I use MoCA. Should I consider powerline?

[infotime]

I've done one network with MoCA. It worked but I feel like I'm cheating. I'm looking at the TP-Link Powerline adapters on Amazon and they're pretty well reviewed. 800+ reviews averaging 4.5 stars for the 500 Mbps model.

It sounds a little too good to be true. $60 for a pair of adapters to build an "ethernet" backbone for a home multi-access point wireless system. As long is the backbone is faster than the ISP then in most of the installs I'd have in mind it wouldn't matter. A helluva lot easier and faster than cutting drywall and fishing cable.

Seems to me that multiple mildly powered access points placed throughout the home would be better than one monster router. I liken it to speakers. You don't put a single 500 watt speaker in the center of the house and turn it up really loud so you can hear it all over. Your ears would bleed sitting next to it. Instead you put just what you need in each room whether it's speakers or WiFi.

 

[thiggins]

Look at the TP-LINK TP-WPA4420KIT. Combines an AV500 adapter and combo AV500 Adapter/ 2.4 GHz N300 access point for $60.

At that price you would think it would be meh performance. But I have tested for an upcoming review and found pretty consistent 90 Mbps powerline performance and the AP is 60 - 80 Mbps.

 

[infotime]

Wow, that's interesting. A main unit and a couple of combo units strategically placed around the house and you might have a nice system. Question: I suppose you would degrade the backbone performance by having multiple powerline adapters, correct?


Sent from my iPhone using Tapatalk

 

[thiggins]

Wow, that's interesting. A main unit and a couple of combo units strategically placed around the house and you might have a nice system. Question: I suppose you would degrade the backbone performance by having multiple powerline adapters, correct?

Multiple adapters share the total available bandwidth, just like any communication channel.

 

[infotime]

Been looking at the TP-Link stuff on and off when I have a chance since yesterday. I'm having trouble finding a reason to not use this or something like it in a large house that might need three or four access points. What am I missing?

 

[thiggins]

Why not just try it? For $60, it's worth the experiment. And Amazon return policy is liberal if they don't work out.

The catch right now is that you have to buy the AP in a bundle with an adapter. If you are doing multiple APs, you'll end up with extra adapters.

The other catch is that the AP is 2.4 GHz only.

 

[azazel1024]

It is probably worth the experiment. Downside is, because it is shared, speeds could be an issue. If the only thing you would ever use it for is extending your internet connection wirelessly and you don't or would never have access to a fast connection, it should work fine.

Shared won't really matter. However, if you have local resources on your network you want to access, the shared nature of powerline might bite you in the butt.

Obviously if cat 5e/6 can't be run to the locations you'd want the access points, you are SOL.

However, one thing I never see anyone mentioning is using existing phone wiring. That is usually cat3, but once you change the jack from an RJ11 to an RJ45 it can usually support 10/100 speeds, which likely will be faster than powerline and IS NOT shared, at least not as soon as you seperate the phone wiring off the phone bus to a patch panel and then in to a switch.

I've seen a couple of people do this with existing phone wiring and it worked okay. It isn't gigabit speeds, but I've seen modest lengths of cat3 phone wiring support 100Mbps just fine.

You'll have to make sure it is actual cat 3 and not jacketed station wiring (if not twisted, the cross talk is probably going to be too high for 100Mbps, but you can do 10Mbps over it).

Anyway, it might be worth a shot at an even cheaper price.

 

[stevech]

I'd like to know, and would expect, if the 500Mbps (raw bit rate) power line can, on average, across different attenuation and noise conditions in many residences, sustain an average net yield of 200Mbps or so. Not in just one above average situation (cross-power-phases, dimmer noise, etc.).

Knowing power wiring challenges, I'd doubt it. Likely, these products use rate-adaptation (based on SINR) as does 802.11/WiFi, slowing as the SINR degrades.

Has there been any test results with realistic conditions?

 

[thiggins]

I'd like to know, and would expect, if the 500Mbps (raw bit rate) power line can, on average, across different attenuation and noise conditions in many residences, sustain an average net yield of 200Mbps or so. Not in just one above average situation (cross-power-phases, dimmer noise, etc.).

Knowing power wiring challenges, I'd doubt it. Likely, these products use rate-adaptation (based on SINR) as does 802.11/WiFi, slowing as the SINR degrades.

Has there been any test results with realistic conditions?

Right now, 90 - 100 Mbps is more like it for AV500. We'll see what AV2 MIMO (AV1000) products do once they start shipping.

 

[azazel1024]

AV2 MIMO has a much better chance of managing it from what I know of it. It bumps things up to using the ground wire too (which I don't think any earlier form does) as well as a number of other changes.

I'd be pleasantly suprised if it hit over 100Mbps in typical setups though...but maybe. With ideal ones, it might hit over 200Mbps (but I'd be suprised).

Ground wires are generally noiser than power or neutral...of course with the exception of times when neutral and ground are bonded in the panel...in which case neutral is generally just as noisy as ground...and I question what using ground would do for you, though I'd imagine you could use it as a comparative even if it isn't an isolated signal path from neutral.

 

[stevech]

by "ground" wire.. do you mean the Neutral as in No. American wiring?
Hot (black), Neutral (white) and green (ground).

Ground may be connected to all sorts of water pipes, earth ground rods, etc. Seems like a bad medium for RF.

 

[stevech]

Right now, 90 - 100 Mbps is more like it for AV500. We'll see what AV2 MIMO (AV1000) products do once they start shipping.

Wow. If the IP layer yield is under 100Mbps, for a 500Mbps link layer burst rate, the marketing guys will have a field day with this!
Product package: "GET 500Mbps WITH OURS!!!"

 

[sfx2000]

hmmm... if the noise is known, then I can calculate the DC offset from ground, and modulate from there at the most basic level. All wireless is like this, and MoCa, HomePlug, etc... it's RF... then it's down the what RF we can push down the wire...

interesting... if not optimal

 

[stevech]

DC offset, etc. concept doesn't work in this RF singnal case. We're talking about noise summed in with the desired signal.

 

[thoma]

To quote from the design manual, Homeplug AV and IEEE 1901: A Handbook for PLC Designers and Users (p. 313):

Note that only two unique signals can be transmitted on three wires due to Kirchoff's voltage law: the signal third wire pair is the sum of the signals on the other two wire pairs.

On receive, there are up to four possible receive channels: Line-Neutral, Line-Ground, Neutral-Ground, and common mode. Note that the three wire pairs are subject to Kirchoff's voltage law, but in practice the third wire pair does provide additional diversity evidently because of non-ideal implementations of the coupling to the powerline wiring. Common mode is the signal between true earth ground and the three wire pairs and is generally only available in products that have a relatively large effective ground plane, such as a flat screen television.

Send two signals, but receive up to four signals. But neutral is tied to ground at the service panel! The frame of the flat-screen television is tied to the ground pin on the plug! Why do these signals exist at all? But they do.

These signals may be weak, and they may be noisy, but they exist. And that means that you can throw it all into the signal-processing algorithms and come out with additional bandwidth. It's like magic.

You really have to think like a wireless designer. If you think about it, in wireless MIMO, the antenna arrays are only a few inches apart. Yet even this small spatial offset creates enough diversity in the channel characteristics to exploit through signal processing.

To quote a company that specializes in MIMO:

Spatial Multiplexing is often the technique that people find difficult to believe if not grasp. http://silvustechnologies.com/technology/introduction-to-mimo/

 

[azazel1024]

In the US with 3-phase and 120v, the hot/power wire is black, neutral is white and ground is bare copper (green is only ground within an appliance/fixture. Its bare copper in the Romex/residential wiring). In most situations ground is bonded to the neutral bus in the main breaker box (and you generally do NOT want to do that in sub panels).

I can see with different signal paths how you potentially gain something from using the ground wire too for carrying a signal as it gives you something to compare the main signal to. Ground tends to be a bit noisier than neutral, but neutral, due to the nature of the bonding (in part) is also pretty noisy.

 

[stevech]

High frequency RF data on power line is always going to be problematic. IF not on day 1, it'll be at day n, as signal attenuating things are plugged in, or as dimmers and motors make high freq. noise.
But IP on power wiring has its place and ranking.