (keeping it ticking over) would be more like 10w.  It also handles the power on process, providing a soft start rather than a huge current in rush.

Here's a picture of an S1 Pre being assembled in the factory:

Transformer in the bottom (even larger in the monoblock) with capacitors above it and inputs/outputs to the right in a faraday cage.  In the monoblock we looked at the SMPS would be directly above the silver cage.

In the top box, the PCB mounted on the inside front of the unit contains an ARM processor that handles input selection and volume control.  To the right, located as far away from the ARM CPU as possible is the sandwich of PCBs (five I think) that are mounted on a large brass plate all of which is sprung with a resonant frequency of 12Hz to negate microphony.  This is the main preamp section.  There's also over 40 discrete power regulators in there! 

If you look closely (look behind the long ribbon cable in the top box) you can see that the metal panel of the far side has metal milled out of it.  That's partly to reduce weight so that it passes a 15 degree topple test and also to avoid internal reflections that a flat piece of metal can produce.  The attention to detail is just crazy.

A full set of pre and power takes 3 weeks for two people to assemble!

Sorry for drifting off topic, Ken.

Mike-B posted:

Graham Clarke posted:

Ken has a dedicated mains spur for his system.  Assuming the switch is NOT on that spur, do you still feel that changing the switch PS will make an audible difference?  My gut feeling would be "no", but then I've never tried this.

Hi Graham,  the SMPS is powering the switch circuits & the DC contains a level of switching noise that varies from 50kHz to several MHz & as I understand it this is the primary noise pollution,  into the LAN. 

Switching noise does also get fed back into the mains & I guess this is where your question is coming from, so yes,  it would not make much difference on that side if the SMPS gets its power from a different mains supply to the audio equipment.    

My SMPS gets it power via my UPS & with its isolation transformer & C&D mode choke, little to zero noise is getting back into my mains anyhow.  Despite this,  the addition of my iPower unit was noticeable.    

Interesting...  So the issue is more the noise into the LAN.  Have you experimented with a LAN isolator?  I see that they're available from that rain forest web place for a little over £100.

james n posted:

Ken and anyone else worried about their home network and audio playback...

I'd have thought one way to test out whether reducing network noise is influencing replay quality on the network streamers is to put a few favourite tracks you'd normally play over the network on a USB pen drive and play them via the NDX / NDS (whichever flavour you have).

Unplug the network connection - does it sound better / worse / no different ?

This sounds very sensible to me? though i suspect i will have to listen to a somewhat wider range of music than a 'few favourite tracks' -- but i believe thats what you meant anyhow...

enjoy

ken

Graham Clarke posted:

 

Sorry for drifting off topic, Ken.

not at all Graham -- very interesting indeed. we want more!!! are you going to post some more -- or are they already in your thread in hifi corner?

enjoy

ken

james n posted:

Ken and anyone else worried about their home network and audio playback...

I'd have thought one way to test out whether reducing network noise is influencing replay quality on the network streamers is to put a few favourite tracks you'd normally play over the network on a USB pen drive and play them via the NDX / NDS (whichever flavour you have).

Unplug the network connection - does it sound better / worse / no different ?

That's a test many of us have performed.  I used it to tune the RFI suppression on my network; eventually I found the right combination of ferrites on network and audio cables to get to a point where I couldn't distinguish between network and USB stick.

Yes you're right to point out that you need to disconnect the network to ensure a valid test (and power down the network SMPS in case they are injecting noise into the mains).

Huge posted:

james n posted:

Ken and anyone else worried about their home network and audio playback...

I'd have thought one way to test out whether reducing network noise is influencing replay quality on the network streamers is to put a few favourite tracks you'd normally play over the network on a USB pen drive and play them via the NDX / NDS (whichever flavour you have).

Unplug the network connection - does it sound better / worse / no different ?

That's a test many of us have performed.  I used it to tune the RFI suppression on my network; eventually I found the right combination of ferrites on network and audio cables to get to a point where I couldn't distinguish between network and USB stick.

Yes you're right to point out that you need to disconnect the network to ensure a valid test (and power down the network SMPS in case they are injecting noise into the mains).

for a number of reasons, this is turning out to be potentially a very useful test -- for a variety of other purposes -- and not just in conection with noise and SMPS issues...

Great, guys...!

enjoy

ken

Graham Clarke posted:

 

Interesting...  So the issue is more the noise into the LAN.  Have you experimented with a LAN isolator?  I see that they're available from that rain forest web place for a little over £100.

No I haven't, I prefer to suppress noise at source.  Medical grade isolators are for hazard rather than noise & are miss-sold for audio applications. I tried the little black connector on a wire thing & was not impressed.   RF level noise can carry by induction through all sorts of devices & isolators are not always effective.    Besides which if the noise starts from the switch,  then it means all ethernet branches off that switch need to be isolated  ...........  not the best solution IMO.    

ken c posted:

Graham Clarke posted:

 

Sorry for drifting off topic, Ken.

not at all Graham -- very interesting indeed. we want more!!! are you going to post some more -- or are they already in your thread in hifi corner?

enjoy

ken

That's pretty much everything, Ken.

Mike-B posted:

Graham Clarke posted:

 

Interesting...  So the issue is more the noise into the LAN.  Have you experimented with a LAN isolator?  I see that they're available from that rain forest web place for a little over £100.

No I haven't, I prefer to suppress noise at source.  Medical grade isolators are for hazard rather than noise & are miss-sold for audio applications. I tried the little black connector on a wire thing & was not impressed.   RF level noise can carry by induction through all sorts of devices & isolators are not always effective.    Besides which if the noise starts from the switch,  then it means all ethernet branches off that switch need to be isolated  ...........  not the best solution IMO.    

I understand the logic of suppressing noise at source but if you were to use an isolator why would you need to use it anywhere except immediately before the streamer, given that's the only audio device(s) connected to the network.

Graham Clarke posted:

.........   I understand the logic of suppressing noise at source but if you were to use an isolator why would you need to use it anywhere except immediately before the streamer, given that's the only audio device(s) connected to the network.

I use an isolator (one from Baske medical) and found it beneficial in my streamer days. My switch (GS105) and other network hardware is at the end of a 30m length of bog standard Cat5e  in another room in the house. I've tried different power supplies on the switch (plug top SMPS, Linear and a heavy duty linear lab supply) and Audioquest and Meicord cables with no real differences for me. Apart from the Isolator, the only thing i've consistently found to be avoided (in my system) is a switch near the hi-fi with a short ethernet cable to the streamer (something Linn recommended) - it always hardens the sound which at first can sound more detailed and exciting but with further listening is just a bit gritty and tiring. Not nice. As always it's worth experimenting as our environments and networks are a lot different. 

James

is it an accepted 'fact' that all USB sticks sound the same when music content is played into e.g. NDS?

enjoy

ken

Hi Ken,  I'm sure someone somewhere will say there is a difference, there have been a few comments on this forum.  Problem is most people just use USB flash drives for storing/playing music & although I expect some people will offer an opinion,  I doubt any have actually done a real flash drive make/model/size SQ comparison test.

All this talk of electrical noise on the Ethernet cable fails to address the elephant in the room, TCP/IP noise. Depending on media transfer rate on the network I have observed two quite different modes of TCP operation. One creates more two way network traffic than the other.. and to my ears they 'sound' different. The method is for the most part determined by the TCP data rate from the server and how the streamer responds.

In my listening tests more rapid bursty transfers sound slightly better than slower more consistent transfers. The latter requires the streamer/renderer to create far more network traffic back to the server.

i think, yet I need to do more measurements, this accounts for differeing qualities in part from different media server /OS stacks and the varying quality from streaming services such as Tidal.

ok not quite as exciting as gravational waves.. but to date this phenomenon has been largely ignored as dragons have been chased elsewhere. I will create a dedicated post at some point with findings... but I will wait until the current beta firmware is released.

Simon 

I need to do more measurements ........... I will create a dedicated post at some point with findings

..... ....    looking forward to this Simon

ken c posted:

is it an accepted 'fact' that all USB sticks sound the same when music content is played into e.g. NDS?

enjoy

ken

Probably - but then you'd be introducing yet another variable. Just pick one and go with that. 

Simon-in-Suffolk posted:

All this talk of electrical noise on the Ethernet cable fails to address the elephant in the room, TCP/IP noise. Depending on media transfer rate on the network I have observed two quite different modes of TCP operation. One creates more two way network traffic than the other.. and to my ears they 'sound' different. The method is for the most part determined by the TCP data rate from the server and how the streamer responds.

In my listening tests more rapid bursty transfers sound slightly better than slower more consistent transfers. The latter requires the streamer/renderer to create far more network traffic back to the server.

i think, yet I need to do more measurements, this accounts for differeing qualities in part from different media server /OS stacks and the varying quality from streaming services such as Tidal.

ok not quite as exciting as gravational waves.. but to date this phenomenon has been largely ignored as dragons have been chased elsewhere. I will create a dedicated post at some point with findings... but I will wait until the current beta firmware is released.

Simon 

Interesting Simon. Are you looking at internal noise generated by the streamer card too ? I'd certainly expect to see a different noise spectrum depending on what the card was doing (which would correlate with the subjective differences you've observed with differing levels of network traffic). Will be interested in what you find.

James

James - my measurements are based on looking at network data traffic to and from the renderer and subjective assessments on any changes to the resultant audio. I am not using a spectrum analyser to look at noise within the streamer / streamer powerlines - simply because I don't own a sufficiently capable spectrum analyser - if I had the cash I might be tempted with that 552 that is appeared in an almost new used state at my local emporium

But yes the implication is that the differences heard are due to the network card and TCP stack working harder on the renderer in different modes and therefore causing differing perturbations.

This observation was undertaken in more detail in response to  some  beta testing I volunteered to undertake for Naim.

Thanks Simon - should be interesting reading 

James