I have perfected my network for the nth time


This is the complete summation of my efforts to optimize my network for streaming music. I started out with a combination of wired ethernet and fiber optic and ended with a completely wired network with passive filtration. This article represents only one of the myriad ways that exist to set up a network around myriad component choices. Your network will be a variation on mine as all systems are ultimately unique. It also represents a “no stone unturned” effort around the realization that “everything matters.” As such, I hope it will be a reference to those newly starting out, or to those currently in the process of searching for the best sound.

Let me humbly share my journey of tweaking my streaming network that I built around a Small Green Computer bundle. Back when I initially set up this configuration I modeled it exactly after Michael Lavorgna’s recommendations, now of Twittering Machines. It was stated that wired ethernet contains a multitude of music destroying noises that must be dealt with. Who was I to argue. At the time I didn’t even know what an ethernet switch was (a free- lance musician and private music teacher leading a sheltered existence) .

The bold type face in the following diagrams indicate elements that will be changed and explained as you progress through the text:

FO through wall > ONT,smps > 8tac tenrehte > router,smps > cat 8 > $15 switch,smps > cat 8 > i5 Transporter,smps.

Same switch > cat 8 > FMC,smps > Fiber optic > FMC,smps > cat 8 > ultraRendu with LPS > Pangea Audio Premier SE MKII USB Cable > Benchmark DAC3B.

I had completely solved all power problems in my system to achieve a superior black background. The above still had grunge :

1. I added a LPS with a Y adapter to power both FMC = better

2. I added a LPS to the Transporter = better

3. I replaced the LPS power chords with Pangea Audio AC-14SE MkII = better

4. I wrapped my fiber optical cable in bubble wrap. It seems they are sensitive to vibrations. = better

5. I replaced the 3 post switch cat8 ethernet cables with Pangea Premier SE = way way way better, but STILL some remaining niggly grunge!

6. I had not bought into the $700-$10,000+ ethernet switch mania. I had bought my $15 plastic covered switch from Home Depot. I did try a more expensive switch and sent it back because I could hear no improvement. I rewired post switch and removed the FMCs. LISTENED. Then I added the FMCs back in. They were definitely improving the sound, by a lot!

I ordered a iFi iPower X Ultra Low Noise AC/DC Power Supply for the switch. I took out the FMCs again and installed the new PS. WHAM!! The sound came into complete focus. I reinstalled the FMCs and now a huge amount of grunge had disappeared.

7. The new Pangea Audio Premier XL MKII USB Cable came on the market, which separates out the 5v current wire from the signal wires. Because my Benchmark DAC was using the 5v current, I replaced my previous Pangea Audio Premier SE MKII USB Cable (same wire and connectors) and wiped another smear of noise from the window. This is the single biggest improvement of all these suggestions.

8. Shutting the WIFI off in my primary router and using a wireless access point to regain WIFI for the house. It has resulted principally in opening up the sound-stage : width, depth, with clearer separation. This was the elephant masking the remaining network problems. I subsequently had to go back into my router when I realized I also had to disable the broadcasting of the router name and also disable the WAN feature to get the complete quiet I was looking for. The result was truly  startling.

9. I lived with this WIFI tweak for two weeks and I gradually became aware of a persistent coloration that I eventually came to realize was masking the harmonic structure of the music. Once I defined the problem there was no ignoring it. The problem had to be solved.

I started digesting the suggestions in this thread for optimizing an optical filter as I watched the $$$ mount while listing things to buy. I remembered that Mr. Lavorgna moved on by replacing his optical filter with a GigaFOIL. I found a thread on this very subject from that time period. Some level-headed fellow suggested removing the filter for a listen before proceeding. He had become unconvinced of the need for filtering. It was simple to connect an ethernet cable directly from my switch to the Sonore ultraRendu. I pushed play on Qobuz not knowing what to expect. Not only was the coloration banished, but the musicians exploded into the room, the result of expanded dynamic freedom . I was sitting there like a one-year-old that had seen his first Jack-in-the-Box! I stayed up late last evening listening to favorites with profound new revelations at every turn of phrase.

I decided to redo the test I had done previously comparing Dudamel conducting Ives Symphony 2 : Qobuz 24/96 vs CD 16/44.1. The first time I reported that they sounded identical. Now they did not. The Qobuz sounded like a SACD and the CD sounded like a CD, which is how it should sound when comparing different resolutions. The CD sounded like the colored generic-optical stream and the all-wired stream of Qobuz sounded much more open and revealing of a multitude of detail.

My network is now simplified being all wired ethernet with no optical:

FO through wall > ONT,smps > 8tac tenrehte > router,smps > cat 8 > $15 switch with iFi PS > Pangea ethernet > i5 Transporter with LPS

Same switch > Pangea ethernet > ultraRendu with LPS > Pangea Audio Premier XL MKII USB Cable > Benchmark DAC3B.

 

10. Let’s do some testing. I stumbled on this looking for other advice. It was reported that the ethernet cable direction makes a difference. My findings: definitely!!!!!

Test piece : Qobuz stream 16/44.1 : Elliott Carter "Horn Concerto" which is a highly detailed and dramatic work sure to reveal any differences.

First, I reversed cat8 between ONT and router and a veil was removed!!

Second, I reversed Pangea ethernet between switch and Sonore ultraRendu, and I had it right the first time.

Third, I reversed Pangea between switch and Transporter i5 resulting in severe sonic degradation. Back to the way it was originally.

I don’t have the courage to reverse the 50’ cat8 running under my floor and suspended with cable ties, but extrapolating from what I heard, I think I got it right(grin).

Three out of four isn't bad, but that one reversal has taken things to a new level of dynamic freedom and sonic clarity.

I was grateful for this new discovery because it gave me a foundation to further explore ethernet cabling in my system. Initially I researched out an inexpensive Chinese brand of cat8 that had excellent published specs and just went with it. When I substituted Pangea cables post switch (Every component in my system has to be best value for the dollar, systematically synergistic, and musical sounding. No budget busting or retirement account draining devices allowed!) I got a nice jump in clarity and left it at that. Now, when I had just removed the optical filter from my network, I had an extra Pangea and decided to try it between my ONT and router.

This involved some work as the cable was short necessitating me moving my router much closer to the ONT. I removed the Ethernet cable whose direction had been determined, with a cable whose direction was not determined. I had to try it both ways and then go back and forth between cables a couple of times, and the Pangea won. The Chinese cable sounds wonderful but just a little smooth and a little less open on the top. This is subtle but clear to me, and the more open cable leads to less fatigue in longer listening sessions.

This suggests that the remaining 50' Chinese cable MAY need to be replaced and I ordered a 20m Supra cat8 and Linkup cat8.

11. Ladies and Gentlemen : the results of The Great Ethernet Cable Shootout. The contenders:

a. Pangea Premier SE cat8 26awg Cardas Grade One silver-coated copper, audiophile claims .6m (Audio Advisor)

b. Gigaware cat6 24awg 15m (Radio Shack)

c. Chinese No-name cat8 26awg 15m published specs (Amazon)

d. Linkup cat8 22awg 15m published specs (Amazon)

e. Supra cat8+ 26awg 20m audiophile claims, published specs (ebay)

All cables were individually listening tested for directionality. It was conclusively unanimous, all of the ethernet cables sounded better one way than the other. When installed backwards the principal result was a shrinking of the sound-stage. It seemed the longer the cable the more dramatic the effect. Since none of the cables had marked arrows, the chances of correct installation when testing a new cable was 50%, and that is what I achieved. With a little experience I could hear the narrow sound-stage in less than a minute. These results are corroborated in Audio Quests ethernet cable literature :

Directionality
All audio cables are directional. The correct direction is determined by listening to every batch of metal conductors used in every AudioQuest audio cable. Arrows are clearly marked on the connectors to ensure superior sound quality. For best results, have the arrow pointing in the direction of the flow of music. For example, NAS to Router, Router to Network Player.

After I had corrected the reversed cable in my network, and replaced the short Chinese cat8 with Pangea, the sonic improvements led me to suspect the 15m Chinese cat8 as a potential bottleneck in my network.

ONT > Pangea > router > WINNER of the Shootout > switch > Pangea > Sonore ultraRendu

A. I installed the Gigaware cat6 that I used when I first started to experiment with streaming years ago. The sound was horrible. It reduced the sound of $$$$ equipment to the sound of $$ equipment.

B. The Chinese cat8 was good sounding and I could get Qobuz 24/96 tracks to sound like SACD. When after weeks of reading this thread and working on my network I listened to silver discs again, I found a subtle difference which I felt I needed to pursue, leading to this test.

C. The Linkup cat8 is an industrial strength cable, unusually thick and stiff. My initial impression was it is going to need significant break in time. After 24 hours it sounded better but in no way settled. My impression was that it is a system dependent cable. I am pretty sure it will never be the right fit for me.

D. The SLAM-DUNK WINNER is SUPRA 8. When I dropped this into my system the instincts leading to this test were completely justified. Indeed, the Chinese cable was holding the sound hostage. The sound-stage opened to new unheard of width and depth. The dynamics were so free and the musicians were so present, I wondered why I could not smell their sweat. I did not think to look at my speakers because it was obvious there was no music was emanating from them!

Upon reflection there is a synergy between the Pangea and Supra, a perfect marriage. Each allow the other complete freedom to do its thing. Since I will never hear a complete loom of either Pangea or Supra, I cannot conjecture what they would sound like. I don’t care, because when you have found the love of your life, you will only confuse yourself or do harm by continuing to play the field.

There are so many other brands/models to choose from. Many of you have your favorites. For me the above setup is a cost-effective value-for-performance solution that I always search for.

While I was testing, it was easy to run each ethernet cable to my new wireless access point and stream music to my laptop > iFi usb filter > AudioQuest Dragonfly Red > headphones. In each case, the sonic signature of the cables described above was broadcast over WIFI. This suggests there is no free lunch and a quality cable is needed.

 

12. I gathered the pieces to replace the smps for my ISP ONT and router:

a. iFi elite 12v 4a PS

b. DC cable Y adapter 18awg

c. Pangea 14awg PC

d. Brickwall surge protector

The installation of them opened the soundstage to cavernous dimensions, brought unbelievable separation to complex music, and a wonderfully beautiful tonality.

Ladies and Gentlemen, I declare my network optimized!

 

  1. I am back. After getting through #12 above, I was a tired and worn out audiophile and needed to return to being a melomaniac, my chief feature. After three months of joyfully listening to music on LPs, silver discs, and Qobuz, the audiophile lights started to flash into my attention. I was listening to a CD of Brahm’s Symhony 2, Dohnanyi, Cleveland Orchestra on London. The performance was immersive and the sound was definably solid in audiophile terms. I wondered how it would compare to the streamed 16/44.1 version. Here we go down the rabbit hole. There was no favorable comparison. The streamed sound was bleached and the dynamics impeded. Could this be be the “ethernet noise” that I had so far been unable to quantify? Could I do a quick test on my hypothesis that did not use fiber optics?

 My network with possible filtration points: 1,2 etc.

ONT > cat 8 > (5) router (4) > cat 8 > $15 switch > cat 8 > (3) i5 Transporter

Same switch > cat 8 > (2) ultraRendu (1) > USB cable > DAC

I remembered that I had an iFi iSilencer USB noise filter that I had purchased to use with my Audio Quest Dragonfly Red DAC some years ago. I had considered trying it in the big rig several times but considered it unworthy due to prejudice (no one is immune). When I tried it in position 1 above there was a definite improvement. iFi has a new version called iFi iSilencer+, which I ordered and was rewarded with another big step in enhanced performance.

Was I done? I noticed iFi had an iFi LAN iSilencer in their catalogue. I reached out to Ifi support and asked what I could expect if I piggy-backed these in the same network. I got an immediate, thoroughly helpful, and cheerful reply stating I would probably experience “diminishing returns”, but you never really know until you try for yourself. Off went the order. I tried it in position 2 and it was not optimal to have two filters on the in and out of the same device. There was a slight but discernable squashing of the sound-stage. When I moved the LAN filter to position 3 I got a huge improvement that was additive. I moved it from position 3 to 4 and heard a new kind of good while sacrificing the previous good. I ordered a second LAN filter.

Trying LAN filters in positions 3 and 4 and the USB filter in position 1, I was rewarded with an other additive upgrade in performance! I moved one filter from position 4 to 5 and it just sounded wrong.  Back to 4 and I was done filtering for $250 plus tax and no new boxes, PS, PC, etc.

  1. While I was waiting for the various filters to arrive I realized I had not yet experimented with mechanical isolation with my router, PS, surge protector (12 above). I bought another bag of hockey pucks and used stacks of 2 under devices until they were stable and achieved another huge improvement! I also tested elevating the PCs and 20m Supra CAT 8+ ethernet cable from the floors around the router and in the listening room. I was rewarded again. DO IT!

When you are working on your network and pondering your progress, be sure to compare “apples to apples” or CD verses 16/44.1 as your principle test. When they match you have truly achieved something. There are many files on the streaming services that are not the same as the CD so you have to use your experience to make sure you are on the right track. I am now getting as close to equality comparing CD to streaming 16/44.1 as I am likely to get and have probably achieved parity. The improvements I have brought to CD level streaming have scaled nicely to higher resolution streams, opening up sonic vistas that were subtly masked before and now stunning to witness. After all, what is a direct comparison to these resolutions? My streaming has become a truly satisfying experience.

Ladies and Gentlemen, for the second time, I declare my network optimized (until the audiophile lights start flashing)!

 

Now I am going to jump into the weeds and try to draw some conclusions :

a. There are two (three) ways to skin this network cat : All wired or conversion to all optical, both can sound equally good, although I personally have not heard optimized optical. (In addition there probably are those who have optimized WIFI setups.)

b. I think another reason for my success with all wired ethernet is the installation of a whole-house surge protector. This not only has lowered the line voltage from 126v to around 122v, but also seriously lowered the electrical noise floor. This has benefits for a network being installed in multiple rooms and on different electrical legs.

c. I am sensing that before the WIFI tweak (ref 6 above), the optical filter could remove some, but not all of the upstream noise, suggesting a confirmation of the original poster’s assertion of signal damage. This could and should be investigated by someone with the right equipment and expertise. What happens to the square wave when subjected to intense RFI in these router combos?

d. Wired network optimization now consists of the tried and true audiophile tweaks of quality cabling, ALL THINGS POWER, mechanical isolation, and passive filtering. In addition optical network optimization needs the highest quality converters with additional LPS and PCs, transceivers, and optical cable. This can be effective, but is it necessary? It surely is expensive. (Again, BOTH need to use a separate Wireless-Access-Point.)

e. IMHO, for those newly setting up their network, I highly recommend all Pangea digital cables, the highest model only, available with a 30 day return. Give them a listen against the mega-buck cables or the cheapies you may be using and see what you think.

This post reveals that I make no changes in my system unless I have clearly defined a problem that needs a solution. I listen to music with my trained and experienced musician’s ears. Problems always surface when I get frustrated trying to parse more meaning out of the performance/sound. My system has been meticulously tweaked. When one first starts tweaking, the sonic changes are minute because there are still so many problems masking clarity. However, as you get to the end, each new tweak is a revelation/game-changer. They are harder to find but Oh!-so-rewarding. If I can find anything new I’ll report back.

This was a four year journey. I have learned a ton along the way. I hope it is helpful to someone else.

singingg

@kennyc 

I applaud your efforts but as you can see, you’ve pretty much wasted your time and energy. It’s best to ignore folks like fredrik ….their singular purpose is to provoke and cause discord. Unfortunately, Mods here are much more tolerant and that’s why they can get away with their continuous nuisance and bullying. 

Birds gotta fly
Trolls gotta troll.

@lalitk +1 and well said.  Again I quote Pink Floyd from The Wall — “It’s not easy banging your heart against some mad bugger’s wall.”

Daily gibberish of singingg 6

 

Now that we have fought WW III & IV on this thread, I hope we can stop feeding the troll and refrain from further psycho-analysis, and get back to the discussion of the OP.

 

  1. Here is another input about PARITY from Michael Lavorgna:

 

Review: Technics SL-G700M2 Network/SACD Player

https://twitteringmachines.com/review-technics-sl-g700m2-network-sacd-player/

 

It is not easily achievable but well worth pursuing for the sonic rewards and the education of the journey. You can see many have not faced the challenge head on yet.

 

 

Now folks, I as a layman, am going to jump into the weeds and try to discuss a couple of issues, both of which have become conflated, and have been plaguing everyone. I invite polite constructive discussion. I am way less than an expert and just want to learn.

 

 

  1. My final statement on ethernet cables. Cable Bidirectionality by Design vs. Efficiency of Signal Propagation=Signal Directionality. Take a look at the diagram: 

A. Ethernet cables need a two-way communication accomplish their task. This is easily seen in the diagram. The green wires on pins 1& 2 always Transmit, and the orange wires on pins 3 & 6 always Receive. If you reverse the direction of the wire, pin 1 will always connect to pin 1. The design is for bidirectional communication.

 

  1. Since all the wires in the cable are metal, whether copper or silver plated copper or pure silver, they have been drawn through dies creating surface anomalies and therefore exhibit better signal propagation in one direction verses the other. This is established fact. (I am sure a simple search can pull up numerous studies.)

 

Now say we are transmitting music from a switch to a renderer and the signal is traveling from left to right in the diagram, if the green wires direction are aligned with the signal direction, then it MAY sound better. We then test the sound of the resulting music comparing one direction to the other. I have done this and reported the results in the OP. Now it is your turn to test and report. (How to align the other wires for the ultimate design, I leave to the engineers.) In conclusion, don’t equate bidirectionaliy with directionality.

 

  1. The topic of the sound of ethernet switches continues. I think an intellectual error has been made and continues to be propagated in conflating the sound of a simple switch (doing its switching) vs using a switch as as a device to solve the bugaboo of ethernet noise (or other issues).

 

It took me a long time to finally identify and quantify the destructive nature of said noise (13 in the OP). In fact, I published a post entitled “The Myth of Ethernet Noise,” which I have withdrawn. I have tried to keep the switch simple ($15, believe it or not, and my streaming is sounding A+ with this device). The only switch upgrade was a high quality PS which really enhanced its performance through giving it the energy it needed. It has been my approach to find the actual source of the signal degeneration and creatively find its solution there. This is not stated to ruffle any feathers, because I am as aware as any of the myriad design choices and solutions for setting up a network/system. Just become aware that you are asking the switch to switch and solve other issues that may be best addressed elsewhere, OR NOT, depending on YOUR SYSTEM DESIGN APPROACH and WHAT YOUR EXPERIMENTAION LEADS YOU TO CONCLUDE.

 

To take one example:

 

When people replace switch X with switch Y “that does more things,” and then conflate, they say that switch Y sounds better than switch X. However, with pinpointing, the actual issue might be with the router spilling noise. Is it better to address the noise at the router with passive or optical filtration, or let the switch at the next junction down the road deal with it in its way? Are both points equally successful at addressing the issue, or is one better than the other? You have to experiment and make these design choices for your system. So does switch Y sound better than switch X when it is switching?

 

I have not yet found that switches sound different if well powered. However, that could change at any time if I HEAR something that jolts me awake.

 

CHEERS!

@singingg 

All very valuable points here, indeed. As to switches: There is value in reclocking the signal since the noise incursions tend to do damage to the actual shape of the sine wave of the signal thereby rendering the exact determination of peaks and troughs more imprecise. Reclockers  such as InnuOS Phoenix Net or LhyAudio SW-6/8/10 have a demonstrable positive impact on SQ.

Since all the wires in the cable are metal, whether copper or silver plated copper or pure silver, they have been drawn through dies creating surface anomalies and therefore exhibit better signal propagation in one direction verses the other. This is established fact. (I am sure a simple search can pull up numerous studies.)

FWIW, I lean heavily towards subjectivism but Brother, this is NOT established fact. The late Chris Sommovigo for example did not believe in the concept. He is not alone among wire (audio cable) experts.