From Newsgroup: rec.arts.sf.written

On Mon, 01 Sep 2025 20:53:29 -0400, Joy Beeson
<jbeeson@invalid.net.invalid> wrote:

Friday, 29 August 2025

Can't send this because the internet is down again. We'd be
having some sharp words with BrightSpeed, were it not that
we have already told Surf Internet to install fiber-optic
cable.

The cable was buried a few days ago. I was in Fort Wayne at
the time, but not reading. The new router is at the other
end of the house, and the signal in my office is weak and
intermittent.

Current waiting-room reading is the March/April 2024 Analog,
beginning with "Ganny goes to war" by David Gerrold.

Finding out, about halfway through, that Ganny's official
name is Hazel Stone is a nice easter egg, but it reminds
those who get it that this story is a story. (I peeked: the
last line is "We are going to be all right.")

I loved _The Rolling Stones_ as a teenager, and wished that
the book were a great deal longer.
--
Joy Beeson
joy beeson at centurylink dot net
http://wlweather.net/PAGEJOY/
http://wlweather.net/PAGESEW/

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On Fri, 26 Sep 2025 10:04:25 -0400, Joy Beeson
<jbeeson@invalid.net.invalid> wrote:

On Mon, 01 Sep 2025 20:53:29 -0400, Joy Beeson
<jbeeson@invalid.net.invalid> wrote:

Friday, 29 August 2025

Can't send this because the internet is down again. We'd be
having some sharp words with BrightSpeed, were it not that
we have already told Surf Internet to install fiber-optic
cable.

The cable was buried a few days ago. I was in Fort Wayne at
the time, but not reading. The new router is at the other
end of the house, and the signal in my office is weak and
intermittent.

And I am sure you are exploring fixes for that.
When one of my devices insisted that it would only connect via
Ethernet to the Gateway in a distant room, I got an
Ethernet-over-Power starter pack (two devices that plug into the wall
and provide both an Ethernet port and a power socket to replace the
one occupied) and it is working just fine.
Some notes:
-- The device being connected uses it only for streaming audio from my
main computer (which itself uses WiFi). A more demanding device might
or might not work so well.
-- the two A/C grounded circuits used are in the same breaker box [1]
and not used for much else, so are not very busy. Busy circuits might
affect performance.
But I have also seen offers for WiFi repeaters. The repeaters, of
course, take time to receive and then send the data; the
Ethernet-over-Power does not.
[1] When I bought it this was phrased in the descriptions in such a
way that it sounded like both had to be on the same /circuit/, but
being in the same /breaker box/ is enough. Being on circuits in
/different/ breaker boxes, it appears, will not work.
--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Paul S Person <psperson@old.netcom.invalid> wrote:

But I have also seen offers for WiFi repeaters. The repeaters, of
course, take time to receive and then send the data; the
Ethernet-over-Power does not.

[1] When I bought it this was phrased in the descriptions in such a
way that it sounded like both had to be on the same /circuit/, but
being in the same /breaker box/ is enough. Being on circuits in
/different/ breaker boxes, it appears, will not work.

They have to be on the same leg. If you've got single-phase service
like most us, then all the devices need to be on either even-numbered
or odd-numbered breakers unless there is enough RF leakage between the
two legs. You can get a coupling device to create enough leakage if you
need it.

But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On 9/26/2025 12:21 PM, Scott Dorsey wrote:

Paul S Person <psperson@old.netcom.invalid> wrote:

But I have also seen offers for WiFi repeaters. The repeaters, of
course, take time to receive and then send the data; the
Ethernet-over-Power does not.

[1] When I bought it this was phrased in the descriptions in such a
way that it sounded like both had to be on the same /circuit/, but
being in the same /breaker box/ is enough. Being on circuits in
/different/ breaker boxes, it appears, will not work.

They have to be on the same leg. If you've got single-phase service
like most us, then all the devices need to be on either even-numbered
or odd-numbered breakers unless there is enough RF leakage between the
two legs. You can get a coupling device to create enough leakage if you
need it.

But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.
--scott

Or just don't use WiFi? I have a physical fiber-optic internet
connection and uploads & downloads just /SCREAM/. (But then I also have
an ISP that doesn't throttle my internet connection the way most
telecoms do.)
--
I've done good in this world. Now I'm tired and just want to be a cranky
dirty old man.
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Dimensional Traveler <dtravel@sonic.net> writes:

On 9/26/2025 12:21 PM, Scott Dorsey wrote:
Paul S Person <psperson@old.netcom.invalid> wrote:

But I have also seen offers for WiFi repeaters. The repeaters, of
course, take time to receive and then send the data; the
Ethernet-over-Power does not.

[...]

They have to be on the same leg. If you've got single-phase service

like most us, then all the devices need to be on either even-numbered
or odd-numbered breakers unless there is enough RF leakage between the
two legs. You can get a coupling device to create enough leakage if you
need it.
But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.
--scott

Or just don't use WiFi? I have a physical fiber-optic internet
connection and uploads & downloads just /SCREAM/. (But then I also
have an ISP that doesn't throttle my internet connection the way most telecoms do.)

Does that mean you have installed a home optical network? I know some
people have. I went for a cat-6 network in my house and don't use WiFi
(except on occasions for a visitor). Unfortunately I have HFC, not
fibre and it isless reliable, especially as they have electronics in the
path without backup power so a one point failure takes all downstream
out in the segment.
--
Stephen Harker sjharker@aussiebroadband.com.au
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Dimensional Traveler <dtravel@sonic.net> wrote:

On 9/26/2025 12:21 PM, Scott Dorsey wrote:

But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.

Or just don't use WiFi? I have a physical fiber-optic internet
connection and uploads & downloads just /SCREAM/. (But then I also have
an ISP that doesn't throttle my internet connection the way most
telecoms do.)

Ethernet-over-power isn't wifi. It's much, much slower than wifi.
Also less secure. May or may not be less reliable.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Stephen Harker <sjharker@aussiebroadband.com.au> writes:

Dimensional Traveler <dtravel@sonic.net> writes:

On 9/26/2025 12:21 PM, Scott Dorsey wrote:
Paul S Person <psperson@old.netcom.invalid> wrote:

But I have also seen offers for WiFi repeaters. The repeaters, of
course, take time to receive and then send the data; the
Ethernet-over-Power does not.

[...]

They have to be on the same leg. If you've got single-phase service >>> like most us, then all the devices need to be on either even-numbered

or odd-numbered breakers unless there is enough RF leakage between the
two legs. You can get a coupling device to create enough leakage if you >>> need it.
But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.
--scott

Or just don't use WiFi? I have a physical fiber-optic internet
connection and uploads & downloads just /SCREAM/. (But then I also
have an ISP that doesn't throttle my internet connection the way most
telecoms do.)

Does that mean you have installed a home optical network?

That's not likely. The ISP connection may provide up to
1Gb/sec speeds (100M bytes/sec). Any internal wired
network (Cat 5 or higher) can easly support 1Gb/sec.

WiFi, is generally limited to somewhat less, depending
on WiFi generation, distance from repeater, type of
configuration (e.g. mesh vs. single-point) and
endpoint capability.


--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On 9/27/2025 12:27 AM, Stephen Harker wrote:

Dimensional Traveler <dtravel@sonic.net> writes:

On 9/26/2025 12:21 PM, Scott Dorsey wrote:
Paul S Person <psperson@old.netcom.invalid> wrote:

But I have also seen offers for WiFi repeaters. The repeaters, of
course, take time to receive and then send the data; the
Ethernet-over-Power does not.

[...]

They have to be on the same leg. If you've got single-phase service

like most us, then all the devices need to be on either even-numbered
or odd-numbered breakers unless there is enough RF leakage between the
two legs. You can get a coupling device to create enough leakage if you >>> need it.
But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.
--scott

Or just don't use WiFi? I have a physical fiber-optic internet
connection and uploads & downloads just /SCREAM/. (But then I also
have an ISP that doesn't throttle my internet connection the way most
telecoms do.)

Does that mean you have installed a home optical network? I know some
people have. I went for a cat-6 network in my house and don't use WiFi (except on occasions for a visitor). Unfortunately I have HFC, not
fibre and it isless reliable, especially as they have electronics in the
path without backup power so a one point failure takes all downstream
out in the segment.

No "network", just a fiber optic connection from the ISP to a
modem/router with a wired connection from there to my computer. (Though
the modem/router can also connect via wire to a couple other computers
as well. I occasionally connect my work laptop's docking station to it.)
--
I've done good in this world. Now I'm tired and just want to be a cranky
dirty old man.
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On Fri, 26 Sep 2025 15:21:51 -0400 (EDT), kludge@panix.com (Scott
Dorsey) wrote:

Paul S Person <psperson@old.netcom.invalid> wrote:

But I have also seen offers for WiFi repeaters. The repeaters, of
course, take time to receive and then send the data; the >>Ethernet-over-Power does not.

[1] When I bought it this was phrased in the descriptions in such a
way that it sounded like both had to be on the same /circuit/, but
being in the same /breaker box/ is enough. Being on circuits in
/different/ breaker boxes, it appears, will not work.

They have to be on the same leg. If you've got single-phase service
like most us, then all the devices need to be on either even-numbered
or odd-numbered breakers unless there is enough RF leakage between the
two legs. You can get a coupling device to create enough leakage if you
need it.

As best I can tell, one is odd and one is even. And they are on
different sides as well.
The actual requirement that was causing confusion was that they /not/
be in separate breaker boxes, or perhaps that they not be joined only
at the power source for the building (in my case, a telephone pole
with a transformer on it). The former of which is, I suspect, more
likely to affect a business than it is a home. The latter might affect
a large industrial building.

But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.

These plug into the wall socket and provide a socket for whatever is
displaced precisely to minimize that problem, as least in comparison
with plugging into a surge protector. Apparently, plugging one into a
surge protector really does produce a lot of ... stuff. Something to
do with surge protectors working, in part, by not connecting the
outputs directly to the input.
The speed promised was 300Mbps (by comparison, my ISP promises "up to
940Mbps, and my various speed tests suggest download speed over WiFi
tops out at about 75Mbps, although a web-based test showed 47.5Mbps
just now). When I installed it, it's own software measured the speed
of the network /itself/ at 600Mbps, presumably because nothing else
much is using the circuits involved.
This leads, of course, to the question: if the actual WiFi speed at
the WiFi is about 75Mbps, and the machine hosting the music server
connects with WiFi, what does having 300Mbps, 600Mbps, 940Mbps, or 1,000,000Mbps matter? Does it affect how many items can be drawing on
the same network at the same time?
That said, keep in mind the limits of my useage here: one person, at
most five devices on WiFi (and one on Ethernet, over the power lines).
I am not putting any major strain on the system.
Still, a repeater WiFi box might work better for the OP. I have no
experience with them, but that they delay things a bit is apparently well-known. Whether the "bit" is something anyone would notice I have
no idea.
Or just going into the router and making sure the WiFi is set to put
out as strong a signal as possible. Although the installers here
installed the Gateway that way, presumably to minimize complaints.
--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On Sat, 27 Sep 2025 09:11:00 -0400 (EDT), kludge@panix.com (Scott
Dorsey) wrote:

Dimensional Traveler <dtravel@sonic.net> wrote:

On 9/26/2025 12:21 PM, Scott Dorsey wrote:

But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.

Or just don't use WiFi? I have a physical fiber-optic internet
connection and uploads & downloads just /SCREAM/. (But then I also have >>an ISP that doesn't throttle my internet connection the way most
telecoms do.)

Ethernet-over-power isn't wifi. It's much, much slower than wifi.
Also less secure. May or may not be less reliable.

You may be right -- if, say, AX can top 700Mbps.
It is possible that my adherence to b/g/n (which produced a WiFi that
worked as it had before, only faster) has blinded me to the true
speeds possible.
--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Paul S Person <psperson@old.netcom.invalid> writes:

On Fri, 26 Sep 2025 15:21:51 -0400 (EDT), kludge@panix.com (Scott
Dorsey) wrote:

Paul S Person <psperson@old.netcom.invalid> wrote:

But I have also seen offers for WiFi repeaters. The repeaters, of
course, take time to receive and then send the data; the >>>Ethernet-over-Power does not.

[1] When I bought it this was phrased in the descriptions in such a
way that it sounded like both had to be on the same /circuit/, but
being in the same /breaker box/ is enough. Being on circuits in >>>/different/ breaker boxes, it appears, will not work.

They have to be on the same leg. If you've got single-phase service
like most us, then all the devices need to be on either even-numbered
or odd-numbered breakers unless there is enough RF leakage between the
two legs. You can get a coupling device to create enough leakage if you >>need it.

As best I can tell, one is odd and one is even. And they are on
different sides as well.

Neither criteria is definitive, it really depends on the
equipment manufacturer.

A typical panel or subpanel enclosure supports both 120VAC
and 240VAC breakers on either side of the panel, consequently,
whether they are on different sides of the breaker panel is
not a characteristic that can be relied upon to determine which
leg of the 240V single-phase service a particular breaker
is served by.

Generally speaking, on each side, the single-width breaker
positions will alternate between the two current carrying
conductors.

An example busbar for a low-end breaker panel common in
residential use (I prefer the QO-series panels, personally).

https://www.amazon.com/Square-Homeline-SC1624M100-Replacement-Breaker/dp/B0758MXG5Y

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

scott@slp53.sl.home (Scott Lurndal) writes:

Stephen Harker <sjharker@aussiebroadband.com.au> writes:

Dimensional Traveler <dtravel@sonic.net> writes:

On 9/26/2025 12:21 PM, Scott Dorsey wrote:
Paul S Person <psperson@old.netcom.invalid> wrote:

But I have also seen offers for WiFi repeaters. The repeaters, of
course, take time to receive and then send the data; the
Ethernet-over-Power does not.

[...]

They have to be on the same leg. If you've got single-phase service >>>> like most us, then all the devices need to be on either even-numbered

or odd-numbered breakers unless there is enough RF leakage between the >>>> two legs. You can get a coupling device to create enough leakage if you >>>> need it.
But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.
--scott

Or just don't use WiFi? I have a physical fiber-optic internet
connection and uploads & downloads just /SCREAM/. (But then I also
have an ISP that doesn't throttle my internet connection the way most
telecoms do.)

Does that mean you have installed a home optical network?

That's not likely. The ISP connection may provide up to
1Gb/sec speeds (100M bytes/sec). Any internal wired
network (Cat 5 or higher) can easly support 1Gb/sec.

I was curious as I have seen a few discussions on a home optical network
which usually stop when they get a quote. Some appareently proceeded.
Some modem routers now come with higher network NICx 2.5 Gigabites being
more common on recent consumer equipment. The NICs are also expensive.
It can be a temptation, but the cost is currently prohibitive to most.
My printers have 100Mb/sec as does my old iBook G4 (usually I boot it
once a week to update Debian Linux) and my TV. I have several computers
with Gigabit speeds.

The cat-6 network was said to support Gb/sec up to some distance
(possibly 10 m), similarly for cat-5e, cat-5 was more questionable.
There were more weasel words for multi-Gb/sec speeds (up to 10 m I
think, which is fine for a house).

WiFi, is generally limited to somewhat less, depending
on WiFi generation, distance from repeater, type of
configuration (e.g. mesh vs. single-point) and
endpoint capability.

Yes, it is essentially a shared bandwidth, which is why wired
connections to fixed items is useful.
--
Stephen Harker sjharker@aussiebroadband.com.au
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Dimensional Traveler <dtravel@sonic.net> writes:

On 9/27/2025 12:27 AM, Stephen Harker wrote:

Dimensional Traveler <dtravel@sonic.net> writes:

On 9/26/2025 12:21 PM, Scott Dorsey wrote:
Paul S Person <psperson@old.netcom.invalid> wrote:

But I have also seen offers for WiFi repeaters. The repeaters, of
course, take time to receive and then send the data; the
Ethernet-over-Power does not.

[...]

They have to be on the same leg. If you've got single-phase service >>>> like most us, then all the devices need to be on either even-numbered

or odd-numbered breakers unless there is enough RF leakage between the >>>> two legs. You can get a coupling device to create enough leakage if you >>>> need it.
But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.
--scott

Or just don't use WiFi? I have a physical fiber-optic internet
connection and uploads & downloads just /SCREAM/. (But then I also
have an ISP that doesn't throttle my internet connection the way most
telecoms do.)

Does that mean you have installed a home optical network? I know
some
people have. I went for a cat-6 network in my house and don't use WiFi
(except on occasions for a visitor). Unfortunately I have HFC, not
fibre and it isless reliable, especially as they have electronics in the
path without backup power so a one point failure takes all downstream
out in the segment.

No "network", just a fiber optic connection from the ISP to a
modem/router with a wired connection from there to my computer.
(Though the modem/router can also connect via wire to a couple other computers as well. I occasionally connect my work laptop's docking
station to it.)

Fair enough, networks can be useful, but the use case is not always there.
--
Stephen Harker sjharker@aussiebroadband.com.au
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Stephen Harker <sjharker@aussiebroadband.com.au> wrote:

I was curious as I have seen a few discussions on a home optical network >which usually stop when they get a quote. Some appareently proceeded.
Some modem routers now come with higher network NICx 2.5 Gigabites being
more common on recent consumer equipment. The NICs are also expensive.
It can be a temptation, but the cost is currently prohibitive to most.
My printers have 100Mb/sec as does my old iBook G4 (usually I boot it
once a week to update Debian Linux) and my TV. I have several computers
with Gigabit speeds.

There's no real advantage in optical within a building any more since
you can just run 10G on ordinary Cat6E these days. So many folks have
optical to their home and then copper or even just wifi within the home.

But, I have a friend who has a barn and a shed and a detached garage
at some distance to his house, so he trenched fibre to two of the
outbuildings and used media converters to get gigabit service in them.
The third outbuilding had a conduit out to it for power so he used the
existing pull cord left from running the power to pull in a small
2-fibre cable (and of course a new pull cord as well).

Fibre is a big win for long distances and the ability to run fibre
alongside power lines is also a win.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

kludge@panix.com (Scott Dorsey) writes:

Stephen Harker <sjharker@aussiebroadband.com.au> wrote:

I was curious as I have seen a few discussions on a home optical network >>which usually stop when they get a quote. Some appareently proceeded.
Some modem routers now come with higher network NICx 2.5 Gigabites being >>more common on recent consumer equipment. The NICs are also expensive.
It can be a temptation, but the cost is currently prohibitive to most.
My printers have 100Mb/sec as does my old iBook G4 (usually I boot it
once a week to update Debian Linux) and my TV. I have several computers >>with Gigabit speeds.

There's no real advantage in optical within a building any more since
you can just run 10G on ordinary Cat6E these days. So many folks have >optical to their home and then copper or even just wifi within the home.

Indeed, and unless you live in a castle, even Cat5 will happily
support 1000BaseT, so long as it was competently installed
for the typical home-run lengths involved.

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Scott Lurndal <slp53@pacbell.net> wrote:

kludge@panix.com (Scott Dorsey) writes:

Stephen Harker <sjharker@aussiebroadband.com.au> wrote:

I was curious as I have seen a few discussions on a home optical network >>>which usually stop when they get a quote. Some appareently proceeded. >>>Some modem routers now come with higher network NICx 2.5 Gigabites being >>>more common on recent consumer equipment. The NICs are also expensive. >>>It can be a temptation, but the cost is currently prohibitive to most.
My printers have 100Mb/sec as does my old iBook G4 (usually I boot it >>>once a week to update Debian Linux) and my TV. I have several computers >>>with Gigabit speeds.

There's no real advantage in optical within a building any more since
you can just run 10G on ordinary Cat6E these days. So many folks have >>optical to their home and then copper or even just wifi within the home.

Indeed, and unless you live in a castle, even Cat5 will happily
support 1000BaseT, so long as it was competently installed
for the typical home-run lengths involved.

There was a day, though, not too long ago, when if you wanted more than 10Mb you needed fibre. We still have FDDI jacks in a lot of buildings at work.
They sit there unlamented and collecting dust, replaced by inexpensive copper cable.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Stephen Harker <sjharker@aussiebroadband.com.au> wrote:

I was curious as I have seen a few discussions on a home optical network >which usually stop when they get a quote. Some appareently proceeded.
Some modem routers now come with higher network NICx 2.5 Gigabites being
more common on recent consumer equipment. The NICs are also expensive.
It can be a temptation, but the cost is currently prohibitive to most.

Most new mid-range desktop motherboards comes with 2.5G NICs these
days, I suspect it adds something 10c to the BOM for the manufacturer.
Upper mid-range may have a 5G port, high-end 10G or multiple.

Likewise, many mid-range home routers have two 2.5G port (in and
"primary computer"), as you go up in the range usually all (1+4) ports
go 2.5G and then at the high end there may be some 10G ports too or
more internal ports (say 1+8).

Obviously discrete PCIe cards or switches cost more but we're still
only talking perhaps $20-$25 for a 2.5G card and $65 for a 5-port 2.5G
switch from well-known brands (prices from Newegg.com).

[...]

The cat-6 network was said to support Gb/sec up to some distance
(possibly 10 m), similarly for cat-5e, cat-5 was more questionable.
There were more weasel words for multi-Gb/sec speeds (up to 10 m I
think, which is fine for a house).

CAT-5E supports 2.5GBASE-T for at least 100m (and usually more). The
sales pitch for 2.5G was "faster without needing to upgrade wiring".
Often 5GBASE-T will run 50m+ on 5E cabling but I don't think this is
guaranteed unlike the other distances I mention.

CAT-6 supports 5G to at least 100m, 10G to at least 60m.
CAT-6A supports 10G to at least 100m.

TECHNICALLY all those numbers are for "structured cabling" where max
10m? can be flexible stranded cable (patch cables) with the remaining
length done using less flexible solid-core cables (for fixed
installation) and with IIRC max four? junctions but in practice none
of that tends to matter (it's just not guaranteed - but if it tests
fine it's OK).
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Paul S Person <psperson@old.netcom.invalid> wrote:

On Sat, 27 Sep 2025 09:11:00 -0400 (EDT), kludge@panix.com (Scott
Dorsey) wrote:

Dimensional Traveler <dtravel@sonic.net> wrote:

On 9/26/2025 12:21 PM, Scott Dorsey wrote:

But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.

Or just don't use WiFi? I have a physical fiber-optic internet >>>connection and uploads & downloads just /SCREAM/. (But then I also have >>>an ISP that doesn't throttle my internet connection the way most >>>telecoms do.)

Ethernet-over-power isn't wifi. It's much, much slower than wifi.
Also less secure. May or may not be less reliable.

You may be right -- if, say, AX can top 700Mbps.

On 2.4GHz or 5GHz?

There are 802.11ax aka Wifi 6 implementations that can do ~4.8Gbps
link speed over a single 5GHz radio to a single client. Both sides
needs a beefy 4x4 MU-MIMO antenna array, in practice going above 2-3
Gbps is unlikely since most consumer devices only have 3+3 antennas
and "ultralight" (or external antennas) devices may only have 2+2.

Maximum channel width is much smaller on 2.4GHz so the maxim link
speed there is ~1.1Gbps. So you'd need probably at least a 3x3
implementation to match that 700 Mbps using "mere" Wifi 6 over 2.4GHz.

The link speeds aren't THAT much higher than 802.11ac/Wifi 5 but in
practice the speeds tends to be way higher, in many cases Wifi 6 is
where it finally started utilizing most of the link speed properly. So
getting up to at least ax is really worthwhile.

All the above include the encoding overhead so the real maximum speeds
are 10-20% lower, this is similar to powerline which AFAIK also quote
numbers that include overhead BUT unlike wired which provides a much
more "real" value.

And obviously all these assumes good reception, so don't expect these
numbers through say a whole bunch of thick cement walls. But 1Gbps+ on
the 5GHz band can be doable with some walls in between with a good
client.

Wifi 6E adds the 6 GHz band which is less busy but throughput
limitations on the lower bands is unchanged. A few wifi cards already
support this band and it's also REALLY nice for wireless backhaul
(gives a nice mostly unused high-bandwidth link to the Access
Point/Mesh node if you can't get wired to the AP/node).

Wifi 7 adds new encoding which improves speeds on all three bands over
6/6E, lots of minor improvement and moving a lot of things from
"optional" to "mandatory".

It is possible that my adherence to b/g/n (which produced a WiFi that
worked as it had before, only faster) has blinded me to the true
speeds possible.

The first hit on Newegg for "wifi router" was a high-end model capable
of an aggregated max link speed of... 19Gbps! If you want to look at
details it was the ASUS RT-BE96U for $569.

Yes, that's over three radios on different frequency bands and include
overhead (see above). VERY few items is going to use multiple bands so
it's more relevant to look at each individual radio/band instead.

The max link rates for the radios are 1.5Gbps on 2.4GHz, 5.7Gbps on
5GHz and 11.5Gbps on 6GHz.

It can probably do at well 85%-90% of this with a recent Wifi 7 4x4
MU-MIMO card at clost range with no obstructions and not TOO much
noise on the channels it uses - which is not uncommon on the far less
used 6 GHz band.

Obviously it won't do that through multiple walls at a long distance
but it's certainly possible to multi-gigabit over either 5 or 6GHz on
the wrong side of several walls with a recent Wifi 7 laptop. And if
you're sitting in the same room? 9+ Gbps is definitely doable.

But the short version is, yes, 2+ Gbps is available on fairly
reasonable router/AP/mesh nodes - it only gets crazy expensive if you
want THE BEST.

The crazy part - I just realized that this wasn't even ASUS top-endc
consumer router!

The actual ASUS top-end router is the GT-BE98 - it has roughly similar
specs but with a fourth radio giving it two radios on the 5GHz band
where most clients will be for the near future.

It of course keeps the dual 10G (WAN & switch/"main computer") but
replace "lowly" 1G ports used for the second WAN port and additional
internal ports with 2.5G instead. So, it has a total of 25 Gbps
wireless (really 20+) and 62 Gbps wired (wired is full duplex, so the
two 10G ports alone gives 40G).

Does it have enough power to actuall route this? Probably, it's all in
hardware and those tends to run at wirespeed, the CPU is just there to
program the hardware and take occasional decision.

And the even crazier fact - all these are half of what the standard
permits, the Wifi specs allow 8x8 antenna configuration which would
allow Wifi6/6E link speeds of 9.6 Gbps and Wifi7 link speeds of 23
Gbps. Per radio.

But I don't anyone ever created silicone that support antenna
configuration larger than 4x4? - and both sides needs to have it for
it to be usable. Right now most seems to agree on 4x4 being high-end,
3x3 mainstream and 2x2 for "ultralight or low budget".
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On Mon, 29 Sep 2025 18:22:04 -0000 (UTC), Torbjorn Lindgren
<tl@none.invalid> wrote:

Paul S Person <psperson@old.netcom.invalid> wrote:

On Sat, 27 Sep 2025 09:11:00 -0400 (EDT), kludge@panix.com (Scott
Dorsey) wrote:

Dimensional Traveler <dtravel@sonic.net> wrote:

On 9/26/2025 12:21 PM, Scott Dorsey wrote:

But really, ethernet-over-power is such a bad idea just because it
produces so much RF trash. The throughput isn't much to write home
about either.

Or just don't use WiFi? I have a physical fiber-optic internet >>>>connection and uploads & downloads just /SCREAM/. (But then I also have >>>>an ISP that doesn't throttle my internet connection the way most >>>>telecoms do.)

Ethernet-over-power isn't wifi. It's much, much slower than wifi.
Also less secure. May or may not be less reliable.

You may be right -- if, say, AX can top 700Mbps.

On 2.4GHz or 5GHz?

On the 2.4GHz network. I turned off the 5GHz when I restricted the
2.4GHz network to b/g/n.
The problem solved was very strange: it is as if the network were
tuned to do QoS in reverse. Among other problems.
Indeed, the router's /best/ feature is that, because a phone company I
have used for decades now serves as my ISP as well, it cost nothing.
Can't argue with free, as long as it works. (The modem was also free.)
Thanks for the information I am snipping below. If I understand it
correctly, I could get 1.18Gbps out of my network with ax. This would
explain the interest in 1Gb fiber optic connections.
And, as I have noted, that might help with a larger household with
many devices on and in use simultaneously.
--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Paul S Person <psperson@old.netcom.invalid> wrote:

On the 2.4GHz network. I turned off the 5GHz when I restricted the
2.4GHz network to b/g/n.

Why did you turn off 5GHz? It's a serious win in most cases because
of the shorter range.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On Tue, 30 Sep 2025 18:52:42 -0400 (EDT), kludge@panix.com (Scott
Dorsey) wrote:

Paul S Person <psperson@old.netcom.invalid> wrote:

On the 2.4GHz network. I turned off the 5GHz when I restricted the
2.4GHz network to b/g/n.

Why did you turn off 5GHz? It's a serious win in most cases because
of the shorter range.

Because the WiFi network, as set up by the installers, DID NOT WORK
PROPERLY. As I noted in the part you snipped.
I know I have devices that could (or at least their replacements can)
do ax (and the other "a" items on 5Ghz). I do not know that I have any
that could (or can now) use that network. I have never turned it back
on to see if it pops up on any list of Wifi networks.
One of the other problems (besides the upside-down QOS) developed when
I tried connecting the HP Pavilion (which died and was replaced with
the Naviskauto for playing discs and the Chromebook for streaming)
with an Ethernet cable, as both were in the same general area.
Although I turned the WiFi off, there were problems that looked to me
as if the router were trying (and failing) to connect via WiFi. I
could have tried diabling the WiFi, but decided to just remove the
cable.
Some of this may be the router. When I check it out, it frequently
shows devices (such as a Kindle or the Fire HD6) as on Ethernet when
they have no Ethernet port and so cannot possibly be plugged in.
Apparently, it runs out of space on the WiFi page and then, rather
than add more slots for WiFi, puts the latecomers on the Ethernet
page. Or something -- I'm not sure there is any reasonable way to know
what it is doing.
Despite all this, my pokey b/g/n 2.4GHz works very well, very well
indeed.
--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Paul S Person <psperson@old.netcom.invalid> wrote:

On Tue, 30 Sep 2025 18:52:42 -0400 (EDT), kludge@panix.com (Scott
Dorsey) wrote:

Paul S Person <psperson@old.netcom.invalid> wrote:

On the 2.4GHz network. I turned off the 5GHz when I restricted the
2.4GHz network to b/g/n.

Why did you turn off 5GHz? It's a serious win in most cases because
of the shorter range.

Because the WiFi network, as set up by the installers, DID NOT WORK
PROPERLY. As I noted in the part you snipped.

So turning off the higher performing bands is going to make it work properly?
I think not. But scanning the band and seeing where your interference issues are and avoiding those channels might help something.

I know I have devices that could (or at least their replacements can)
do ax (and the other "a" items on 5Ghz). I do not know that I have any
that could (or can now) use that network. I have never turned it back
on to see if it pops up on any list of Wifi networks.

a is very very common. ax is not very common.

One of the other problems (besides the upside-down QOS) developed when
I tried connecting the HP Pavilion (which died and was replaced with
the Naviskauto for playing discs and the Chromebook for streaming)
with an Ethernet cable, as both were in the same general area.
Although I turned the WiFi off, there were problems that looked to me
as if the router were trying (and failing) to connect via WiFi. I
could have tried diabling the WiFi, but decided to just remove the
cable.

The router is trying to connect out to a computer via wifi? Huh?

Some of this may be the router. When I check it out, it frequently
shows devices (such as a Kindle or the Fire HD6) as on Ethernet when
they have no Ethernet port and so cannot possibly be plugged in.
Apparently, it runs out of space on the WiFi page and then, rather
than add more slots for WiFi, puts the latecomers on the Ethernet
page. Or something -- I'm not sure there is any reasonable way to know
what it is doing.=20

Have you considered using a proper AP instead of the AP built into the router if you're having so many issues? An AP from Ubiquiti won't cost you much. --scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On Sat, 27 Sep 2025 14:46:12 GMT, scott@slp53.sl.home (Scott Lurndal)
wrote:

That's not likely. The ISP connection may provide up to
1Gb/sec speeds (100M bytes/sec). Any internal wired
network (Cat 5 or higher) can easly support 1Gb/sec.

Cat 6 certainly can - though I've not seen Cat 5 getting that high.

(Been there done that did the upgrade)

WiFi, is generally limited to somewhat less, depending
on WiFi generation, distance from repeater, type of
configuration (e.g. mesh vs. single-point) and
endpoint capability.

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

The Horny Goat <lcraver@home.ca> writes:

On Sat, 27 Sep 2025 14:46:12 GMT, scott@slp53.sl.home (Scott Lurndal)
wrote:

That's not likely. The ISP connection may provide up to
1Gb/sec speeds (100M bytes/sec). Any internal wired
network (Cat 5 or higher) can easly support 1Gb/sec.

Cat 6 certainly can - though I've not seen Cat 5 getting that high.

I run my entire house on Cat5e using a 1Gbe rack-mount
managed switch and telcom patch panel.

Full bandwidth is available at any device with a 1Gb ethernet port.

Likewise, most of the company office wiring is 5e and it supports
1Gb just fine.

The previous owner wired with cat 3, I haven't actually
tried any of that yet to see how it would stand up to 1Gbe.
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

The Horny Goat <lcraver@home.ca> wrote:

On Sat, 27 Sep 2025 14:46:12 GMT, scott@slp53.sl.home (Scott Lurndal)
wrote:

That's not likely. The ISP connection may provide up to
1Gb/sec speeds (100M bytes/sec). Any internal wired
network (Cat 5 or higher) can easly support 1Gb/sec.

Cat 6 certainly can - though I've not seen Cat 5 getting that high.

1000BASE-T (IE "gigabit ethernet over copper") is DEFINED as as
running at an usable symbol rate of 1000 Mbps on Cat 5E (or higher), specifically for a minimum of 100m of which at least 90m must be solid
wire (less flexible, patch cables use stranded).

Yes, there's some overhead but it's not that hard to exceed 940 Mbps
after framing overhead, gap spacing, IP, TCP (or UDP) and protocol
overhead. You can get a bit closer to the raw speed if you run Jumbo
Frames (IE up to 9k per frame instead of 1500 bytes). The 1000 Mbps is
also bidirectional, you can also easily far exceed 800 in both directions simultaneously - it's less than 2x940 Mbps but not that much less.

The newer and faster 2.5GBASE-T has the exact same cable requirements,
just multiply all the numbers above with 2.5. As I mentioned elsewhere
that's the sales-pitch for 2.5G, use modern technology to run things
faster on existing category 5E cables that are in walls everywhere.
Basically they took the 10GBASE-T standard and run it at quarter the
speed.

Now, both sides DID say CAT 5, not 5e but the difference between CAT 5
and 5e is fairly small, basically tightening up the (far-end?)
crosstalk requirement - the standard committe found they needed this
to hit 100m with 1000BASE-T and that almost all CAT 5 installation
actually match the new criterias - not that there was many out there
at that point and those that couldn't, generally worked at slightly
reduced distances.

As a result the original CAT 5 specification actually ended up
deprecated in favor of 5e a few years later, I don't believe any of
the earlier or later cable specs ever got deprecated because, well,
they're all usable and common in a way CAT 5 never was.
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On Wed, 1 Oct 2025 12:57:35 -0400 (EDT), kludge@panix.com (Scott
Dorsey) wrote:

Paul S Person <psperson@old.netcom.invalid> wrote:

On Tue, 30 Sep 2025 18:52:42 -0400 (EDT), kludge@panix.com (Scott
Dorsey) wrote:

Paul S Person <psperson@old.netcom.invalid> wrote:

On the 2.4GHz network. I turned off the 5GHz when I restricted the >>>>2.4GHz network to b/g/n.

Why did you turn off 5GHz? It's a serious win in most cases because
of the shorter range.

Because the WiFi network, as set up by the installers, DID NOT WORK >>PROPERLY. As I noted in the part you snipped.

So turning off the higher performing bands is going to make it work properly? >I think not. But scanning the band and seeing where your interference issues >are and avoiding those channels might help something.

I turned it off in a deliberate attempt to get the WiFi set up as it
had been with DSL, and that worked so well I saw no reason to test
further.
This was not interference. This was being able to download a movie
very quickly but at the same time having my Web browser slow to a
crawl. Proper QOS gives priority to interactive tasks, like browsers,
not to background tasks, like downloading movies. As I said, it was as
if QOS was inverted.

I know I have devices that could (or at least their replacements can)
do ax (and the other "a" items on 5Ghz). I do not know that I have any
that could (or can now) use that network. I have never turned it back
on to see if it pops up on any list of Wifi networks.

a is very very common. ax is not very common.

I am working from memory here, and my memory says that ax is the
/only/ "a" type available at 2.4GHz. All the other "a" types are 5GHz
only. (If I got the one unique 2.4GHz type wrong, feel free to replace
"ax" with that "a" type -- the point remains).

One of the other problems (besides the upside-down QOS) developed when
I tried connecting the HP Pavilion (which died and was replaced with
the Naviskauto for playing discs and the Chromebook for streaming)
with an Ethernet cable, as both were in the same general area.
Although I turned the WiFi off, there were problems that looked to me
as if the router were trying (and failing) to connect via WiFi. I
could have tried diabling the WiFi, but decided to just remove the
cable.

The router is trying to connect out to a computer via wifi? Huh?

Which is why my diagnosis may be wrong and why disabling the WiFi
(rather than just turning it off) might have worked better -- if the
router was trying to wake up the WiFi and disabling it would prevent
it. Which is a lot of "ifs".
And also one of the reasons why I find the Gateway ... useable. But
only when restricted in what it can do.

Some of this may be the router. When I check it out, it frequently
shows devices (such as a Kindle or the Fire HD6) as on Ethernet when
they have no Ethernet port and so cannot possibly be plugged in. >>Apparently, it runs out of space on the WiFi page and then, rather
than add more slots for WiFi, puts the latecomers on the Ethernet
page. Or something -- I'm not sure there is any reasonable way to know
what it is doing.=20

Have you considered using a proper AP instead of the AP built into the router >if you're having so many issues? An AP from Ubiquiti won't cost you much.

The issues all disappeared when I shut down the 5GHz band and
restricted the WiFi to b/g/n. There is nothing to fix.
Well, normally: yesterday morning I found my Coolby Laptop (currently
working fine as a replacement for the HP Envy, which died a while
back) unable to connect to the internet.
I rebooted the modem and the router. I shut down the Coolby and then
booted it. I also verified that the Chromebook was connected to the
WiFi and the Internet. While trying to get to the window that would
allow me to manipulate the WiFi device itself, I ran across a "reset
WiFi" button and tried that.
Another reboot and re-entry of the WiFi magic number later, and I was
back up on the WiFi and the Internet.
So the usual nonsense still happens occasionally. But that has always
happened, from time to time, and always will.
Like Windows 11 creating a printer port USB002 out of thin air and
then deciding that my USB-to-LPT adapter was connected to it instead
of USB001, thus making the printer mysteriously unusable (ie, the
messages describing this, even in the log file, were not helpful):
once I finally got the Printers window and checked the ports,
switching the port to USB002 fixed the printing problem.
It's all just part of running Windows.
--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

scott@slp53.sl.home (Scott Lurndal) writes:

The Horny Goat <lcraver@home.ca> writes:

On Sat, 27 Sep 2025 14:46:12 GMT, scott@slp53.sl.home (Scott Lurndal) >>wrote:

That's not likely. The ISP connection may provide up to
1Gb/sec speeds (100M bytes/sec). Any internal wired
network (Cat 5 or higher) can easly support 1Gb/sec.

Cat 6 certainly can - though I've not seen Cat 5 getting that high.

I run my entire house on Cat5e using a 1Gbe rack-mount
managed switch and telcom patch panel.

Full bandwidth is available at any device with a 1Gb ethernet port.

Likewise, most of the company office wiring is 5e and it supports
1Gb just fine.

The previous owner wired with cat 3, I haven't actually
tried any of that yet to see how it would stand up to 1Gbe.

As mentioned, I have cat6 to around 21 ports, the 24 port switch is
guaranteed to provide 1GB/sec to all ports (which I have seen. Cat 5e
is said to work, but cat 5 I am unsure about.

I have worked in buildings with optical fibre to all equipment, but the
media converters in my area were limite to 100 MB/sec. It would be
expensive to replace them. The big gain was for security.
--
Stephen Harker sjharker@aussiebroadband.com.au
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

sjharker@aussiebroadband.com.au (Stephen Harker) writes:

scott@slp53.sl.home (Scott Lurndal) writes:

The Horny Goat <lcraver@home.ca> writes:

On Sat, 27 Sep 2025 14:46:12 GMT, scott@slp53.sl.home (Scott Lurndal) >>>wrote:

That's not likely. The ISP connection may provide up to
1Gb/sec speeds (100M bytes/sec). Any internal wired
network (Cat 5 or higher) can easly support 1Gb/sec.

Cat 6 certainly can - though I've not seen Cat 5 getting that high.

I run my entire house on Cat5e using a 1Gbe rack-mount
managed switch and telcom patch panel.

Full bandwidth is available at any device with a 1Gb ethernet port.

Likewise, most of the company office wiring is 5e and it supports
1Gb just fine.

The previous owner wired with cat 3, I haven't actually
tried any of that yet to see how it would stand up to 1Gbe.

As mentioned, I have cat6 to around 21 ports, the 24 port switch is >guaranteed to provide 1GB/sec to all ports (which I have seen. Cat 5e
is said to work, but cat 5 I am unsure about.

Most of my employer networking is 10Gbe with 200/400Gbe backbones.

But cat five is fine for less than 100m. It was only avialable
for a couple of years before 5e rolled out, so most cat 5
installations are really cat5e.


--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Scott Lurndal <slp53@pacbell.net> wrote:

But cat five is fine for less than 100m. It was only avialable
for a couple of years before 5e rolled out, so most cat 5
installations are really cat5e.

My house is wired with cat5, because not long after cat5e came out I
bought several thousand feet of Belden cat5 in the original boxes
for less than the scrap metal price, at a hamfest.

It's just fine at gigabit. I wouldn't want to try 2.5G, but gigabit
runs of under 200 ft. into an unmanaged Allied Telesis switch are all
just fine.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On Thu, 02 Oct 2025 10:07:09 -0700, Paul S Person
<psperson@old.netcom.invalid> wrote:

This was not interference. This was being able to download a movie
very quickly but at the same time having my Web browser slow to a
crawl. Proper QOS gives priority to interactive tasks, like browsers,
not to background tasks, like downloading movies. As I said, it was as
if QOS was inverted.

For those wondering how I determined this, here is what I found:
1. I was browsing with Edge on the HP Envy and Windows 10.
2. I was downloading with an app on the Fire HD 6, with a different OS
3. This pretty clearly prevents the problem from being on either
computer -- that is, it isn't as if Win10 were doing this; it has no
control over the HD 6.
4. When I paused the download, the browser picked right up and worked
as usual.
5. When I restarted the download, the browser's performance declined
again.
From which I inferred that the network was being as if QOS was
inverted.
And, while I searched thoroughly, I never found a QOS setting on the
Gateway that had any effect on this. The Gateway itself basically
asserted that QOS was set above its level.
What was really happening I, of course, have no idea.
Also, I regard networking as infrastructure: my priority and goal was
to make it /useable/, not to use the latest and greatest features as
such.
--
"Here lies the Tuscan poet Aretino,
Who evil spoke of everyone but God,
Giving as his excuse, 'I never knew him.'"
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On Thu, 02 Oct 2025 20:42:29 GMT, scott@slp53.sl.home (Scott Lurndal)
wrote:

sjharker@aussiebroadband.com.au (Stephen Harker) writes:

scott@slp53.sl.home (Scott Lurndal) writes:

The Horny Goat <lcraver@home.ca> writes:

On Sat, 27 Sep 2025 14:46:12 GMT, scott@slp53.sl.home (Scott Lurndal) >>>>wrote:

That's not likely. The ISP connection may provide up to
1Gb/sec speeds (100M bytes/sec). Any internal wired
network (Cat 5 or higher) can easly support 1Gb/sec.

Cat 6 certainly can - though I've not seen Cat 5 getting that high.

I run my entire house on Cat5e using a 1Gbe rack-mount
managed switch and telcom patch panel.

Full bandwidth is available at any device with a 1Gb ethernet port.

Likewise, most of the company office wiring is 5e and it supports
1Gb just fine.

The previous owner wired with cat 3, I haven't actually
tried any of that yet to see how it would stand up to 1Gbe.

As mentioned, I have cat6 to around 21 ports, the 24 port switch is >>guaranteed to provide 1GB/sec to all ports (which I have seen. Cat 5e
is said to work, but cat 5 I am unsure about.

Most of my employer networking is 10Gbe with 200/400Gbe backbones.

But cat five is fine for less than 100m. It was only avialable
for a couple of years before 5e rolled out, so most cat 5
installations are really cat5e.

Hmm. Before I got my new router about a year ago I never got over
750-800/sec

Interesting. (In the end my son - who is an electrical engineer for a
major Amazon contractor who builds warehouses for them - bought a
large roll of cable and made custom length cables for each computer in
our house)
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

The Horny Goat <lcraver@home.ca> writes:

On Thu, 02 Oct 2025 20:42:29 GMT, scott@slp53.sl.home (Scott Lurndal)
wrote:

sjharker@aussiebroadband.com.au (Stephen Harker) writes: >>>scott@slp53.sl.home (Scott Lurndal) writes:

The Horny Goat <lcraver@home.ca> writes:

On Sat, 27 Sep 2025 14:46:12 GMT, scott@slp53.sl.home (Scott Lurndal) >>>>>wrote:

That's not likely. The ISP connection may provide up to
1Gb/sec speeds (100M bytes/sec). Any internal wired
network (Cat 5 or higher) can easly support 1Gb/sec.

Cat 6 certainly can - though I've not seen Cat 5 getting that high.

I run my entire house on Cat5e using a 1Gbe rack-mount
managed switch and telcom patch panel.

Full bandwidth is available at any device with a 1Gb ethernet port.

Likewise, most of the company office wiring is 5e and it supports
1Gb just fine.

The previous owner wired with cat 3, I haven't actually
tried any of that yet to see how it would stand up to 1Gbe.

As mentioned, I have cat6 to around 21 ports, the 24 port switch is >>>guaranteed to provide 1GB/sec to all ports (which I have seen. Cat 5e >>>is said to work, but cat 5 I am unsure about.

Most of my employer networking is 10Gbe with 200/400Gbe backbones.

But cat five is fine for less than 100m. It was only avialable
for a couple of years before 5e rolled out, so most cat 5
installations are really cat5e.

Hmm. Before I got my new router about a year ago I never got over
750-800/sec

What is your unit of transfer in that estimate? 750Kbytes
would indicate a 10Mbit/sec speed.

All it takes is one 4-wire patch cable in the path to limit you to
10BaseT (10 Mbit/1Mbyte).

Interesting. (In the end my son - who is an electrical engineer for a
major Amazon contractor who builds warehouses for them - bought a
large roll of cable and made custom length cables for each computer in
our house)

What kind of cable? Cat5/5e? Stranded? Solid? How was
it wired; in-wall to a patch panel? Or did he crimp RJ45 connectors
on each end of the cable?

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

Scott Lurndal <slp53@pacbell.net> wrote:

The Horny Goat <lcraver@home.ca> writes:

Hmm. Before I got my new router about a year ago I never got over >>750-800/sec

What is your unit of transfer in that estimate? 750Kbytes
would indicate a 10Mbit/sec speed.

All it takes is one 4-wire patch cable in the path to limit you to
10BaseT (10 Mbit/1Mbyte).

100Base-TX (the common one, 100Mbp/12MBs) will run over any two pair
in an CAT 5 cable (yes, the standard requires CAT 5) Since it only
uses two pairs (one in each direction) you can run two 100Base-TX over
a single CAT 5 with splitters at both ends, BTDT...

There's other 100BASE-Tx standards that runs over different amount of
and quality cables but they're VANISHINGLY rare. Specifically
100BASE-T1 runs 15m over one full-duplex CAT 5 pair, 100BASE-T2 runs
100m over CAT3 by using two full-duplex pairs, 100BASE-T4 runs 100m
over CAT3 via 4 pairs, 2 in each direction.

The T1/T2 variants died because the use of full duplex made them
expensive (gigabit and up uses full duplex but by that time computing
had become cheap) while T4 died due to a lack of interest, there
wasn't enough "free" CAT 3 in building to make it sell over the
simpler TX. And people definitely did use splitters for 100BASE-TX
which further tilted it towards this.

Interesting. (In the end my son - who is an electrical engineer for a
major Amazon contractor who builds warehouses for them - bought a
large roll of cable and made custom length cables for each computer in
our house)

What kind of cable? Cat5/5e? Stranded? Solid? How was
it wired; in-wall to a patch panel? Or did he crimp RJ45 connectors
on each end of the cable?

Yeah, stranded and solid cable really do require different connectors!

There's "combo" connectors that works on both but they're way more
expensive and rare.

In practice most loose connectors are only rated for stranded (patch)
cables, most patch panels are only rated for solid cables because
that's how the usual "structured cabling" setup ends up using them.
Anything else is "specialty".

Also, as I mentioned elsewhere the standard techically says it's max
10m? stranded combined in the 100m total though in practice I've not
see issues due to this with modern cables/connector (they tend to far
exceed the minimum requirement).

Another pitfall I've seen lots of people fall into is, well, following
video tutorials too closely!

Those often show 5-10cm of each pair untwisted and then inserted into
the connector and if actually you do that, well, it'll probably run at 10MBASE-T or at BEST 100BASE-TX.

It's important to keep the untwisted portion short to avoid cross-talk overwhelming the signal, they often show long untwisted sections
because it's very hard to see/demonstrate if you don't do that! but
they often either "forget" to tell you that it can't actually be done
that way or doesn't emphasis it enough.

Manually crimping 1G/2.5G connectors isn't hard even with budget tools
and connectors if you know how to do it. 10G is much more demanding,
you need a good tool and good CAT6/6A connectors and the untwisted
section needs to really short. For 10G getting that part right will
probably be 90% of the time needed to do the connector unless you're
very experienced.
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

On Fri, 03 Oct 2025 22:04:26 GMT, scott@slp53.sl.home (Scott Lurndal)
wrote:

Hmm. Before I got my new router about a year ago I never got over >>750-800/sec

What is your unit of transfer in that estimate? 750Kbytes
would indicate a 10Mbit/sec speed.

All it takes is one 4-wire patch cable in the path to limit you to
10BaseT (10 Mbit/1Mbyte).

Just checked on my ISP's "speed test" and that's mbps (mega bits per
second)
--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: rec.arts.sf.written

The Horny Goat <lcraver@home.ca> writes:

On Fri, 03 Oct 2025 22:04:26 GMT, scott@slp53.sl.home (Scott Lurndal)
wrote:

Hmm. Before I got my new router about a year ago I never got over >>>750-800/sec

What is your unit of transfer in that estimate? 750Kbytes
would indicate a 10Mbit/sec speed.

All it takes is one 4-wire patch cable in the path to limit you to
10BaseT (10 Mbit/1Mbyte).

Just checked on my ISP's "speed test" and that's mbps (mega bits per
second)

Which is pretty much full speed for a 1Gb LAN connection, given
packet overhead et alia.
--- Synchronet 3.21a-Linux NewsLink 1.2