Wavelength Ethernet Spacelator

[Miska]

3 hours ago, marce said:

PLT, power line communication is adding EMC to the mains, so you distribute it everywhere, it can couple all over the place and filtering it is not as easy as some presume…

A lot of you get excited about a few pA of noise from a SMPS, yet will quite happily add far more noise to the mains using PLT, it is a bit of a joke…

Anyway if you want to get a broader view of the problem here is some info from EMCIA:

https://www.ban-plt.org.uk/emcia.php

 

Already at the beginning it starts talking about technologies using power lines to deliver internet access and not about local networking:

"as an option for the delivery of broadband Internet" ... "One mechanism for providing broadband Internet access might be thought to be PLT, and similar digital communications technologies, that use the electric power lines (the "mains") to, and within, homes to carry the high speed data signals used by broadband Internet."

 

But other than that, the rant sounds much more like industry political battle where traditional telcos are worried that power companies would become competitors. Power companies are anyway already using PLT technologies for their smart grid over longer distances than the home networking devices can do. For home networking the power levels needed to carry the data are so low that it doesn't pose much problem.

 

We are now specifically talking about https://en.wikipedia.org/wiki/IEEE_1901 and https://en.wikipedia.org/wiki/G.hn

 

And no, I don't get excited about SMPS noise, because I have a probably closer to 50 SMPS thingies running in the house.

 

[Superdad]

4 hours ago, marce said:

PLT, power line communication is adding EMC to the mains, so you distribute it everywhere, it can couple all over the place and filtering it is not as easy as some presume…

A lot of you get excited about a few pA of noise from a SMPS, yet will quite happily add far more noise to the mains using PLT, it is a bit of a joke…

 

And I seem to recall that PLT (of utility companies) in the USA uses a 19Khz carrier!

[marce]

1 hour ago, Miska said:

 

Already at the beginning it starts talking about technologies using power lines to deliver internet access and not about local networking:

"as an option for the delivery of broadband Internet" ... "One mechanism for providing broadband Internet access might be thought to be PLT, and similar digital communications technologies, that use the electric power lines (the "mains") to, and within, homes to carry the high speed data signals used by broadband Internet."

 

But other than that, the rant sounds much more like industry political battle where traditional telcos are worried that power companies would become competitors. Power companies are anyway already using PLT technologies for their smart grid over longer distances than the home networking devices can do. For home networking the power levels needed to carry the data are so low that it doesn't pose much problem.

 

We are now specifically talking about https://en.wikipedia.org/wiki/IEEE_1901 and https://en.wikipedia.org/wiki/G.hn

 

And no, I don't get excited about SMPS noise, because I have a probably closer to 50 SMPS thingies running in the house.

 

I would read the whole of the info there is more out there, having had  a vested interest in EMC, I have to disagree. You are adding noise to your mains wiring system, a nice collection of long un-shielded cables meandering all around your domestic environment. Noise that will couple through most basic filters into anything plugged into the mains.

Here.s an historic thread from DIY Audio, where I rant and provide some practical (and tested) methods of mitigating this noise from 2015.

https://www.diyaudio.com/forums/pc-based/278547-quick-audioquest-jitterbug-review.html#post4422693

 

And here from 2012 is some stuff on switchers noise and EMC

https://www.diyaudio.com/forums/power-supplies/193705-switcher-emc-design.html

 

Now lets get practical over power levels...

Laptop, 10mW, router 100mW wireless not a lot of power...

The effects of noise on sensitive circuits...

As some example have shown on this forum, use a high impedance probe and very low current interference can create a large voltage (Ohms law I believe). Within a group of circuitry (even within a single device) there are high and low impedance nodes. Guess which nodes are sensitive to even very low levels of induced noise... The high impedance inputs and to make matters worse these are often critical inputs (power supply feedback pins are one example, reference inputs to converters etc.) so any induced noise can however low can have a large effect on the circuit operation. There are numerous ways of protection, EMC protection at inputs and outputs, good power delivery system design and protecting the sensitive nodes, one of the more common ones for very sensitive analogue is removing the ground from under the device pins (SMD) to reduce capacitance.

Capacitance, perfect for coupling RF noise or any high frequency noise.........

So no I disagree, its an easy way of avoiding installing a proper LAN, I'm running CAT6 everywhere and wireless.

 

 

[marce]

33 minutes ago, Superdad said:

 

And I seem to recall that PLT (of utility companies) in the USA uses a 19Khz carrier!

Luckily like me I would guess those frequencies are just a memory from our youth? but again not a good frequency for added noise.

[Miska]

4 hours ago, marce said:

Now lets get practical over power levels...

Laptop, 10mW, router 100mW wireless not a lot of power...

 

So far I have seen much more problems out of cellular phone transmitting at it's full 2W power near sensitive equipment, especially because it is pulsating packet transimission. And large electric motors starting that you can even notice as lights blink. Also when I turn on my power amp lights blink. Comparing than anything from power line comms...

 

Those MHz frequencies don't really get much through with normal linear or switching power supplies.

 

If you are more worried, you can always use mains regenerator or off-line UPS for audio equipment.

 

And what are you going to do about electric company's smart grid data on the network anyway?

 

4 hours ago, marce said:

You are adding noise to your mains wiring system, a nice collection of long un-shielded cables meandering all around your domestic environment.

 

That is more of a problem in a different way, having couple of kilometers of air wires on the poles anyway picks up all kinds of things, including lightning induced electricity.

 

[Miska]

4 hours ago, Superdad said:

 

And I seem to recall that PLT (of utility companies) in the USA uses a 19Khz carrier!

 

That can actually get through LPSU mains transformer, while something like 25 MHz doesn't really go...

 

[cjf]

11 hours ago, octaviars said:

 

Not that it perhaps matter but Roon switched from UDP to TCP in build 234.

 

Today’s release of the core includes a re-design of RAAT’s audio streaming protocol that uses TCP instead of UDP to transmit the audio stream.

Hrrrrmm, thanks for the heads up. I'll have to poke around and check it out.

 

Interesting that they would move back to TCP from UDP (I believe they were TCP in the past/initially also) considering that there is far less overhead with UDP (ie...its faster).

 

I would argue that the need for TCP is unnecessary for our purposes much like it is unnecessary for for real-time voice applications. If errors occur during transmission, the healing capabilities of TCP take too long and result in in-coherent sound or echoes...etc.

 

With UDP the errors just get discarded which would seem more ideal for HQ audio purposes (since it is real-time playback essentially). I would rather hear a "burp", if things went wrong,  then out of sync re-transmitted TCP packets.

 

Good info..thanks again

[marce]

9 hours ago, Miska said:

 

So far I have seen much more problems out of cellular phone transmitting at it's full 2W power near sensitive equipment, especially because it is pulsating packet transimission. And large electric motors starting that you can even notice as lights blink. Also when I turn on my power amp lights blink. Comparing than anything from power line comms...

 

Those MHz frequencies don't really get much through with normal linear or switching power supplies.

 

If you are more worried, you can always use mains regenerator or off-line UPS for audio equipment.

 

And what are you going to do about electric company's smart grid data on the network anyway?

 

 

That is more of a problem in a different way, having couple of kilometers of air wires on the poles anyway picks up all kinds of things, including lightning induced electricity.

 

Maybe you should look at who published the first paper regarding this I posted.

 

9 hours ago, Miska said:

 

That can actually get through LPSU mains transformer, while something like 25 MHz doesn't really go...

 

Capacitive coupling I mentioned it earlier.

[marce]

Interesting reaction to somthing that should be of concern to many on this site... 
Considering we had a long thread about the benefits of screened isolation transformers, many threads on the evils of a bit of leakage current, noise and how to go to extremes to minimise it.
I put up some concern about ethernet over mains cables, adding EMC and a lot of you start dissing the information, yet its from some of the most recognised people in EMC in the world!
I get the impressions if a more subjective member had put the info up it would have been better recieved.
Someone tells you that DA (dielectric absorption) has an effect at DC you lap it up, someone points out this is
silly and they get crucified. Science (physics) and engineering gets dissed unless its twisted to fit in
with a lot of the beliefs.
And a lot will follow the teachings of a Farmer and his box of dirt that will magically
attract noise from a system rather than the real facts on what a real physical ground can and cannot do...
I don't understand the attitude, either you are concerned about noise on your mains or your not...

[Miska]

I measure lot of devices and have quite a bit of measurement equipment. And I have done quite a bit of measurement equipment.

 

For any gear you must begin with the assumption that mains is in any case horrendously dirty. You can start with assumption that in worst case it will roughly stay within +-50V of noise in range from 50 Hz to 1 MHz, and will contain noise up to GHz range at lower levels. And you may gets spikes that reach kV levels. Then you are good.

 

Standard RFI/EMI filters that attenuate noise about 30 dB are pretty good already. I've got Siemens ones in the equipment I've built.

 

@marce can you please post measurements where this stuff really shows up in analog output of various DACs or similar that we can verify it being a real problem?

 

I'm not recommending to use HomePlugs or similar if there are other ways that suffice, but sometimes there are no other reasonable ways to do it, there's no need to do lot of fear mongering about it. In any case you have no control over what other peoples like neighbors, nature or electricity company does on the grid. If you cannot access your music, you certainly have another kind of problem. Or if you have a computer that sounds like vacuum cleaner in your listening room that's certainly noisy in a different kind of way.

 

[Miska]

Here's output noise spectrum of a standard type of LPSU based on LM317:

 

Now combine that with standard PSRR of a good opamp that is 120 dB and it is not that bad at all.

 

 

If there are good practical and objective results from audio gear, I'm happy to see.