A novel way to massively improve the SQ of computer audio streaming

[austinpop]

1 hour ago, auricgoldfinger said:

@mozes and I can vouch for this cable if you decide not to take the DIY route.  The price is very reasonable given the current CAD/USD exchange rate.

 

https://audiosensibility.com/blog/sotm-audio/#!/Signature-Silver-SOtM-DC-Cable/p/102414197/category=27007760

 

Me too. I went ahead and JSSG 360'd this for a further boost in SQ.

 

Maybe the holy grail is a 7N UPOCC silver star quad with JSSG 360?

[mozes]

4 hours ago, auricgoldfinger said:

@mozes and I can vouch for this cable if you decide not to take the DIY route.  The price is very reasonable given the current CAD/USD exchange rate.

 

https://audiosensibility.com/blog/sotm-audio/#!/Signature-Silver-SOtM-DC-Cable/p/102414197/category=27007760

Agree highly recommended!

[jean-michel6]

3 hours ago, Johnseye said:

 

 

 

Yes, I have an LPS 1.2 and 1.  The specs are a little different so can support different deivces, and I was considering the SOtM master clock at the time which would have afforded a discount.

 

Since you three have some experience with the Hirose connector, do you think it will work with the Gotham Star Quad ?  My goal will be to replace all DC cables with it.  Also, can you crimp it with a coax crimper or does it need something special?

 

 

Thanks.  I have a similar silver cable from SOtM.  Not as good as that one though, the gauge is smaller.  I may just JSSG 360 that cable if the Hirose is too difficult to work with.

I have made one of this cable with an Hirose sotm connector and canare 4S6 JSSG360 .  I think it should be feasible with the gotham as the conductors have about the same section as the one in the canare .

However it is a pain ... to do it . 

I have use the hirose connector with pin to be soldered . You must also use heat shrink tubing to protect the connections as they are very close to each other .

[jean-michel6]

3 hours ago, austinpop said:

 

Me too. I went ahead and JSSG 360'd this for a further boost in SQ.

 

Maybe the holy grail is a 7N UPOCC silver star quad with JSSG 360?

 

I agree !

 

Somebody to do it ?

 

[elan120]

Network switch clock mod - part-2 (Big thanks to Alex C.)

 

With input and help from @Superdad, I decided taking some time this morning to give Crystek CCHD-575 oscillator a test drive, replacing the Conner Winfield OH4 I retrofitted a week ago.  Again, after having the oscillator mounted on the carrier board, its measurement look good, and the rest was quite straightforward.  After the new oscillator was retrofitted, initial listening impression is very positive, but I will wait for some additional warm-up and burn-in time to see if there are going to be changes...hopefully even more improvements.

 

Below are some pictures if anyone is interested in finer details...

 

Crystek CCHD-575 initial power up measurement:

 

Completed carrier board:

 

Completed new oscillator retrofit:

 

Carrier board close up:

 

[Superdad]

8 minutes ago, elan120 said:

Network switch clock mod - part-2 (Big thanks to Alex C.)

 

Hi Kevin:  Looks great!  Please remind us what model of Ethernet switch your mod is in.  And have you done anything else to that switch besides replace the clock?

[elan120]

4 minutes ago, Superdad said:

 

Hi Kevin:  Looks great!  Please remind us what model of Ethernet switch your mod is in.  And have you done anything else to that switch besides replace the clock?

Hi Alex,

 

This is the ZyXEL GS-108B V3 switch, and it is being powered by LPS-1.  I have not done other mods yet as I plan on comparing this modded switch with your suggested Cisco WS-C2960-8TC-L first, and if this turnout to be better, I will change out the two switch mode regulators.  In the meantime, once I received the Cisco switch, I will see if the same clock mod can also be implemented as well as better power source.

[RickyV]

2 hours ago, elan120 said:

Network switch clock mod - part-2 (Big thanks to Alex C.)

 

With input and help from @Superdad, I decided taking some time this morning to give Crystek CCHD-575 oscillator a test drive, replacing the Conner Winfield OH4 I retrofitted a week ago.  Again, after having the oscillator mounted on the carrier board, its measurement look good, and the rest was quite straightforward.  After the new oscillator was retrofitted, initial listening impression is very positive, but I will wait for some additional warm-up and burn-in time to see if there are going to be changes...hopefully even more improvements.

 

Below are some pictures if anyone is interested in finer details...

 

Crystek CCHD-575 initial power up measurement:

 

Completed carrier board:

 

Completed new oscillator retrofit:

 

Carrier board close up:

 

Good job Elan120, very intresting.

[cat6man]

On 7/5/2018 at 3:06 AM, jean-michel6 said:

I do had a dx-USB HD with sCLK 2224 and I did upgraded it to sCLKex.  

It is now powered by an lps 1.2

My dealer here in france bought back from me at half cost the sCLk224 board and i did purchase the sCLKex. 

I am using it to input into AES-EBU in an AMR dp 777 se dac. 

I did got some SQ improvement going from sCLK 2224 to sCLKex but it is not very big. 

In my opinion it is a worthwhile investment if you plan to use the two additional clock output of the sCLKex and adding later a 10mhz ref clock which I have not done yet. 

One other route you may try. 

I did try in my system the new matrix USB-Spdif converter.  

It is better than the sotm with sCLKex board and cost is small .... cheaper than sCLK ex 

 

Merci Jean-Michel,

 

The matrix x-spdif 2 looks like an interesting option for me to try, as you say that it outperformed the dx-usbHD with the sCLKex clock and is cheaper.  I assume you were comparing the aes/ebu outputs of the two and both were powered by the lps1.2 (also what I use for  the dx-usbHD).  You noted that the change from the sCLK 2224 to sCLKex was small.  How would you describe the improvement going from sCLKex to matrix x-spdif2?

[seeteeyou]

22 hours ago, elan120 said:

I have not done other mods yet as I plan on comparing this modded switch with your suggested Cisco WS-C2960-8TC-L first, and if this turnout to be better, I will change out the two switch mode regulators.

 

It's an interesting switch to mod since there were so many Naim fans who loved that as @charlesphoto mentioned below

 

https://www.computeraudiophile.com/forums/topic/37034-smps-and-grounding/?page=38&tab=comments#comment-816134

 

Even Netgear (without network leakage issues) seemed to be no match

 

http://forums.naimaudio.com/topic/cisco-switch
http://forums.naimaudio.com/topic/cisco-2960-set-up

http://forums.naimaudio.com/topic/cisco-2960-router-confusion

http://forums.naimaudio.com/topic/cisco-2960-recommendation

http://forums.naimaudio.com/topic/cisco-2960-8tc-vs-paul-pang-tcxo

http://forums.naimaudio.com/topic/cisco-catalyst-ws-c2960-8tc-l-2960

http://forums.naimaudio.com/topic/cisco-router-to-match-cisco-ws-c2960-8tc-l

Quote

GS105 was my first switch introduction, yes it was an improvement, but the 2960 blew it away with ease.

 

Now the question is what voltage(s) would that switch actually needs. Cisco Power Supply 341-0208-01 should be providing 48V since WS-C2960-8TC-L only needs 20W while the current draw is 0.5 to 0.25 A

 

https://www.cisco.com/c/en/us/td/docs/switches/lan/catalyst2960/hardware/installation/guide/2960_hg/higspec.html

 

Besides, we could actually find two system clocks instead of one

 

 

 

In other words, we'll have lots of fun if both of them were replaced with better ones.

[jean-michel6]

22 hours ago, cat6man said:

 

Merci Jean-Michel,

 

The matrix x-spdif 2 looks like an interesting option for me to try, as you say that it outperformed the dx-usbHD with the sCLKex clock and is cheaper.  I assume you were comparing the aes/ebu outputs of the two and both were powered by the lps1.2 (also what I use for  the dx-usbHD).  You noted that the change from the sCLK 2224 to sCLKex was small.  How would you describe the improvement going from sCLKex to matrix x-spdif2?

Yes I was comparing both unit on their aes/ebu output and both were powered by lps 1 . I did receive my lps 1.2after I made this comparison.  

 

The matrix was was more transparent than the sotm and medium high and top end was smoother .Overall it was nicer and sounding more analogic. Soundstage was also improved. 

 

For the time being I still run with the sotm because I am using the 2 extra clock of the clockboard and I want also to test the external 10mhz ref clock. 

But in the near future I intend to move to the matrix. 

[Superdad]

2 hours ago, seeteeyou said:

Now the question is what voltage(s) would that switch actually needs. Cisco Power Supply 341-0208-01 should be providing 48V since WS-C2960-8TC-L only needs 20W while the current draw is 0.5 to 0.25 A

 

Well per @lmitche‘s urging I recently bought that exact model of Cisco Catalyst switch.  It is a heck of a piece—with way more in it than a $20 Netgear—not surprising since its original list price was probably $600+. ( @JohnSwenson thinks one of the big chips in it may be one  that he worked on at LSI/Avago—I need to send him close up photos to be sure.)

 

I’ve been running microRendu > ISO REGEN > modified SU-1 > Holo Spring L3 (the first 3 each with their own LPS-1.2) fed from a 25-foot BlueJeans/ Belden CAT6a—direct from my 4GHz iMac, no swtich, bridging trick.

 

So I put the Cisco on the floor under the audio rack, plugged the long Ethernet cable into port 1 and a 2-foot BJC CAT6 into port 8 to feed the microRendu.  Did not expect much since when I did this with a Netgear FS105 it was a wash or maybe worse than without.  

But whoa!  This was immediately better! More lively and musically engaging. Very nice.

 

The LiteOn SMPS in this switch is 12V/0.6A (if I recall correctly; 12V for sure, and while the supply might be 1A, Larry told me draw is 0.6A), so I could easily run it from another LPS-1.2.  But I’m busy, and since we are working on our own from-the-ground-up highly isolated/specially clocked switch (see EtherREGEN), I don’t know that I’ll bother at this point.

 

This certainly does bode very well for our own unit.  While I’ve be confident that EtherREGEN will technically perform as planned (certainly better than the modified switches being sold), I was not sure how much sonic impact it would have for the average user.  But what I heard with this Cisco switch was not subtle, and it beat the direct connection from the computer!  

Color me excited...

 

—Alex C.

[austinpop]

29 minutes ago, Superdad said:

 

Well per @lmitche‘s urging I recently bought that exact model of Cisco Catalyst switch.  It is a heck of a piece—with way more in it than a $20 Netgear—not surprising since its original list price was probably $600+. ( @JohnSwenson thinks one of the big chips in it may be one  that he worked on at LSI/Avago—I need to send him close up photos to be sure.)

 

I’ve been running microRendu > ISO REGEN > modified SU-1 > Holo Spring L3 (the first 3 each with their own LPS-1.2) fed from a 25-foot BlueJeans/ Belden CAT6a—direct from my 4GHz iMac, no swtich, bridging trick.

 

So I put the Cisco on the floor under the audio rack, plugged the long Ethernet cable into port 1 and a 2-foot BJC CAT6 into port 8 to feed the microRendu.  Did not expect much since when I did this with a Netgear FS105 it was a wash or maybe worse than without.  

But whoa!  This was immediately better! More lively and musically engaging. Very nice.

 

The LiteOn SMPS in this switch is 12V/0.6A (if I recall correctly; 12V for sure, and while the supply might be 1A, Larry told me draw is 0.6A), so I could easily run it from another LPS-1.2.  But I’m busy, and since we are working on our own from-the-ground-up highly isolated/specially clocked switch (see EtherREGEN), I don’t know that I’ll bother at this point.

 

This certainly does bode very well for our own unit.  While I’ve be confident that EtherREGEN will technically perform as planned (certainly better than the modified switches being sold), I was not sure how much sonic impact it would have for the average user.  But what I heard with this Cisco switch was not subtle, and it beat the direct connection from the computer!  

Color me excited...

 

—Alex C.

 

Welcome to the club, Alex. I'm glad you've now heard "what's possible."

[afrancois]

8 minutes ago, austinpop said:

 

Welcome to the club, Alex. I'm glad you've now heard "what's possible."

Just ordered a WS-C2960-8TC-L on usedcisco.de.

I want to know if it's better than my AQVOX :-)

[tboooe]

39 minutes ago, Superdad said:

So I put the Cisco on the floor under the audio rack, plugged the long Ethernet cable into port 1 and a 2-foot BJC CAT6 into port 8 to feed the microRendu.  Did not expect much since when I did this with a Netgear FS105 it was a wash or maybe worse than without.  

But whoa!  This was immediately better! More lively and musically engaging. Very nice.

Ok I'm sold.  Just picked up this switch from eBay for $26.  Very curious how this compares to my Cisco SG300-10 switch.  

[Superdad]

15 minutes ago, tboooe said:

Ok I'm sold.  Just picked up this switch from eBay for $26.  Very curious how this compares to my Cisco SG300-10 switch.  

 

Well I have an SG300-10 (for the rest of my network) and I distinctly preferred the direct Ethernet connection over it (though admittedly it just has its cheap wall wart PS).

Thus my surprise at the very positive result of the Catalyst 2960 above the direct connection.

And for $26 there is near-zero risk!  Do report back with your own impressions Tommy. 

[tboooe]

18 minutes ago, Superdad said:

 

Well I have an SG300-10 (for the rest of my network) and I distinctly preferred the direct Ethernet connection over it (though admittedly it just has its cheap wall wart PS).

Thus my surprise at the very positive result of the Catalyst 2960 above the direct connection.

And for $26 there is near-zero risk!  Do report back with your own impressions Tommy. 

I definitely will Alex.  What sold me was that you felt this switch sounded better than a direct connection between the server and microrendu.  I use a direct connection between  HQP and NAA pc.  Like you, I didn't feel the Netgear FS/GS made any difference.  My SG300 is being powered by a linear power supply and I use the SFP port to a fiber NIC in my server.

[Johnseye]

1 hour ago, Superdad said:

 

Well per @lmitche‘s urging I recently bought that exact model of Cisco Catalyst switch.  It is a heck of a piece—with way more in it than a $20 Netgear—not surprising since its original list price was probably $600+. ( @JohnSwenson thinks one of the big chips in it may be one  that he worked on at LSI/Avago—I need to send him close up photos to be sure.)

 

I’ve been running microRendu > ISO REGEN > modified SU-1 > Holo Spring L3 (the first 3 each with their own LPS-1.2) fed from a 25-foot BlueJeans/ Belden CAT6a—direct from my 4GHz iMac, no swtich, bridging trick.

 

So I put the Cisco on the floor under the audio rack, plugged the long Ethernet cable into port 1 and a 2-foot BJC CAT6 into port 8 to feed the microRendu.  Did not expect much since when I did this with a Netgear FS105 it was a wash or maybe worse than without.  

But whoa!  This was immediately better! More lively and musically engaging. Very nice.

 

The LiteOn SMPS in this switch is 12V/0.6A (if I recall correctly; 12V for sure, and while the supply might be 1A, Larry told me draw is 0.6A), so I could easily run it from another LPS-1.2.  But I’m busy, and since we are working on our own from-the-ground-up highly isolated/specially clocked switch (see EtherREGEN), I don’t know that I’ll bother at this point.

 

This certainly does bode very well for our own unit.  While I’ve be confident that EtherREGEN will technically perform as planned (certainly better than the modified switches being sold), I was not sure how much sonic impact it would have for the average user.  But what I heard with this Cisco switch was not subtle, and it beat the direct connection from the computer!  

Color me excited...

 

—Alex C.

 

Alex, you're referring to this model. 10/100?

 

WS-C2960-8TC-L

 

Have you or @lmitche tried a comparable gig Catalyst? 

[seeteeyou]

Someone was indeed getting gig Catalyst WS-C2960G-8TC-L and I didn't see any complaints

 

https://forums.naimaudio.com/topic/cisco-2960-again-recommended-settings

 

Both WS-C2960G-8TC-L and WS-C2960-8TC-L are sharing the same PSU model 341-0208-01 from Lite-On

 

https://supportforums.cisco.com/t5/lan-switching-and-routing/2960-port-issue/td-p/1647404

Quote

Motherboard assembly number     : 73-10613-08
Power supply part number        : 341-0208-01
Motherboard serial number       : FOC12123DXJ
Power supply serial number      : LIT12090BZC
Model revision number           : A0
Motherboard revision number     : B0
Model number                    : WS-C2960G-8TC-L

 

https://supportforums.cisco.com/t5/lan-switching-and-routing/high-latency-issue-in-lan-between-two-switches/td-p/2231639

Quote

Motherboard assembly number     : 73-10611-07
Power supply part number        : 341-0208-01
Motherboard serial number       : FOC140359C9
Power supply serial number      : LIT13510LWX
Model revision number           : C0
Motherboard revision number     : D0
Model number                    : WS-C2960-8TC-L

 

Former one draws 0.8 to 0.4 A while latter one draws 0.5 to 0.25 A respectively

 

https://www.cisco.com/c/en/us/td/docs/switches/lan/catalyst2960/hardware/installation/guide/2960_hg/higspec.html

 

BTW, those old caps inside that Lite-On 341-0208-01 might need a few days to get better according to Peder, and then a few people weren't 100% happy about their own Cisco because those caps could be lemons or something after all these years

 

https://forums.naimaudio.com/topic/cisco-router-to-match-cisco-ws-c2960-8tc-l?page=2

Quote

Many people on this forum say the same,better after 4-5 days.....most in the music flow area,the naturalness.Think Naim, when you put the power on again,even used equipment take at least 3 days for full performance.Exactly,just because the Cisco is used, you perhaps need a longer warm uptime for the caps.

 

I have had contact with over 50 people who has the Cisco 2960, only 3-4 are not 100% positive about it.Maybe your Cisco isn't "fresch", something is wrong with it.As you remember I say that the Company I buy my Cisco from,they say that some Cisco's failed in the 26-step testprogram.Maybe your Cisco need a recapp

 

In that case, maybe replacing the PSU with LT3045 should do the trick since I didn't see any kinda caps on the motherboard itself. Other than that, French Rooster actually paid more for a sealed 2960 and it seemed to make a huge difference to his setup sans FMC

 

https://forums.naimaudio.com/topic/cisco-switch?page=8

[austinpop]

5 hours ago, tboooe said:

Ok I'm sold.  Just picked up this switch from eBay for $26.  Very curious how this compares to my Cisco SG300-10 switch.  

 

Dang it - now I'm curious how this compares to my TLS OCXO switch...

[Johnseye]

19 minutes ago, austinpop said:

 

Dang it - now I'm curious how this compares to my TLS OCXO switch...

 

Exactly what I was thinking.  Time for a test.

 

It seems the gig switch is about $100 on ebay.  I think I'll try the 10/100.

[austinpop]

6 minutes ago, Johnseye said:

 

Exactly what I was thinking.  Time for a test.

 

It seems the gig switch is about $100 on ebay.  I think I'll try the 10/100.

 

Oh good - you try it first! My expectation is that the TLS switch will still sound better, but one never knows until one tries it.

[seeteeyou]

Only one gig Catalyst left from MN, 75 bucks shipped within US and seller seemed to have 100% positive feedback under the belt

 

https://www.ebay.com/itm/Cisco-WS-C2960G-8TC-L-2960-Series-8-Port-Gigabit-Switch-10-100-1000-CTC/391985472133

 

Fortunately it's only $60 a pop in China

 

https://item.taobao.com/item.htm?id=557845901834

https://item.taobao.com/item.htm?id=558899927588

https://item.taobao.com/item.htm?id=562616562000

https://item.taobao.com/item.htm?id=564834073722

https://item.taobao.com/item.htm?id=571566562569

 

That SFP port should be interesting because we could grab two 2960 just like @charlesphoto did, then they'll replace FMC while we're still taking advantage of what fiber optic cables could offer.

[austinpop]

On 6/16/2018 at 12:19 PM, flkin said:

Hi Rajiv, try the PoE. It's cheap and easy and doesn't need any DIY skills. 

 

Parts required:

 

https://www.amazon.com/dp/B01LX1WL2Z

https://www.amazon.com/gp/aw/d/B00QV1F0RU

 

Best of all, it's a clear step-up from my old star quad Canare with Oyaide connectors. 

 

Something I've been working on myself for a test, try connecting the unused ethernet ends of both ends of the PoE using a spare ethernet cable you may have around. You'd need female/female ethernet connectors like this to help :

 

https://www.amazon.com/PLUSPOE-Coupler-Ethernet-Inline-Connector/dp/B076C1F38Z/ref=mp_s_a_1_16?ie=UTF8&qid=1529169225&sr=8-16&pi=AC_SX236_SY340_QL65&keywords=ethernet+connector+female+to+female&dpPl=1&dpID=416s9a2v7rL&ref=plSrch

 

This will form a loop of the unused wires and should act as a pseudo-JSSG. I've found further improvements with this technique. Far easier than any DIY cable ? for me at least. 

 

Just a quick update on DC cables. I had not forgotten this experiment. It just so happened that the Amazon seller referenced above ships from China, so this took the scenic route to get to me!

 

No matter. I was now able to compare DC cables between LPS-1.2 and TLS OCXO switch:

1. My baseline: JSSG 360ed Audio Sensibility 7N UPOCC Silver
2. Ghent JSSG Canare 4S6 cable
3. PoE adapters with Ugreen Cat 7
4. PoE adapter with Supra Cat 8

Well, it’s really no contest. #1 sounded best, followed quite closely by #2. Both #3 & #4 were indistinguishable from each other, and lagged considerably behind 1 & 2. The biggest complaint about the PoE config was a distinct lack of bass. 

 

I’ll  let it burn in a few days and revisit.

[tboooe]

11 hours ago, austinpop said:

 

Oh good - you try it first! My expectation is that the TLS switch will still sound better, but one never knows until one tries it.

Curious why one would want to try the gigabit version of the Catalyst switch?  It seems that in general 10/100 hardware sounds better.