I've bought an ifi Audio iPower SMPS for my Netgear GS105v4 switch.
With its active noise cancellation & 12 element output array its claimed the iPower SMPS has an exceptionally low noise floor @ ~1uV, compared to a low noise linear power supply with ~20uV & a consumer grade SMPS that's typically 1000uV.
My present LAN set up is a bit OTT w.r.t. to suppressing RFI, switching & other random electrical noises; I've tried to reduce it as much as possible with screened cables, ferrite & correct LAN grounding.
The power to the LAN switch SMPS is via an APC 350 UPS that includes an isolation transformer & common mode choke. Its plugged into a four way power strip together phone & broadband SMPS's & a Russ Andrews Silencer, the short power strip cable has a ferrite clamp on each end. I have assumed the UPS isolation transformer & choke, plus the clamp & ferrite should go some way to reduce the power supply switching noise feeding back into mains power. This new SMPS is intended to reduce noise on DC power to the Netgear GS105v4 switch.
Measuring the iPower in comparison to the Netgear SMPS showed the Netgear voltage varied & the iPower was very stable to >0.01VDC. I also used my little hand held 'scope & that was not able to show much more than the voltmeter other than see a mV-AC element, but am not sure what that is showing other than question the iFi claim of better than 1uV noise floor.
I first measured VDC & Hz with no load & then loaded the output with a 22ohm resistor, that's 6.5watt & 63% more than the Netgear switch max.
iPower
No Load – 12.22VDC 6 .6mVAC 0Hz
6.5w Load - 12.18VDC 6.6mVAC 0Hz
Netgear
No Load - 12.14/12.16VDC 9.9/15.3mVAC Hz started as 50Hz & settled at 0Hz
6.5w Load - 11.70/11.78VDC 10.1/18.2mVAC 0Hz
OK that's the numbers, what about the SQ.
First impression is something has changed, not at all sure what, its subtle & not easy to define other than its different. It will probably need a few hours concentrated listening, so I will do that this evening & post again tomorrow.
Posted on: 16 January 2016 by Mike-B
Hi Dave, taking it from the top: A SMPS can inject noise back into the mains, not all do this, most do to some extent, some can be particularly bad & wall warts seem to have gained the "bad" reputation, maybe unjustified & something from the early days & the pooh has stuck. To balance this up, your TV & other such devices will invariably have them & all the Linn units have them these days, so they are not all bad.
Very few people are bothered with them as they are not aware of the effects, hifi peeps seem to be more aware but again many are not that bothered & I doubt many can actually show what the noise is actually doing. The big complainers are the radio hams, they can both hear it & their 'scopes can see it.
There are regulations for SMPS noise suppression, but as I understand it, they limit noise rather than eliminate. The noise is from the switching process, very rapid on/off/on & usually from 50kHz to 1mHz.
This experiment is aimed at eliminating switching noise on my LAN (NDX-switch-NAS). I was aware that as good as my NDX SQ is, playing from a USB was slightly better than NAS when I removed the ethernet from the NDX, I then found the same effect with the ethernet still connected but the LAN switch powered off. So I assumed this small SQ change was caused by the SMPS noise & looked around for something to fix'um. I could not easily install a linear unit as they are quite large, this iPower unit is around the same size as the Netgear SMPS so worth the £40 to experiment. It passed the USB -v- NAS comparison test, so its a keeper.
Regarding plugging it into the same power strip as the amps., I can't say other than what the iFi company claim, so I would expect it to not cause a problem. My SMPS is powered via my UPS & that has an isolation transformer so in theory its isolated from the mains, however it is possible the noise could carry thru the transformer.
Re the need for screened cables, I'm not sure about that, it might be the screen carries the noise (but mine was correctly grounded so it shouldn't be the case) Whatever I've changed my screened Cat7 ethernets for non-screened Cat6 ........... more on that later.
Posted on: 16 January 2016 by nigelb
I've been listening to the iFi iPower and in summary at £40 it is one on the best VFM upgrades I have made.
Like Mike sayes the improvements are subtle but they are unmistakable and very worthwhile. I will try to detail what improvements I have heard.
Firstly there is a little more space and air around instruments and lead instruments/vocals don't over-dominate which was (is now) apparent previously. There is a little better balance with backing/supporting instruments/vocals more evident. In some ways, music is presented in a slightly more relaxed manner on some tracks, in some ways less impressive but better able to draw you in and therefore all the more impressive if that makes any sense. Vocals are a tad cleaner with slightly better articulation and hence lyrics a tad more comprehensible. Bass is slightly firmer on some tracks and easier to follow when the music gets complex. One thing the iPower couldn't do was tame extreme sibilance on one particular test track but I think this was poorly recorded in the first place so no hope I would say. I am still listening to the iPower as I type and am hearing again that better balance of instruments I have already mentioned. Quite impressive IMHO.
I say again these are not night and day difference with the iPower, subtle at least, but improvements are noticeable and worth having.
You can't hear 'noise' when it invades your network but you sure can hear the improvement when said noise is reduced.
Best £40 I have spent for sure.
Posted on: 16 January 2016 by Mike-B
Great report Nigel, you have heard it just like myself - as you say its not night & day, but maybe a bit more than subtle once your ears are tuned in.
The new Meicord ethernets arrived today & that's added some more, including the confusion factor as now I'm not sure if the change is the iPower or Meicord.
Posted on: 16 January 2016 by nigelb
Mike - Thanks for the tip about the iFi iPower, great value upgrade IMHO.
I would be interested to hear more about your ethernet cable changes. Why you changed (other than the broken clips), what the new cables have given in terms of SQ and the pros and cons of screened vs non-screened Ethernet cables.
Still think you should have gone for AudioQuest Vodka though. 
Posted on: 16 January 2016 by Mike-B
Nigel, I changed the ethernet because of broken plug clips plus the cable is a bit stiff & physically unforgiving & IMO it would not get any better even with a new plugs. It was bite the bullet time.
I'm not convinced ethernet has much effect on SQ, so TBH I was not expecting much. The renderer plays from the buffer not the cable & each packet is received & verified as intact, or re-transmitted. It may be that cable/plug termination quality, twisted pair deformation etc. can result in signal transfer issues, maybe noise is added & (maybe) could be audible.
Anyhow, to the Meicord Opal's: They are thinner than the old Supra Cat7, but surprisingly stiff compared to another Cat6 I have. The plugs are excellent quality & the clips appear a lot more robust.
The change in sound was surprising as I really wasn't expecting too much if anything. The problem is the SQ changes with the iPower & now the Meicord's are a bit blurred. However I pretty certain the Meicord's have bought a change for the better with clarity, definition & dynamics. Anyhow, about to start another listening session. Maybe some more later
Posted on: 16 January 2016 by nigelb
Mike - look forward to your further observations on Ethernet cables.
Back to the iPower. I have been so impressed with it considering its low cost, I was wondering if there could also be benefits using it on other network devices using SMPSs. I have a Unitiserve and already have....ahem......a pukka linear power supply on it which I must say had moved the SQ on markedly. However I note that my BT Home Hub 5 router had a crappy SMPS on it. Do you think there are any technical reasons why a better SMPS (like the iPower) might bring SQ benefits on a router? Can a router with an average SMPS introduce noise into a network?
If Simon in Suffolk is watching (and others with experience here) I would be interested in your thoughts too.
Sorry to bug you but I really think you are on to something by reducing noise from network device SMPSs without resorting to linear power supplies.
Posted on: 16 January 2016 by Dave***t
Hi Dave, taking it from the top: [snip]
Thanks for the generous response, Mike. And thanks to both you and Nigel for sharing conclusions. I think I'll get a USB stick out and try a few things out myself. Be interesting to see if anything else crops up in this thread.
I confess that the issue of whether the cable shields carry noise and how to ensure they're properly grounded is something I'm similarly unclear on. But I think I remember an attempt (by Huge, perhaps?) to put together some kind of guide. Will have a search.
I think this kind of stuff is something a lot of us would probably do well to pay more attention to, even (or especially) if space/practicality means that we can't always ensure optimal cable dressing or plug siting etc. Potential free upgrades, effectively. The best kind 
Posted on: 16 January 2016 by Mike-B
I was wondering if there could also be benefits using it on other network devices using SMPSs. .............. I note that my BT Home Hub 5 router had a crappy SMPS on it. Do you think there are any technical reasons why a better SMPS (like the iPower) might bring SQ benefits on a router? Can a router with an average SMPS introduce noise into a network?
The BT SMPS is labelled 12V 1500mA. = 18W. I've found some www forum info that says HH5 idles ~5.4W.
The iPower is rated at 12V 1.1A = 13.2W so it should be OK
I'm not sure it will bring any changes to your system SQ unless (maybe) its plugged into the same & close to the Naim power circuit
Posted on: 16 January 2016 by nigelb
Hi Dave, taking it from the top: [snip]
Thanks for the generous response, Mike. And thanks to both you and Nigel for sharing conclusions. I think I'll get a USB stick out and try a few things out myself. Be interesting to see if anything else crops up in this thread.
I confess that the issue of whether the cable shields carry noise and how to ensure they're properly grounded is something I'm similarly unclear on. But I think I remember an attempt (by Huge, perhaps?) to put together some kind of guide. Will have a search.
I think this kind of stuff is something a lot of us would probably do well to pay more attention to, even (or especially) if space/practicality means that we can't always ensure optimal cable dressing or plug siting etc. Potential free upgrades, effectively. The best kind
My past and especially my more recent experiences of 'cleaning up' my LAN have brought significant improvements to the SQ of my streaming audio. That makes me think that there maybe many sub-optimal LANs out there (including mine to some extent) that are letting down Naim systems upstream of the LAN.
I am not sure if it has been done on here before but there might be some merit in forumites compiling 10 dos and don'ts to optimise a LAN. I realise there is a lot of argument out there over LANs, ethernet cables and 0's and 1's generally but a properly implemented LAN has a significant influence on upstream SQ IMHO. So I think a simple guide on how to implement a LAN could be very useful and provide a free of charge, or a low cost, upgrade to many.
If there is interest out there for such a guide, maybe those with experience could contribute their (say) 10 tips to optimise your LAN and see where it goes. I might need to start a new thread to see if there is any interest.
Many (most?) of us these days have streaming systems but many (including me) have little knowledge on how to implement an optimal LAN - the building block for a great streaming system.
Any thoughts out there?
Posted on: 16 January 2016 by nigelb
I was wondering if there could also be benefits using it on other network devices using SMPSs. .............. I note that my BT Home Hub 5 router had a crappy SMPS on it. Do you think there are any technical reasons why a better SMPS (like the iPower) might bring SQ benefits on a router? Can a router with an average SMPS introduce noise into a network?
The BT SMPS is labelled 12V 1500mA. = 18W. I've found some www forum info that says HH5 idles ~5.4W.
The iPower is rated at 12V 1.1A = 13.2W so it should be OK
I'm not sure it will bring any changes to your system SQ unless (maybe) its plugged into the same & close to the Naim power circuit
Thanks Mike. The iPower appears to have the electrical capability to power the BT HH5. I was wondering if the BT SMPS/Hub has the potential to introduce noise, and if the noise-cancelling technology of the iPower could usefully reduce it. I already have my network devices (switch, router, server) separated physically and electrically (as much as possible) from my Naim black boxes.
I guess the only way is to try it.
Posted on: 17 January 2016 by Dave***t
I'm trying to piece together the reasoning behind the latter parts of that guide, when it comes to STP, UTP and unshielded couplers.
As a simple rule of thumb for LAN grounding, am I right in thinking that what you want is just one grounding point (probably at the streamer, in this case)?
Which would imply that in a network involving at least some STP cables, which has more than one grounded piece of equipment, and assuming that the switch(es) carry the shield through (ie have metal port surrounds), you'd want a UTP cable or an unshielded coupler at some point between the two grounded pieces...?
It all seems quite arcane, but I sense there's something fairly simple behind it.
If I've got it somewhere near right, that'd make sense of the fact that the switch I have closest to the Naim has an exterior ground screw on it like some phono preamps - it'd be for instances where nothing connected to it has its own ground. So not needed here, because the Naim is grounded.
Apologies, BTW, if this is getting off topic. Could make a new thread if it's doing anyone's head in.
Posted on: 17 January 2016 by Mike-B
...................... Anyhow, to the Meicord Opal's: They are thinner than the old Supra Cat7, but surprisingly stiff compared to another Cat6 I have. The plugs are excellent quality & the clips appear a lot more robust.
The change in sound was surprising as I really wasn't expecting too much if anything. The problem is the SQ changes with the iPower & now the Meicord's are a bit blurred. However I pretty certain the Meicord's have bought a change for the better with clarity, definition & dynamics. Anyhow, about to start another listening session. Maybe some more later
Re the Meicord ethernet cables. I am a little surprised because I can hear a difference but am less surprised now after reading up on a possible reason why this might be.
Meicord as a company/brand was started by 2 two radio engineers who experimented with the ethernet & RJ45 connection methods & found that optimising the twisted pair routing into the RJ45 pins improved the sound. As a result Meicord use wire management formers to route the individual wires & these have different Left+Right orientation for each end of the cable. The importance of wire routing & its affects on NEXT (near end cross talk) & RL (reflection loss) & resultant cable performance is well understood & can be found in many paces around the www. Meicord also discovered that UTP sounds better than STP, but are still experimenting.
Another company that have stressed the importance of terminations is Blue Jeans Cables, & both Blue Jeans & Meicord have published tests on other commercial brand patch cords that show few actually met (passed) the performance criteria for the Cable Category standard; both found Cat-5 cables that fail Cat-5 standards & Cat-6 that don't meet Cat-6 & most shocking some Cat-6 fail to pass even at Cat-5.
Interesting to note that Blue Jeans & Meicord supply a cable test data sheet with each cable – Meicord's is a generic test for that specific cable length (not the purchased actual cable)
The Meicord's have a different sound in my system & are more open than the old Supra Cat-7. I said in the previous post they had clarity, definition & dynamics, but that's part of the more open presentation I think; as is the bass texture & treble clarity. Mrs Mike thinks they are slightly different & described it as smoother.
Other than the cable engineering stuff I have no other explanation, but I'm a dyed in the wool objectivist, so what do I know.
Anyhow, this is an iPower SMPS thread, I've digressed into thread drift.
Posted on: 17 January 2016 by Mike-B
Thanks Mike - very interesting. If I may stay off topic a little longer, have you now gone for screened or non-screened ethernet cables and why? Apologies if this has been explained earlier but I am getting a little confused over the relative merits of screened vs non-screened.
Sorry Nigel I found an error in my post & had to edit it, & as it was past the 15 minute limit I had to copy/delete/paste/edit, so its now out of sync with your post.
The Meicord's are UTP - not screened. STP - screened cables - are not used commercially except in very noisy environments. Its seems STP is is something the audio & AV community have taken on board as a fad (for want of a better word) & anything lower that Cat7 & 600 MHz is almost unworthy. However all audio renderers only support (accept) 100 MHz, so anything faster is surplus to requirements; as I said in another post, its like having a Ferrari for a shopping trip.
I had assumed a screened cable was good for my system because as a marine electrical engineer in a previous life, in military applications it was required to separate, screen & clamp all sorts of different radio, radar & voltage frequencies, so a screened ethernet was a natural thing to do when I first started streaming.
I am now also aware that the screen can carry various noise & its possible it can have an effect on something, SQ maybe. But don't misunderstand me I am not saying STP is bad or inferior in any way, I'm quite looking forward maybe someday to test some AQ Vodka
Posted on: 20 January 2016 by Mike-B
Anyhow, to the Meicord Opal's: They are thinner than the old Supra Cat7, but surprisingly stiff compared to another Cat6 I have. The plugs are excellent quality & the clips appear a lot more robust.
The change in sound was surprising as I really wasn't expecting too much if anything. The problem is the SQ changes with the iPower & now the Meicord's are a bit blurred. However I pretty certain the Meicord's have bought a change for the better with clarity, definition & dynamics. Anyhow, about to start another listening session. Maybe some more later
To wrap this ethernet thread deviation up .......... I temporally hooked up the old 3m of Supra Cat7 between switch & NDX & re-familiarised myself, then reconnected the Meicord Opal Cat-6 (4m).
As a cable sceptic I'm surprised to say they do have a different sound. Its more open than the Cat-7 & with a better sense of clarity & definition in the mid/high ranges & sound stage, bass is a touch stronger, not the amount/volume but more in the texture & I'm hearing more the longer I listen.
As an objectivist I am just a little surprised, but am less surprised after reading up on a possible reason why this might be. Apart from high quality cable & in-house designed RJ45 plugs, the all important cable to RJ45 connection is optimised with left & right handed wire management formers to route the individual twisted pairs correctly into the RJ45 pins. This twisted pair management minimises negative effects of NEXT (near end cross talk) & RL (return/reflection loss). The theory is well understood & can be found in many paces around the www.
So apart from the Meicord cable engineering stuff, I have no other explanation as to why this gain in SQ should be. , but I'm a dyed in the wool objectivist, so what do I know.
Posted on: 20 January 2016 by Simon-in-Suffolk
Mike, yes, in my experience connectors and connector connections are so important, and carelessly breaking the geometry of the twisted pairs will break the balance and impedance.. and of course if the impedance varies by a significant amount then the reflections and the SWR will increase, though I doubt there will be big issues here ... However I suspect more affects will be down to the geometry errors of the twisted pairs at the connectors... and of course similar things can happen with coax into a phono connector (phono SPDIF)
Simon
Posted on: 20 January 2016 by nigelb
Thanks for rounding off the ethernet bit. Interesting stuff.
All I would say is if you can borrow some, try a nip of Vodka - you might get addicted. If you want high quality plugs, take a look at those on the AQ Vodka. If you are able to try them, I would be very interested in your findings vs Meicord Opals.
I do of course fully understand if you have had enough of this by now.
