Meridian Unplugged - Powered by MrTechGuy.co.uk Forums Discussion Meridian - Streaming LAN isolator anyone?

Hi Hector,

Just to confirm, are you talking about having a network switch with galvanic isolation built in? If so, do you know if such switches exist?

I'm now wondering where isolation is best applied, and if it ideally needs to be applied in several locations in the signal chain. Gut feel suggests it would help to nail the noise earlier in the chain, rather than in the last few cm before the ID4x card. I'll experiment a bit with putting the GISO in different places, although I think I may need a couple of units to get any decisive results. I also wonder if multiple units might bring some negative impact.

I may follow Cujo's lead and try in some good quality linear PSUs on the Wi-Fi router and switch to cut out some of the nasty 500KHz noise there. PSUs share their earth connections with the units they power and, consequently with the earth used in Ethernet cable. Each side of each balanced pair carries a signal that is referenced to earth, so this is the perfect place to inject degrading noise.

Of course, there's also the Wi-Fi router, in my case, sitting half a metre from the switch, happily banging out RF at 2.4GHz and 5GHz. RF loves push fit connection points as a place to work its way into a signal chain. Maybe I should build a Faraday cage and put the switch inside that ... :-)

Dave

...apart from the OP - perhaps you missed that.

With regard to use of this device on the ethernet connections, per my previous post it appears to be reducing noise at 5KHz (which may or may not be present on the system in question).
Any device that affects the response at one frequency will have effects at all frequencies.
So maybe the questions should be:
is Meridian hardware affected by "noise" on ethernet in the 5KHz region? Bob Stuart is not short of experience in this field so it is reasonable to suggest that noise immunity is strong and has been considered in detail.

Clearly reducing any noise is beneficial however if this benefit is then outweighed by the effects of signal/edge distortion at higher frequencies which frankly are more likely to affect sound due to timing and jitter issues that could be introduced. Again Mr Stuart's expertise in low jitter PLLs is second to none so it is difficult to quantify this.
In the absence of being able to make detailed and comparitive electronic measurements on actual M systems, we have to rely on peoples acoustic experience and I am the first to acknowledge that acoustically it is possible to perceive differences that are nigh on impossible to explain technically.

I am mildly concerned at the nature of the SQ improvements perceived. We all love a valve amp - sounds beautiful - but is it accurate? Could this be an artefact of effects to the form of the high frequency data?

At the end of the day at $200 it is not exactly bankrupting for those with systems of many $1,000s. So for ethernet use why not give it a try. A few reports back should fairly quickly confirm whether positive as experienced so far/just different or even negative.

Hi Hector - indeed ethernet Galvanic isolation can be achieved with fibre ethernet, available at suprisingly accessible prices these days.

* edit: correction as per MattF's post below Cu ethernet is Galvanically isolated anyway *

Problem is at some point it has to be turned back to electrical signals for the RJ45 input of the M device.

Arguably there could be a benefit if only the last short leg is electrical if the network and it devices could be shown to cumulatively add noise which could be isolated by a fibre link to a good quality switch serving only the M device(s).

Anyone want to try it?

If differences are real, maybe there is a case for M to design in fibre ethernet ports in future, galvanic isolation achieved and of course the conversion to electrical data is then fully in M's control within the product.

I would imagine a systems designer would steer well clear of these discussions for protection of their sanity, or to resist the temptation to correct all the false 'beliefs'

The best way to achieve galvanic isolation would be to use copper cable and something called 'Ethernet'

This thing called Ethernet has galvanic isolation built in to the specification.

STP/FTP would still have a ground connection - but this can be 'isolated' with a plastic bodied (no shield connection) RJ45-RJ45 coupler (or a short length of UTP)

Try looking up 'Ethernet Magnetics'

Choices include integrated magnetics into the socket, or an on board package.

Nowadays, the 'fun' you can have with multiple double-insulated products with no real ground plane between them kind of makes this sort of thing necessary.
(and just because there's a metal earth pin doesn't mean it's connected to anything)

The only time I'd worry about using an isolator into a Zone there would also be the initials IS and the numbers 0 1 or 2 (yes I know that's mis-using terminology)

Might have overdone the sarcasm somewhat - Must avoid rant about changing external switch-mode supplies but ignoring the product may be stuffed internally with multiple DC-DC switching regulators (blind testing - if you don't know they're there they don't affect the sound?)

Does adding another isolator every 1m make the sound better or worse?

Tip: If using shielded Ethernet - know where your ground points actually are - check if metal earth pins make it through to the device if using wall-warts and even if your switch isn't grounded - with metal RJ45 sockets the ground can be bridged from a device that is grounded (check all the sockets for the shield contacts) Unplugging the one grounded device could end up with everything else floating, so swapping equipment to check effects on sound could introduce changes that you aren't taking into account.

Dave, many thanks for sharing your findings. A really honest and thoughtful post.

One thing I find quite puzzling in many of your points is the 180 that you take. It makes the sound different and or worse and then later it is different and better.

I can't help wonder what psychology and psycho-acoustics are at work here. Having just read some pieces on the latest in placebo science it is clear we don't really know much about how the mental and physical world interact.

I'm not challenging you here btw. I do think it's great when people find good results in what they try.

Personally I think there is always something going on that needs explaining when your initial instinctive reaction is then over-ruled by a longer more focused opinion.

Again...thanks for sharing.

Gruffle,the original poster had no idea about the GISO and I thought that was what was being discussed further along in this thread.


Davy, I agree with your thoughts. I think the GISO has a break in period of 70 hours or so. I too have dedicated power lines, one for the 818v2 and the other for the amplifier..

If I had to say, the system sounds more "organic" with the GISO inserted. My friends all asked why the system sounded so outstanding on visits after the GISO was installed...

MattF - very true. I admit I'd overlooked that Cu Ethernet is indeed galvanically isolated. Good post esp re STP/FTP grounding, though back to the why does it make an audible difference. Galvanic isolation effectively means DC is blocked/AC coupled; and the transformer sizes are optimised for ethernet high frequencies though some LF stuff will still pass through. I guess this is where this "isolator" claims its performance in "cleaning" up the LF "noise" other artefacts to be debated.

I am also reminded that I am incorrect re medical devices and isolators, this has nothing to do with measurement noise but is safety related as network devices are typically not medically approved (ISO13485) so addtional isolation is used via an approved isolator for compliance reasons.

Hi Nick,

Fair point, diplomatically made and well received!

I think listening to recorded music is all about psycho-acoustics and I�m tempted to think that Bob Stuart�s passion for this whole topic is a big factor in why Meridian kit sounds so good. I know my emotions and state of mind affect how I perceive music � some nights, I get home from work, sit down to listen to music and find I just can�t get into it � the response just isn�t there. Other nights, my hair stands on end and I�m up till 2am playing that �last CD before bed�. No doubt, this is down to state of mind, as the system is probably doing the same thing most nights.

That said, I�d like to clarify a couple of things on my earlier post. If anything, I was guilty of being a tad negatively disposed to the GISO in terms of dynamics, as I�ve so often found that devices involving filtering of any kind kill dynamics. Dynamics seem to affect me a lot and I really can�t live with sound that lacks life and energy. So I looked for that problem and jumped on it � not good behaviour.

My experience of adding components or devices that substantially reduce noise, aside from the �blacker blackness� that reviewers always talk about, is that it always makes the music sound more �organised� and easier to hear into the soundstage. If you don�t listen long enough, or if you focus too much on analysing individual elements of the sound, you can easily perceive the sound as having slowed down when, in fact, the sound is simply more coherent, to the extent you perceive more time between sounds. The telling thing for me was taking the GISO out and stepping back to listen to the overall rhythm and drive of the music. Hands-down improvement with the GISO in place, no question.

As for the 180 turn in my perception of dynamics and soundstage, the first was down to a mix of the point above and burn-in. The second was entirely down to burn-in. For the first 24 hours, the soundstage was simply narrower and less well defined with the GISO in place. This was a brand-new device, cold and fresh out of the box. With hindsight, I should have waited a week before posting � I apologise for jumping the gun!

While I listen to music for the sake of music and the emotions that go with it, my engineering genes force me to be painfully empirical when it comes to assessing the impact of new devices. I can�t speak for others, but my nature is such that, if the gain isn�t there, I learn and move on. Nothing lost, just experience gained, so I�m very happy every time that happens. We all think differently, but I always hold hands-on experience of the impact a device to be far more valuable than preconceptions. Now and then, it�s good to nudge the boundaries.

Truth is, I�m still a bit discombobulated by what the GISO has done to my system � a little gizmo like this hanging on the end of �20 worth of CAT7 cable has no right to bring such a big change such a great system. My guess is that there�s a heck of lot still to be learned about digital signal transmission and that HHers and manufacturers alike haven�t even scratched the surface yet. I think we�ll have some fun watching that unfold and experiencing the benefits.

Right, I�ll bugger off and stop bothering everyone now.

Cheers,
Dave

Hi Kingsrule, thanks for sharing the point on burn-in. Helps me believe I'm not loosing my marbles. Interesting to hear you have dedicated mains too. Yes, Organic is probably the single word that best describes the overall feel of the improvement.

please don't!

In spite of any debate technically - you are the only one here who has one, so you are the only person qualified to comment on its audible effects.

Of even greater interest please, is how you feel about it after a while, perhaps by removing it and reinstating it.

As I said earlier for $200 - if it makes a worthwhile difference then it is a no brainer, but I am interested whether after some time with it you still feel it is a significant improvement or whether it just alters the sound.
Next question will be - is it possible to make a better one!!!

The other interesting observation of the feedback so far is that it "removes the digital grunge", "spotify effect", "spitty and gritty" etc.

These are certainly not terms you would usually expect to hear/read (at least on this forum) when describing the SQ of 818/808v5 going into 7200 & 8000 DSP's.

Maybe these are all just relative terms and are only meaningful in the context of this additional box being layered on top of a very high baseline of performance.

Just a dumb question : in my config (MD600 -> C15 -> DAC2x), where would the isolator go ?

To C15 eth, MD600 eth or both or none ?