I just installed a pair of these powerline units, and since I don't see
much info about G.hn devices in this forum, I thought I'd pass on my
initial impressions.
Although I whined a few days ago about the lack of powerline gear that can
operate without mains power, I do have one place in my house where that's
not a problem: I have a pair of printers that are not close to any
ethernet jacks. If the power is out, so are the printers, so there's no
need for them to have a UPS-supported network connection. I've been using
a pair of old Apple Airports to bridge them to my house net, but I want to
retire that gear, so powerline seemed like an option to investigate. I've
been pleased with the Zyxel AP I got recently, so I thought I'd try their
offering in the powerline space. I got a pair of PLA6456 units from
Amazon for $70 plus tax, and plugged in one where the printers are and the
other into an outlet on the same circuit that is close to one of my
network switches.
These units are fairly large, about the same size as a standard dual wall
socket. They did manage to arrange it so the unit can cover just one
socket, letting you plug something else into the other socket --- but then
you'll need several inches clearance beside the socket for the unit's body
and the outgoing ethernet cable. In any case, you probably want to use
the unit's pass-through socket instead, as the manual claims that has
filtering to reduce electrical noise problems.
Setup is pretty much a breeze. All you really need to do is sync the
units to create an encrypted connection, by pressing their "ENCRYPT"
buttons within a minute or two of each other. However, each unit has an
http-accessible configuration page if you feel the need. I'd recommend at
least visiting it for long enough to change the factory-default password.
(A possible annoyance for non-techies is that the units initially get
their IP addresses from your DHCP server, so if you don't know how to look
at the DHCP server's leases then you'll not know which IP address to visit.)
The units boot impressively quickly, passing data less than ten seconds
after power is applied, which is a lot better than any wireless gear I've
used. After seeing that, I made a point of hand-assigning IP addresses to
them, because my DHCP server definitely doesn't boot that fast.
Performance is, well, a bit wonky. I initially got some very inconsistent
results with iperf3 tests, ranging all the way from 50 to 500 Mbps. The
only potentially performance-relevant setting I can find is the "G.hn
profile", which defaults to "PLC 100MHz MIMO Boost", but the manual
recommends changing to "PLC 100MHz Boost" if you have the things plugged
into ungrounded sockets. These sockets are grounded according to my
wiring tester, but nonetheless things got much more stable after I
switched to the latter setting. (It would astonish me not at all to learn
that the grounding is substandard; I live in an old house with a lot of
sketchy wiring.) I am now seeing fairly consistent iperf3 speeds of ~230
Mbps when sending data towards the printers, but rather less consistent
speeds, as low as 110 Mbps, in the other direction. Don't ask me why.
While these speeds are a lot less than what you might expect from the
advertised "2400 Mbps", they're plenty for my need; both printers have
only 100Mbps ethernet ports.
(I think that the "2400 Mbps" advertising-ese really means "1200 Mbps each
direction, full duplex". I did see the units reporting close to that PHY
rate initially, when I had them both plugged into one power strip for
configuration purposes. As-installed, the PHY rate is around 450 Mbps,
so getting half that in real-world throughput is roughly on par with
what you get from WiFi gear. Can't say if more modern electrical wiring
would have given me better results --- but I'm glad I'm not trying to
pass data across different circuits.)
Another thing that disturbed me initially was terrible-looking ping
statistics, along the lines of
rtt min/avg/max/mdev = 1.063/17.413/58.288/10.931 ms
However, closer inspection showed that when no other data is being passed,
the ping round-trip time cycles predictably between 32ms and ~1ms,
repeating over 20 seconds or so. Immediately after running an iperf3
test, it's close to 1ms and stays that way for 10 seconds before rising.
So I think the slow pings are from some weird power-save-while-idle
behavior that probably won't have much impact on real-world performance.
Bottom line is that I'm satisfied with my purchase. These units do what
I needed, and I'm not sure I could have gotten better results for $70
total with new WiFi gear. Plus I've removed one wireless net from my
immediate surroundings, which should help my primary wireless net at least a bit.
much info about G.hn devices in this forum, I thought I'd pass on my
initial impressions.
Although I whined a few days ago about the lack of powerline gear that can
operate without mains power, I do have one place in my house where that's
not a problem: I have a pair of printers that are not close to any
ethernet jacks. If the power is out, so are the printers, so there's no
need for them to have a UPS-supported network connection. I've been using
a pair of old Apple Airports to bridge them to my house net, but I want to
retire that gear, so powerline seemed like an option to investigate. I've
been pleased with the Zyxel AP I got recently, so I thought I'd try their
offering in the powerline space. I got a pair of PLA6456 units from
Amazon for $70 plus tax, and plugged in one where the printers are and the
other into an outlet on the same circuit that is close to one of my
network switches.
These units are fairly large, about the same size as a standard dual wall
socket. They did manage to arrange it so the unit can cover just one
socket, letting you plug something else into the other socket --- but then
you'll need several inches clearance beside the socket for the unit's body
and the outgoing ethernet cable. In any case, you probably want to use
the unit's pass-through socket instead, as the manual claims that has
filtering to reduce electrical noise problems.
Setup is pretty much a breeze. All you really need to do is sync the
units to create an encrypted connection, by pressing their "ENCRYPT"
buttons within a minute or two of each other. However, each unit has an
http-accessible configuration page if you feel the need. I'd recommend at
least visiting it for long enough to change the factory-default password.
(A possible annoyance for non-techies is that the units initially get
their IP addresses from your DHCP server, so if you don't know how to look
at the DHCP server's leases then you'll not know which IP address to visit.)
The units boot impressively quickly, passing data less than ten seconds
after power is applied, which is a lot better than any wireless gear I've
used. After seeing that, I made a point of hand-assigning IP addresses to
them, because my DHCP server definitely doesn't boot that fast.
Performance is, well, a bit wonky. I initially got some very inconsistent
results with iperf3 tests, ranging all the way from 50 to 500 Mbps. The
only potentially performance-relevant setting I can find is the "G.hn
profile", which defaults to "PLC 100MHz MIMO Boost", but the manual
recommends changing to "PLC 100MHz Boost" if you have the things plugged
into ungrounded sockets. These sockets are grounded according to my
wiring tester, but nonetheless things got much more stable after I
switched to the latter setting. (It would astonish me not at all to learn
that the grounding is substandard; I live in an old house with a lot of
sketchy wiring.) I am now seeing fairly consistent iperf3 speeds of ~230
Mbps when sending data towards the printers, but rather less consistent
speeds, as low as 110 Mbps, in the other direction. Don't ask me why.
While these speeds are a lot less than what you might expect from the
advertised "2400 Mbps", they're plenty for my need; both printers have
only 100Mbps ethernet ports.
(I think that the "2400 Mbps" advertising-ese really means "1200 Mbps each
direction, full duplex". I did see the units reporting close to that PHY
rate initially, when I had them both plugged into one power strip for
configuration purposes. As-installed, the PHY rate is around 450 Mbps,
so getting half that in real-world throughput is roughly on par with
what you get from WiFi gear. Can't say if more modern electrical wiring
would have given me better results --- but I'm glad I'm not trying to
pass data across different circuits.)
Another thing that disturbed me initially was terrible-looking ping
statistics, along the lines of
rtt min/avg/max/mdev = 1.063/17.413/58.288/10.931 ms
However, closer inspection showed that when no other data is being passed,
the ping round-trip time cycles predictably between 32ms and ~1ms,
repeating over 20 seconds or so. Immediately after running an iperf3
test, it's close to 1ms and stays that way for 10 seconds before rising.
So I think the slow pings are from some weird power-save-while-idle
behavior that probably won't have much impact on real-world performance.
Bottom line is that I'm satisfied with my purchase. These units do what
I needed, and I'm not sure I could have gotten better results for $70
total with new WiFi gear. Plus I've removed one wireless net from my
immediate surroundings, which should help my primary wireless net at least a bit.