Ifi LAN iPurifier Pro

[Benje]

I acquired one of these (ifi LAN iPurifier PRO) from Addicted To Audio and have used it at home for my ethernet connection from an Ubiquiti Edge 24 switch to my Apple TV 4K (2nd Gen).

 

There are various reports on the performance of these units depending on which forum you choose, and whether you are a 'listener' or a 'science believer'.  I can't give any concrete advice on whether the unit is good or otherwise, make your own decision.  The unit is meant to provide galvanic isolation of the ethernet (which is not a bad thing, could be very good) and some filters (can't comment on that, beyond my knowledge/experience).  I am happy to keep this unit in my system at this stage with one proviso.

The unit can be very hot to touch.  VERY HOT.  However, if you turn off the LED status lights, the unit becomes warm only.  Use the button on the rear and push twice to dim then turn off the display lights, and the unit heat drops back to something reasonable.

 

At some stage, I may come to an opinion on what this unit does with my system for audio and for video (principally via the Apple TV, but possibly also via a SONOS Connect).  The unit is between the Ubiquiti Edge 24 switch and the Apple TV 4K (or Sonos Connect).

 

Benje

[Volunteer]

FYI, galvanic isolation of Ethernet is already inherent in the design of the standard.

[Benje]

10 hours ago, El Tel said:

FYI, galvanic isolation of Ethernet is already inherent in the design of the standard.

Thanks El Tel

 

can you elaborate?  Which standard?  thanks

 

Benje

[Volunteer]

13 hours ago, Benje said:

Thanks El Tel

 

can you elaborate?  Which standard?  thanks

 

Benje

 

That describes how galvanic isolation is achieved through the section on transformer coupling:

https://en.m.wikipedia.org/wiki/Galvanic_isolation

 

The IEEE 802.3 Ethernet standard specifies use of transformer coupling in Ethernet interfaces primarily for galvanic isolation and signal conditioning. Remember, this is on every PHY (Physical Layer Device) active (powered) port that an RJ45 connector plugs into (switch, router, PC, NAS, streamer etc).

 

Transformers are used to isolate the PHY from the transmission medium (the cable), protecting devices from electrical faults which include voltage surges and other electrical artefacts like EMI-imduced noise that some concern themselves with.

 

You can get lost for days in the IEEE Ethernet standards documentation, but the answers are all there if it interests you to go further.

[MattyW]

Yes the iFi LAN Purifier Pro caught my eye for a bit as the price is more reasonable than Network Acoustics products, even before they lifted their prices recently. It's a worthy consideration.

 

I mean I'm a fan of NA product insofar as the sonic results but the pricing is just too much for me these days.

 

I ended up with a Stack Audio Smooth Lan ethernet filter. I'm yet to compare it with my NA Muon Pro or Eno Ag filters though will get there eventually.

Edited February 16 by MattyW

[Volunteer]

On 14/02/2025 at 8:08 PM, Benje said:

and some filters (can't comment on that, beyond my knowledge/experience)

 

The thing to keep front-of-mind with claims of filtering is that there is no end-to-end electrical continuity in an Ethernet network (that's what transformer coupling does - it invokes electromagnetic induction across the windings of the transformers, without direct conductive connection, in order to send the signal). 

 

The design standard also includes the mechanism of regenerating the packet at every transmission (when the packets exit a participating node in the chain between source and destination). So what exactly is being filtered and what benefit does it have when the packet is regenerated fresh at each hop as the last action of the transmitting node?

 

If this is happening as part of the standard, then what is any other filter doing to mitigate or improve anything else? That's a question worthy of anyone's consideration before investing in anything in this esoteric backwater of the audio vista.

 

Notwithstanding, I always say that you should do you, after all, nobody can do you better than you. If anyone wants to test and then perceives a tickle of their fancy, knock yourself out

 

 

[frankn]

On 16/2/2025 at 3:38 PM, El Tel said:

 

 

The design standard also includes the mechanism of regenerating the packet at every transmission (when the packets exit a participating node in the chain between source and destination). So what exactly is being filtered and what benefit does it have when the packet is regenerated fresh at each hop as the last action of the transmitting node?

 

 

@El Tel when you say “regeneration” do you mean error checking protocols / retransmission comments I picked up from Wikipedia?

 

Systems communicating over Ethernet divide a stream of data into shorter pieces called frames. Each frame contains source and destination addresses, and error-checking data so that damaged frames can be detected and discarded; most often, higher-layer protocols trigger retransmission of lost frames. Per the OSI model, Ethernet provides services up to and including the data link layer.[3] The 48-bit MAC address was adopted by other IEEE 802networking standards, including IEEE 802.11 (Wi-Fi), as well as by FDDI. EtherType values are also used in Subnetwork Access Protocol (SNAP) headers.

[Volunteer]

1 hour ago, frankn said:

@El Tel when you say “regeneration” do you mean error checking protocols / retransmission comments I picked up from Wikipedia?

 

Systems communicating over Ethernet divide a stream of data into shorter pieces called frames. Each frame contains source and destination addresses, and error-checking data so that damaged frames can be detected and discarded; most often, higher-layer protocols trigger retransmission of lost frames. Per the OSI model, Ethernet provides services up to and including the data link layer.[3] The 48-bit MAC address was adopted by other IEEE 802networking standards, including IEEE 802.11 (Wi-Fi), as well as by FDDI. EtherType values are also used in Subnetwork Access Protocol (SNAP) headers.

 

I use the words packet and frame interchangeably at times. My bad. There are subtle differences (packets technically are Layer 1 on the OSI, and frames are Layer 2), so to try not to obfuscate, I'll do my best to stick with the correct terminology here, which is frame.

 

Regeneration relates to the fact a frame is not passed along the chain per se. They are generated as new frames at every transmission node along the way.

 

This is probably the simplest summary:

• The switch receives the complete Ethernet frame on one of its input ports and processes it (checksum validation etc).
• The switch examines the frame's destination MAC address in the header.
• It consults its MAC address table to find the correct output port towards the destination MAC address.
• The switch creates a new Ethernet frame with the same data payload as the original, during which it updates the information for the new frame's header with source MAC address now set to the MAC address of its own outgoing switch port (rather than that of the previous sending source/node MAC address).
• The destination MAC address remains the same if it's the final destination (or is set to the next hop's MAC address if routing is involved - in other words a different IP subnet requiring routing off the existing subnet at Layer 3 (IP) of the OSI model).
• A new Frame Check Sequence (FCS) is calculated and appended to ensure data integrity.
• The newly created frame is then transmitted out through the appropriate output port to the next network node or final destination.

 

One can deduce from this that concerns like noise, or compounded voltage drift are dealt with due to the transformer-coupling and creating a new frame at each onward transmission (and transmitting that new frame rather than the same original frame through continuous electrical signal from the original source to the ultimate destination).

 

This is a complex subject, so the best thing to enable people to pick it up if the concepts are new to them is to use a little poetic licence in the simplification of the explanation, it seems like you grasp it just fine, but in case anyone is uncertain maybe think of it as a relay race of some sort. The runners are the Ethernet devices along the way (source device/switches/routers/destination device) and the bottles are the frames that get handed to the next runner through the cable

 

• The source node is the first runner, they write a message, put it in a bottle, and start running (transmitting).
• When reaching the next runner, they don't simply hand over the bottle. 
• Instead, the new runner receives the bottle, reads the message, writes it on a fresh piece of paper, places it in a new bottle and then starts running (transmitting) with the new bottle
• This process repeats for all runners (nodes) until the last runner (destination) receives the message.
• At the finish line, the message in the final bottle is identical to the original, even though it's not the same physical bottle or paper.

 

I apologise if any oversimplification has been inferred as condescending. I am trying to help as many people to follow these processes as easily as possible. I hope people see this as a genuine attempt to democratise the knowledge and not to patronise.

 

Hope this clarifies.

 

[Addicted to music]

There’s plenty of information if you google it,  if you want an explanation on how it’s done search for the ISO 7 layer for networking.

[rand129678]

On 15/2/2025 at 7:11 AM, El Tel said:

FYI, galvanic isolation of Ethernet is already inherent in the design of the standard.

Only if you use unshielded UTP cable. A lot of audiophile ethernet cables are actually shielded, which undoes galvanic isolation.

 

[bob_m_54]

12 hours ago, rand129678 said:

Only if you use unshielded UTP cable. A lot of audiophile ethernet cables are actually shielded, which undoes galvanic isolation.

 

But that is isolation between the chassis' of the components, and may cause hum etc. The data signal will still have galvanic isolation between components. And of course, if the shield is only grounded at one end, as any data cable shielding should be, then there won't be any chassis to chassis connection either.

[Assisi]

On 18/02/2025 at 5:05 PM, El Tel said:

I apologise if any oversimplification has been inferred as condescending. I am trying to help as many people to follow these processes as easily as possible. I hope people see this as a genuine attempt to democratise the knowledge and not to patronise.

 

Hope this clarifies.

 

I appreciate that your response was answering  a question.   My understanding of the complexity of Networking and Ethernet is limited.  I do not at all doubt the integrity of the data at any stage in the network.  At least for me there are no integrity issues.  Everything gets through to the DAC always intact.  Your explanation of the steps involved with the transmission of the signal does not clarify for me my lived experience of the benefits of audio switches for my listening experience.  The reduced noise floor. To me this is the Holy Grail of the audio outcomes in the digital context.

 

@bob_m_54      in another thread when I mentioned a reduced noise floor, indicated that he seemed to be confused about the topic of lower noise floor.  He is not the only person who does not appreciate a reduced noise floor and the listening benefits. 

 

I knock my self out with pleasure every time I turn my system on.

 

John 

[rand129678]

38 minutes ago, bob_m_54 said:

But that is isolation between the chassis' of the components, and may cause hum etc.

Yes ground loop is what you want to avoid.

[bob_m_54]

12 minutes ago, Assisi said:

@bob_m_54      in another thread when I mentioned a reduced noise floor, indicated that he seemed to be confused about the topic of lower noise floor.  He is not the only person who does not appreciate a reduced noise floor and the listening benefits. 

Any noise floor that is transmitted via the data cable, is the noise floor from the original recording, that has been digitised and is "legitimate" data and part of your music file. It isn't added to nor subtracted from the data, by using any specialised equipment. In fact if it is, then that means your data hasn't been transmitted faithfully.

[Assisi]

51 minutes ago, bob_m_54 said:

Any noise floor that is transmitted via the data cable, is the noise floor from the original recording, that has been digitised and is "legitimate" data and part of your music file. It isn't added to nor subtracted from the data, by using any specialised equipment. In fact if it is, then that means your data hasn't been transmitted faithfully.

Whilst there may be noise in some way in the original recording it is in no way the noise that I am referring to.  If there is noise in the recording nothing that I do will remove it.  Some tracks that I do listen to do have noise and I hear it. I have a streaming play list. When or if I do something and I perceive a subtle reduction in the noise floor there can be a listening benefits.  There is no change in the original recording that I may have played years ago. The digital signal information is exact.

 

The noise I am referring to is generated by resonance, interference or vibrations in network components, power etc.  The purpose of a network filter is to reduce the noise.  The noise can travel together with the digital through the network. It can end up in the analogue domain. When you first experience less of the noise I am referring to you will wonder what has happened and why.  I know I did.  I did not understand until some body provided the explanation.  My first experience was from isolation feet under a DAC.

 

You are not the first and will not be the last to query my reference to noise in relation to Ethernet and networking.  One day hopefully you will have a "Hallelujah Moment".  Your listening will be changed forever.  Be warned, the pursuit of a lower noise floor can be addictive and expensive. 

 

John

[Volunteer]

48 minutes ago, Assisi said:

Your explanation of the steps involved with the transmission of the signal does not clarify for me my lived experience of the benefits of audio switches for my listening experience.  

 

I was not clarifying or questioning individual perceptions. I was explaining galvanic isolation and added in clarifications to demystify common misunderstandings on noise propagation and phenomena such as voltage drift which can not be passed with the digital signal through Ethernet interfaces (and the network as a whole) due to it comprising non-contiguous connections (i.e. no electrical continuity from end-to-end due to galvanic isolation). 

 

As a stated benefit (and an assumed additional benefit over normal Ethernet connectivity) of the filter in OPs first post, it was important to reflect that the filter does not provide anything around galvanic isolation that is not already inherent in the IEEE 802.3 Ethernet Design Standards and associated control mechanisms.

 

Your lived experience is not undermined by actual technical explanations of how things actually work. 

 

As I said earlier:

On 16/02/2025 at 4:08 PM, El Tel said:

Notwithstanding, I always say that you should do you, after all, nobody can do you better than you. If anyone wants to test and then perceives a tickle of their fancy, knock yourself out

 

Instead, perhaps consider that the distillation of these standards into easy-to-follow explanations are not directed at anyone or any group, but are intended to aid in untangling any unintended mystery that pervades. If one claims they hear a difference and then states they believe it is due to x or y, and x or y can be articulated properly, then everyone including the claimant can understand whether the reasoning for the difference holds-up to functional scrutiny in the light of a simplified technical description of what x and y actually do.

 

Would it not be a great outcome if, through process-of-elimination, we can arrive at a place where we can identify and agree what can and cannot cause the different experiences and not continually be grasping at straws, be it cables, switches, filters or just the human experience? This is, of course, rhetorical; I don't intend, or need, to defend a simplified, but demonstrably accurate description of how Ethernet works.

 

One can muse, cogitate, reflect and postulate all they like with regards to the cause of any difference they may experience/perceive in changing equipment in or out of any system, but do not be surprised on a public forum, with a multitude of experienced, professional and/or qualified people at our collective disposal, if someone offers a technical clarification to allow people to understand more and to subsequently make their own informed decisions.

 

None of the above impinges on your personal experiences and I accept them as your experience without argument. 

 

My statements above are not personal experiences, they are just a simplified description of technical information that is available to all if they care to look.

 

Time to disengage and move on, methinks.

[MattyW]

I do wish however that the specification could explain why differences are in fact audible to any number of us. I've simply accepted long ago that some things defy explanation and everything matters, even though it shouldn't. I understand completely why there is so much disagreement on the subject. Especially given the technical standards define how it works and their really aught to be no difference.  I believe it to just be one of those areas where online discussion is of limited benefit to anyone involved. Such is life eh?

[Volunteer]

22 minutes ago, MattyW said:

I do wish however that the specification could explain why differences are in fact audible to any number of us. I've simply accepted long ago that some things defy explanation and everything matters, even though it shouldn't. I understand completely why there is so much disagreement on the subject. Especially given the technical standards define how it works and their really aught to be no difference.  I believe it to just be one of those areas where online discussion is of limited benefit to anyone involved. Such is life eh?

 

The whole piece is obfuscated by other factors without. When I say without, I mean outside of the standard's design, recommended implementation of the specifications and intended usage applications, as an example, I refer to using not-fit-for-purpose or even out-of-spec network cables (think shielded patch cables, or data centre specific CAT8 usage or even the proprietary CAT7 spec for example) and introduction of grounding where there was none to begin with. That kind of thing just bypasses the protections afforded by galvanic isolation as per the OP post.

 

Anyway, there are plenty of threads already that talk about experiences. The OP spoke of the stated advantages of a product as per the product documentation and marketing materials, it's important to understand what is and is not possible in such light.

 

My tired disclaimer notwithstanding - do whatever makes you feel warm and fuzzy - but don't get offended if an explanation of the actual mechanisms at play is offered when one starts hypothesising a cause for the differences perceived.

[rantan]

@El Tel

 

Just to add to a previous line to the one in your sig..

 

"sadly they left, telling no one goodbye, Pressed Rat wore red Jodhpurs, the Warthog, a striped tie.

 

It sounds so much better if you can copy Jack's accent and it is a sad tale to be sure.

[bob_m_54]

6 hours ago, Assisi said:

Whilst there may be noise in some way in the original recording it is in no way the noise that I am referring to.  If there is noise in the recording nothing that I do will remove it.  Some tracks that I do listen to do have noise and I hear it. I have a streaming play list. When or if I do something and I perceive a subtle reduction in the noise floor there can be a listening benefits.  There is no change in the original recording that I may have played years ago. The digital signal information is exact.

 

The noise I am referring to is generated by resonance, interference or vibrations in network components, power etc.  The purpose of a network filter is to reduce the noise.  The noise can travel together with the digital through the network. It can end up in the analogue domain. When you first experience less of the noise I am referring to you will wonder what has happened and why.  I know I did.  I did not understand until some body provided the explanation.  My first experience was from isolation feet under a DAC.

 

You are not the first and will not be the last to query my reference to noise in relation to Ethernet and networking.  One day hopefully you will have a "Hallelujah Moment".  Your listening will be changed forever.  Be warned, the pursuit of a lower noise floor can be addictive and expensive. 

 

John

Well I don't have answers for your experience. But if you are hearing differences, then that's a good thing for you.

 

6 hours ago, Assisi said:

The noise I am referring to is generated by resonance, interference or vibrations in network components, power etc.  The purpose of a network filter is to reduce the noise.  The noise can travel together with the digital through the network. It can end up in the analogue domain. When you first experience less of the noise I am referring to you will wonder what has happened and why.  I know I did.  I did not understand until some body provided the explanation.  My first experience was from isolation feet under a DAC.

 

If this were possible, it would also be possible to measure it, and display the results. Have you seen any credible information/evidence, anywhere on the web, to support this? As @El Tel has mentioned, there isn't much point in theorising about the cause of what you experience, we all hear things differently, but numbers look the same to all of us.

Edited February 26 by bob_m_54
error

[Assisi]

1 hour ago, bob_m_54 said:

Well I don't have answers for your experience. But if you are hearing differences, then that's a good thing for you.

 

 

If this were possible, it would also be possible to measure it, and display the results. Have you seen any credible information/evidence, anywhere on the web, to support this? As @El Tel has mentioned, there isn't much point in theorising about the cause of what you experience, we all hear things differently, but numbers look the same to all of us.

As a poster many years ago corrected me it is not about difference.  It is about whether there is a benefit. As I said above once you realise when the noise floor is reduced you have found the Holy Grail of the audio outcomes in the digital context.  The benefit is that the listening pleasure is enhanced.  The purity of just the recording.

 

You and many others seek measurements and tests.  I often think that is just and excuse for no experience.  Just give me the numbers and the graphs.  I can understand with what I think is your avionics and instruments background that the numbers may be important for you.  That is what you are comfortable with.  All of us and I mean all of us are afflicted with the confirmation bias.  It is about what we individually understand.  We all reject what we do not understand or appreciate.  It is about being human.    As somebody once said “I do not listen to an oscilloscope.” I have no idea as to how to undertake measurements even though I have all the gear to do so.  For me it is all about just listening and working out what works with what.  Not the theory.  Am I right? Who knows?

 

I am very aware of  @El Tel's  position on the topic.  More so than you would be aware.  His technical knowledge  probably exceeds that of any other poster on this forum.   I am focused on my lived experience.  That is what is real to me.  I am certain that one day and maybe sooner rather than later that the matter of Low Noise Floor and the listening benefits in the digital context will be an accepted norm.  My lived experience is real.  I know I am not delusional with my perspective or that my experience is a placebo outcome. 

 

An important question is what is your actual experience with digital and streaming?  Is it just theory?

John

[andyr]

25 minutes ago, Assisi said:

An important question is what is your actual experience with digital and streaming?  Is it just theory?

 

Oooohh, maaate - you're swimming in dangerous waters!  

 

[Volunteer]

This thread is not about subjective experience at large. OP mentioned a product and explicitly called-out the manufacturer's documented claims.

 

The claims mentioned were shown to be nothing more than the pre-existing design safeguards inherent in the medium in question.

 

Following a question seeking clarity on frame regeneration, summaries of related mechanisms for the Ethernet standard were provided as extracted from the IEEE technical libraries.

 

Let's not derail this thread with any further subjective inputs; questions and clarification on technical operation of the medium notwithstanding. Nobody is denying anyone else's lived experiences with networking, but there are numerous threads on subjectivity already and very little in terms of concise, simplified and accurate information surrounding the actual operation of Ethernet. Such summaries are vital to allow any inexperienced parties the opportunity to assimilate and assess all objective and subjective inputs to empower them to make their own informed decisions and to experiment for themselves.

[bob_m_54]

13 hours ago, Assisi said:

You and many others seek measurements and tests.  I often think that is just and excuse for no experience.

It's because I know my ears are not even close to the accuracy of calibrated test equipment. And that test equipment can be relied upon to perform the tests with the same accuracy every time, where my ears will give different results, depending on the day and what  noise they have been subjected to during the day. And that's not even talking about the emotional side of being human, and how it affects us.

 

13 hours ago, Assisi said:

An important question is what is your actual experience with digital and streaming?  Is it just theory?

 

I have been streaming digital music that I have ripped from CDs as a source since the mid 90s, when 192Kbps mp3 was considered to be "extreme" quality. My experience in digital signals and transmission methods goes back to the 80s, and in a much more critical environment than moving digital music data between boxes in a home environment.

 

I think that many people that don't see the point of measurements, and adhere to the mantra "I don't need measurements, I just use my ears" just lack understanding in the basic principles of how this technology actually works. And that it may actually show how fallible "just using your ears" can be.

 

But.. I'm not trying to preach to the il-informed, just trying to clarify some misconceptions. But if someone is attempting to advise a new user, that has no experience, always remember to add the caveat "to my ears" when suggesting an "improvement". Never say "this will improve audio" unless you have some legitimate data to back up your claims..

 

Anyway, at the end of the day, just do what you're comfortable with, and what makes you happy.

Edited February 26 by bob_m_54
further info

[Volunteer]

Perhaps it may be of benefit for a purely technical thread to exist that can provide descriptions and answer questions on the subject?

 

It might give experienced people uncomfortable with sharing knowledge in hostile threads the opportunity to do so under the cover afforded by a specialised technical topic.

 

It could also allow subjective threads receiving existential questioning the opportunity to redirect the questions or statements they cannot answer to that technical thread.

 

hmmm.