Is the P5 network socket galvanically isolated

I go to significant lengths to achieve galvanic isolation of my DirectStream Junior DAC, analogue crossover and power amp. I use only a TosLink to feed digital audio into the DS Jnr and have a length of fibre-optic in the network path to the Bridge in it. (The fibre-media converter (FMC) that turns the Ethernet back into electrical form is very close to the DS Jnr and has a very high quality isolating DC power supply by Uptone Audio.)

Now I have a PowerPlant P5 and I’m wondering what to do about its network connection.

• I could interpose another length of fibre-optic with another pair of FMCs but that will need another, very expensive DC power supply for the FMC closest to the P5.
• I could do without the network connection to the P5.
• If I knew that the P5’s network connection is galvanically isolated (there’s no mention in the specifications) I could just plug in a network patch cable and stop thinking about it.

I welcome comments.

On my P10 i can hear difference when LAN is connected so unfortunately it seems it’s not galvaincaly isolated.
Thus during my critical listening i’m unplugging the P10 from LAN as i didn’t want to bring additional complexity to the chain with another PSU and FMCs (similarly as you i have streamer connected to DAC on FMCs with LPS…

It’s really pity as it defeats some of my home automation when powerplant is not on network…

So I’m not the only one! :crazy_face:

Yes, I would prefer to have my P5 on the network. I have a class A amp and like to be able to switch it off remotely if I’ve forgotten to.

Look into this, might be just the ticket for you:

PinkFaun LAN Isolator (just search for it using you search engine of choice)

Not too expensive and It won’t mean adding a box and power supply to your signal chain…


1 Like

thx @scotte1 ! just ordered two of them - one will replace FMC and other will make p10 to stay connected to LAN all the time - happy! :slight_smile:

1 Like

Let us all know if your ears tell you there is an improvement.


Ethernet is inherently galvanically isolated by design. This doesn’t mean that noise can’t be introduced another way though.


Thanks Scotte1.

Let us know if you try 'em and like 'em.


I didn’t know that. So Pink Faun’s LAN isolator is snake oil?

I’d been under the impression that to be galvanically isolated there had to be no electrically conducting elements between the two components to be isolated from each other. I can see from the structure of RJ45 plugs and sockets that there are conducting elements from one RJ45 socket to the other (the sockets’ connectors, the plugs’ connectors and the cable itself). So if Ethernet is galvanically isolated by design then all of those RJ45 sockets must have transformers or optical isolators or something else I’ve not heard of behind them. I’m surprised.

Paul McGowan’s blog post of 11 August dealt with galvanic isolation and he said what I already thought I knew. But perhaps I’m missing something.

I think @jamesh misspoke, at least in this context. I hope he jumps back in to affirm or clarify his statement.

My understanding is that noise that may be carried along with the digits via Ethernet can be “scrubbed” from the signal path using “galvanic isolation”.


Ethernet is transformer-coupled at both ends. This is galvanic isolation but it’s not total. There are differing grades of Ethernet ports. Some use fewer transformers better ones use more and higher grade transformers. Some (frequencies) noise is blocked by transformers other noise is not.

Of course, ideally you wouldn’t connect the P5 to your network but if the convenience out weighs the downside…but I would make sure all other connections, Ethernet and other that carry audio signal and digital bits are in good shape before being too concerned with the P5 connection.

1 Like

Thanks for this^…

To quote from the site you linked:

Ethernet offers galvanic isolation by design. All connections are transformed [sic] coupled. However this does not guarantee that no noise at all might creep into a networked device.

This matches my previous research and understanding.


James is correct in that the ethernet PHY is transformer coupled to the RJ45 cable via magnetics and there’s typically no physical connection other than through a magnetic field, technically galvanically isolated.

Paul and what about ground/shielding? That one would be galvanic i’d assume?

I can’t really speak upon that. I’m no network cabling expert, but brett66 makes a really good point about some ports being higher quality with better transformers. Possibly what they’re doing here.

1 Like

One test is to try multiple types of Ethernet cables. If the sound changes a lot between them all, isolation is not optimal.

What is frustrating is I’ve found no consistency in sound signatures!

Sound signature is a combination of the switch/router, the power supply, and the cable. The same cable may sound good with Switch A while awful with Switch B.

1 Like

oh boy, i shouldn’t red this, now plethora of new combinations to test opens :smiley:


using unshielded ethernet reduces one path for ground noise to be transferred i.e. use UTP cables.
if not in a very electrically noisy environment (industrial, for example) and not using ethernet close to its max length, then shielded is rarely necessary.


Good point. No need to make things worse with shielded Ethernet. In the home environment it will only serve as a path for noise.

1 Like