Mini-Review: Cisco Catalyst 2960 vs EtherREGEN vs EE8 Switches

OK this is where I am at … my good friends wanted to change their digital phone, this is another story suffice to say the only way to get it to work was to consolidate his ‘switches’ and plug the phone into the router. His phone works, but he now wants to Roonify his B&O.

I swapped out his 4 port BT hub and 4 port TP link for my 48 port NETGEAR switch as a temporary arrangement. He would like a 8 port switch. His Roon is on an IMac two floors above the living area that feeds a couple of TVs. His B&O is redundant until I find a feed into its SPDIF input (RPi?). I should not forget to mention that the property is listed, has thick stone walls, and uses an over the mains WiFi connection.
So cutting to the chase, I’m minded to go for a second hand 2960, but he wants something new. From reading this thread the EE8 looks like the solution, but I wonder whether I need a LPS?

I don’t think one would go wrong with either. There’s just no way to tell if an ‘industrial’ versus and ‘audiophile’ switch is better, worse or indifferent in terms of SQ. Functionality, it’s likely a horse a piece so long as one only needs the Cisco as a passive component. As for LPSs, that is a nice to have but not a must, obviously. The design of the SMPS that comes with the EE8 is said to be of high quality and works in tandem with it from a design perspective. The only way to know how much noise is associated with it is to measure it. For consistency I’ve gotten rid of any and every SMPS for audio/networking purposes and replaced with an LPS; it has also made it much simpler in my situation for isolation purposes as well.

1 Like

SWMBO has entered the equation… we wait and see

Potentially a very expensive exercise or experiment!

I have linear power supplies for both the Rendu and DAC (Mytek) in my second system and I deemed the improvements to be worthwhile. However, a swap to LPSs for both of my NAS devices, my Intel NUC (running Roon Rock) my router and my two Cisco 2960 switches would be very expensive and with no guarantee of improvement of any sort let alone a significant improvement.

I’m happy to live with the possibility that my set-up is (very) slightly suboptimal.

1 Like

How true. However, I’ve learned long ago that one doesn’t need expensive LPSs, just well-designed ones, which can be had on the cheap. It’s when one runs into ‘audiophile’ supplies that it can get spendy. Heck, even regulated Jameco or Enercell wall-warts would be a marked improvement. I used ham radio type linear supplies for my network for a while as well (like the Astron units—cheap as they come and well made). Your point is quite valid, though; but compared to what we spend on other tweaks time and again, a few hundred on LPSs doesn’t look so bad.

4 Likes

…or, to prove that there is more than one way to skin a cat…I have just had great success in replacing my LPSU supplied HiFi with a DAC/preamp/power amp run entirely from SMPSs, and it sounds bloody great.

3 Likes

I see Cisco is launching Cisco Catalyst 4 Port Micro Switches. They look to be more compact. They are offering a desk version, might be little easier to place around HIFi gear. I assume since they are Catalyst type they should work just as good as CC2960?

The original Catalyst 2960s are getting a little long in the tooth now, and are being replaced by newer technology such as with the Catalyst 9200s and their new ASIC technology and programmability capability which are quite different under the covers. (Think the equivalent of DSP FPGAs in audio).
I think you will see mini 2960 style switches available for a while, almost certainly at cheaper and cheaper price points… but slowly these older architected devices will disappear.

However technically the newer models will be at least as good with the very limited and basic functionality I suspect most use on their home networks, and due to the demanding EMC requirements they conform to, I suspect newer models will ‘sound’ at least as good from an interaction perspective.

But it is worth noting some of the main benefits of the Catalysts is the network traffic treatment… this happens by default… though you can turn off and modify.
Many of the consumer and so called audiophile switches steer well clear of this sort of network treatment capability.

The other consideration, that some forget, your lowest noise interaction will be with long Ethernet leads. Consider your streamer to switch lead being no shorter than 5m, preferably longer, coiling if necessary. Cat5e is also preferable to Cat 6 in this regard… though either can be shielded which may give a benefit.

1 Like

Thanks Simon, I have missed this point. To reduce noise, we should choose 5m+ lengths of Ethernet cable over much shorter lengths?
Could you explain the rationale? Apologies if you have done so many times before.

Best regards, BF

Sure, it’s probably been a couple of years since I last wrote about this on the forum; when Ethernet serial line switching power can couple into the connected electronics and cause undesired interactions, which is one reason we hear the effects of switch types and Ethernet leads changing the apparent sound from a connected audio device, we want to minimise the switching energy received in the connected device such as the streamer. Therefore we want to take advantage of the voltage drop in the Ethernet lead… the longer it is, the greater the voltage drop, and the smaller the energy seen by the received device.
Cat 5e uses smaller conductors than Cat 6, (typically) and therefore the voltage drop is greater with Cat5e than Cat 6 for a given distance.

In a related consideration, this is one reason why 100BaseT (that Naim use) is preferable to 1000BaseT. Less conductors are used for the serial line encoding… so less noise is created or received.

Of course if your connected device is completely decoupled from your audio device, then this is immaterial, but when DACs are closely integrated and exhibit a degree of coupling to network transports … as in Naim streamers, these considerations can become audibly apparent.

Hopefully that helps

8 Likes

Important to pay attention to specifications as requested by the manufacturer. In my case, dCS, the engineers are calling for UNSHIELDED Cat6. As long as the cable is certified (ie Blue Jeans) dCS engineers discourage audiophile ethernet cables.

Best
Gregg

2 Likes

I do find it strange how few of these audiophile cables, if any, are certified. If you’re trying to sell a premium product, supposedly better, wouldn’t that be the first thing you’d get?

I’m not a fan of these eyewateringly expensive cables, but to be fair, the amount of data they have to carry is very small, and I’m not sure they even need to conform to any Cat in order to perform their intended function.
Some early audiophile network cables used to be described as Ethernet cables by their manufacturers until it was pointed out that Ethernet is a defined standard to which they did not conform. Thus the term ‘streaming cable’ emerged.

1 Like

Gregg - other than for our uses at a network perspective Cat 5e and and Cat 6 are totally identical…so Cat 6 is interchangeable with Cat 5e - however Cat 5e has a greater loss due to its finer wires hence advantageous from a noise perspective for our applications in reducing noise in attached hifi devices, but both are specified for 1Gbps up to 100metres which is the ethernet standard, with Cat 6 being slightly more robust in high RFI commercial settings. But yes Cat 5e is getting slightly harder to get now - and Cat 6 is probably more ubiquitous. The audiophile market doesn’t really drive the network market :slight_smile:

It’s Cat 6A that has a meaningful difference - but again not relevant for our applications

……and yet, we see audiophile streaming cable brands with marketing blurb bragging about use of Cat7 and Cat8 cable.

indeed - what can i say… clearly aimed for the gullible or ill informed…

I have just checked my streamer interface - with Cat5e - which connects to a floating shielded Cat 6 right close by the streamer

 65763365 packets input, 5963770445 bytes, 0 no buffer
 Received 1340698 broadcasts (1291176 multicasts)
 0 runts, 0 giants, 0 throttles
 0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
 0 watchdog, 1291176 multicast, 0 pause input
 0 input packets with dribble condition detected
 285513911 packets output, 233581818807 bytes, 0 underruns
 0 output errors, 0 collisions, 1 interface resets
 0 babbles, 0 late collision, 0 deferred
 0 lost carrier, 0 no carrier, 0 PAUSE output
 0 output buffer failures, 0 output buffers swapped out

And interestingly with the prominently Cat5e on my home network of the 5963770445 bytes received back from my NDX2 - there has been not one single error at the physical link layer …
Kind of puts it into perspective doesn’t it …

2 Likes

Always good to know that you’re not being plagued by runts, giants, throttles and babbles :woozy_face:

2 Likes

For installation I always recommend CAT 6 as it’s specced for 10Gbps up to 55m. Most domestic runs will be shorter, and even if domestic equipment nowadays is unlikely to support it, I expect in time it will.

Of course when buying a patch cable that’s more easily replaced later it doesn’t matter.

Reading @Simon-in-Suffolk responses reminds me of these articles from Ars Technica a while ago.

2 Likes

Interesting articles, especially the latter one.