I bit the bullet & ran some network cables revently
to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?
TIA

--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

On Sunday, 3 May 2020 15:41:00 UTC+1, JimK wrote:
I bit the bullet & ran some network cables revently
to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?
TIA
Check that all eight conductors are present and connected. Some
patch cables only have two pairs fitted and will only operate at
100Mbit/s. Similarly, a broken conductor may cause the connection
to operate at 100Mbit/s.
John

On Sun, 3 May 2020 15:40:55 +0100 (GMT+01:00), Jimk
wrote:

I bit the bullet & ran some network cables revently to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?

Are these commercially made patch cables or home made?

A common mistake when making them at home is not putting the pairs on
the right pins and so the cables won't work over any distance.

Do you get any indication (LED (switch / device) or soft output 'Link
Speed, 1Gb') telling you you are running at Gb speeds and it's not
actually 100Mb (1/3rd of the theoretical max on yer PL adaptors)?

Cheers, T i m

Wrote in message:
On Sunday, 3 May 2020 15:41:00 UTC+1, JimK wrote:
I bit the bullet & ran some network cables revently
to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?
TIA
Check that all eight conductors are present and connected. Some
patch cables only have two pairs fitted and will only operate at
100Mbit/s. Similarly, a broken conductor may cause the connection
to operate at 100Mbit/s.
John


Any way to check that without taking them out again?
--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

On 03/05/2020 15:40, Jimk wrote:
I bit the bullet & ran some network cables revently
to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?

My guess was that you that the AV300 devices were getting real world
throughput of 200Mbps or better, and now you have gone to ethernet, but
for whatever reason it has fallen back to 100Mbps operation.

This would normally be as a result of a caballing issue - either not all
4 pairs properly connected, or, the pairs have been split incorrectly.

Start with a visual inspection of the connectors to verify the proper
wire ordering. Compare with:

https://en.wikipedia.org/wiki/TIA/EIA-568#Wiring

(the B variant is the more common I find)

If all look good, then buzz out each connection with a multimeter.

Note also that most network switches will give and indication of the
speed of the device conected to any port - either with a separate
"gigabit" LED, or changing the colour of the "connected" LED (typically
from amber for 100Mbps, to green for gigabit)


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

T i m Wrote in message:
On Sun, 3 May 2020 15:40:55 +0100 (GMT+01:00), Jimk
wrote:

I bit the bullet & ran some network cables revently to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?

Are these commercially made patch cables or home made?

A common mistake when making them at home is not putting the pairs on
the right pins and so the cables won't work over any distance.

Do you get any indication (LED (switch / device) or soft output 'Link
Speed, 1Gb') telling you you are running at Gb speeds and it's not
actually 100Mb (1/3rd of the theoretical max on yer PL adaptors)?

Cheers, T i m


Patch cables are purchased.

Only thing I can see on the (Netgear) switch is the port LEDs
appear blink faster when it's on the powerline adaptors.


I haven't delved into the switch just plugged it in & it worked....
--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

On Sun, 3 May 2020 15:55:58 +0100 (GMT+01:00), Jimk wrote:

Any way to check that without taking them out again?

One of the cheapo ethernet cable testers will. Around a fiver from
CPC/eBay/Amazon etc(*). Or just loop back each pair (ie one wire
connected to the other wire of the same pair) and check for
continuity at the other end. In theory with the kit powered off and
network cable plugged in there ought to be continity through the
isolating transmormers in each pair. Never tried that so don't know
what to expect.

You mention a coupler and changing that made difference. Are the
plugs and coupler making proper contact on all 8 wires? No iffy plug
end?

Not quite sure of the topography. Is this cable effectively an up
link from the switch that the Pi's are connected to? What is attached
to the upstream end of this uplink cable?

(*) Simple but basic for about £30 you can get rather more
sophisticated tester that shows which end, or how far away, a cable
fault is and any wiring errors. Search SC8108 loads about also can
have up to 8 remote dongles that are numbered that shows on the
tester for cable identification.

--
Cheers
Dave.

On Sun, 3 May 2020 15:40:55 +0100 (GMT+01:00), Jimk
wrote:

I bit the bullet & ran some network cables revently to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?
TIA

In addition to the advice about checking whether operating speed is
10, 100 or 1000 Mbit/sec, you should also check the interface counters
to see if the error rate is significant.

If you're running Linux, then "ifconfig" should give you something
like this - see the "RX errors" and "TX errors" lines:

ens33: flags=4163UP,BROADCAST,RUNNING,MULTICAST mtu 1500
inet 192.168.159.138 netmask 255.255.255.0 broadcast
192.168.159.255
inet6 fe80::20c:29ff:fe3f:43fb prefixlen 64 scopeid
0x20link
ether 00:0c:29:3f:43:fb txqueuelen 1000 (Ethernet)
RX packets 4088 bytes 523484 (511.2 KiB)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 3597 bytes 406309 (396.7 KiB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

Note that many (all?) modern linuxes has depreciated "ifconfig" for
the "ip" command. You can normally install the net-tools package to
obtain ifconfig.

"netstat -i" also shows the counters like this:

Kernel Interface table
Iface MTU RX-OK RX-ERR RX-DRP RX-OVR TX-OK TX-ERR TX-DRP
TX-OVR Flg
ens33 1500 4145 0 0 0 3653 0 0 0
BMRU
lo 65536 8395 0 0 0 8395 0 0 0
LRU

For network equipment, you might be able to use SNMP to read the
counters.

On Sun, 3 May 2020 16:35:21 +0100 (GMT+01:00), Jimk
wrote:

snip

Are these commercially made patch cables or home made?

A common mistake when making them at home is not putting the pairs on
the right pins and so the cables won't work over any distance.

Do you get any indication (LED (switch / device) or soft output 'Link
Speed, 1Gb') telling you you are running at Gb speeds and it's not
actually 100Mb (1/3rd of the theoretical max on yer PL adaptors)?


Patch cables are purchased.

Ok. Are they marked 'Cat5e' or similar OOI?

Only thing I can see on the (Netgear) switch is the port LEDs
appear blink faster when it's on the powerline adaptors.

Which would reflect your observations. Do you have a laptop you could
put on the remote end and see what it says re connection speed?

I haven't delved into the switch just plugged it in & it worked....

Understood, that is the norm with a basic (unmanaged) switch.

Cheers, T i m

Caecilius wrote:
Note that many (all?) modern linuxes has depreciated "ifconfig" for

It's "deprecated" :-)

The command 'ip addr' will produce output somewhat like that produced
by 'ifconfig' if you have ip but not ifconfig.

--
Chris Green
·

"Dave Liquorice" Wrote in message:
On Sun, 3 May 2020 15:55:58 +0100 (GMT+01:00), Jimk wrote:

Any way to check that without taking them out again?

One of the cheapo ethernet cable testers will. Around a fiver from
CPC/eBay/Amazon etc(*). Or just loop back each pair (ie one wire
connected to the other wire of the same pair) and check for
continuity at the other end. In theory with the kit powered off and
network cable plugged in there ought to be continity through the
isolating transmormers in each pair. Never tried that so don't know
what to expect.

You mention a coupler and changing that made difference. Are the
plugs and coupler making proper contact on all 8 wires? No iffy plug
end?

Not quite sure of the topography. Is this cable effectively an up
link from the switch that the Pi's are connected to? What is attached
to the upstream end of this uplink cable?

(*) Simple but basic for about £30 you can get rather more
sophisticated tester that shows which end, or how far away, a cable
fault is and any wiring errors. Search SC8108 loads about also can
have up to 8 remote dongles that are numbered that shows on the
tester for cable identification.


Well - through a blessedly brief process of elimination it was in
fact a dodgy CAT5 cable connector, luckily I had 4 more to do a
substitution with :-)

All up & running now.

Thanks to all
--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

T i m Wrote in message:
On Sun, 3 May 2020 16:35:21 +0100 (GMT+01:00), Jimk
wrote:

snip

Are these commercially made patch cables or home made?

A common mistake when making them at home is not putting the pairs on
the right pins and so the cables won't work over any distance.

Do you get any indication (LED (switch / device) or soft output 'Link
Speed, 1Gb') telling you you are running at Gb speeds and it's not
actually 100Mb (1/3rd of the theoretical max on yer PL adaptors)?


Patch cables are purchased.

Ok. Are they marked 'Cat5e' or similar OOI?

Only thing I can see on the (Netgear) switch is the port LEDs
appear blink faster when it's on the powerline adaptors.

Which would reflect your observations. Do you have a laptop you could
put on the remote end and see what it says re connection speed?

I haven't delved into the switch just plugged it in & it worked....

Understood, that is the norm with a basic (unmanaged) switch.

Cheers, T i m



Turns out dodgy cable connector - sorted now.

Cheers
--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

John Rumm Wrote in message:
On 03/05/2020 15:40, Jimk wrote:
I bit the bullet & ran some network cables revently
to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?

My guess was that you that the AV300 devices were getting real world
throughput of 200Mbps or better, and now you have gone to ethernet, but
for whatever reason it has fallen back to 100Mbps operation.

This would normally be as a result of a caballing issue - either not all
4 pairs properly connected, or, the pairs have been split incorrectly.

Start with a visual inspection of the connectors to verify the proper
wire ordering. Compare with:

https://en.wikipedia.org/wiki/TIA/EIA-568#Wiring

(the B variant is the more common I find)

If all look good, then buzz out each connection with a multimeter.

Note also that most network switches will give and indication of the
speed of the device conected to any port - either with a separate
"gigabit" LED, or changing the colour of the "connected" LED (typically
from amber for 100Mbps, to green for gigabit)



Now that would have slowed me down a bit!

Sorted - dodgy cable connector.

Cheers
--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

Caecilius Wrote in message:
On Sun, 3 May 2020 15:40:55 +0100 (GMT+01:00), Jimk
wrote:

I bit the bullet & ran some network cables revently to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?
TIA

In addition to the advice about checking whether operating speed is
10, 100 or 1000 Mbit/sec, you should also check the interface counters
to see if the error rate is significant.

If you're running Linux, then "ifconfig" should give you something
like this - see the "RX errors" and "TX errors" lines:

ens33: flags=4163UP,BROADCAST,RUNNING,MULTICAST mtu 1500
inet 192.168.159.138 netmask 255.255.255.0 broadcast
192.168.159.255
inet6 fe80::20c:29ff:fe3f:43fb prefixlen 64 scopeid
0x20link
ether 00:0c:29:3f:43:fb txqueuelen 1000 (Ethernet)
RX packets 4088 bytes 523484 (511.2 KiB)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 3597 bytes 406309 (396.7 KiB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

Note that many (all?) modern linuxes has depreciated "ifconfig" for
the "ip" command. You can normally install the net-tools package to
obtain ifconfig.

"netstat -i" also shows the counters like this:

Kernel Interface table
Iface MTU RX-OK RX-ERR RX-DRP RX-OVR TX-OK TX-ERR TX-DRP
TX-OVR Flg
ens33 1500 4145 0 0 0 3653 0 0 0
BMRU
lo 65536 8395 0 0 0 8395 0 0 0
LRU

For network equipment, you might be able to use SNMP to read the
counters.


Noted
Cheers
--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

Some network hardware not as tolerant perhaps, though the whole idea is that
the cable can almost be any bits of wire I assumed. Are we sure the two
joined cables are not faulty.
Brian

--
----- --
This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...

Blind user, so no pictures please
Note this Signature is meaningless.!
"Jimk" wrote in message
...
I bit the bullet & ran some network cables revently to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?
TIA

--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

Two people and a socket with movable link or two people and a very long
continuity tester lead!
Brian

--
----- --
This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...

Blind user, so no pictures please
Note this Signature is meaningless.!
"Jimk" wrote in message
o.uk...
Wrote in message:
On Sunday, 3 May 2020 15:41:00 UTC+1, JimK wrote:
I bit the bullet & ran some network cables revently
to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?
TIA
Check that all eight conductors are present and connected. Some
patch cables only have two pairs fitted and will only operate at
100Mbit/s. Similarly, a broken conductor may cause the connection
to operate at 100Mbit/s.
John


Any way to check that without taking them out again?
--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

On Sun, 3 May 2020 19:01:43 +0100 (GMT+01:00), Jimk
wrote:

T i m Wrote in message:
On Sun, 3 May 2020 16:35:21 +0100 (GMT+01:00), Jimk
wrote:

snip

Are these commercially made patch cables or home made?

A common mistake when making them at home is not putting the pairs on
the right pins and so the cables won't work over any distance.

Do you get any indication (LED (switch / device) or soft output 'Link
Speed, 1Gb') telling you you are running at Gb speeds and it's not
actually 100Mb (1/3rd of the theoretical max on yer PL adaptors)?


Patch cables are purchased.

Ok. Are they marked 'Cat5e' or similar OOI?

Only thing I can see on the (Netgear) switch is the port LEDs
appear blink faster when it's on the powerline adaptors.

Which would reflect your observations. Do you have a laptop you could
put on the remote end and see what it says re connection speed?

I haven't delved into the switch just plugged it in & it worked....

Understood, that is the norm with a basic (unmanaged) switch.


Turns out dodgy cable connector - sorted now.

Gdgd.

Do you intend replacing both leads with a single one at some time?
Personally I prefer to have a continuous piece of cable from end to
end (for the reason you have discovered).

Cheer, T i m

T i m Wrote in message:
On Sun, 3 May 2020 19:01:43 +0100 (GMT+01:00), Jimk
wrote:

T i m Wrote in message:
On Sun, 3 May 2020 16:35:21 +0100 (GMT+01:00), Jimk
wrote:

snip

Are these commercially made patch cables or home made?

A common mistake when making them at home is not putting the pairs on
the right pins and so the cables won't work over any distance.

Do you get any indication (LED (switch / device) or soft output 'Link
Speed, 1Gb') telling you you are running at Gb speeds and it's not
actually 100Mb (1/3rd of the theoretical max on yer PL adaptors)?


Patch cables are purchased.

Ok. Are they marked 'Cat5e' or similar OOI?

Only thing I can see on the (Netgear) switch is the port LEDs
appear blink faster when it's on the powerline adaptors.

Which would reflect your observations. Do you have a laptop you could
put on the remote end and see what it says re connection speed?

I haven't delved into the switch just plugged it in & it worked....

Understood, that is the norm with a basic (unmanaged) switch.


Turns out dodgy cable connector - sorted now.

Gdgd.

Do you intend replacing both leads with a single one at some time?

See how we go.


Personally I prefer to have a continuous piece of cable from end to
end (for the reason you have discovered).

I've got 3 more connectors in stock...

--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

We are now.
Thx

"Brian Gaff \(Sofa 2\)" Wrote in message:
Some network hardware not as tolerant perhaps, though the whole idea is that
the cable can almost be any bits of wire I assumed. Are we sure the two
joined cables are not faulty.
Brian



--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

On 03/05/2020 21:37, Jimk wrote:
We are now.
Thx

"Brian Gaff \(Sofa 2\)" Wrote in message:
Some network hardware not as tolerant perhaps, though the whole idea is that
the cable can almost be any bits of wire I assumed. Are we sure the two
joined cables are not faulty.
Brian




Higher speed use more than one pair of wires in a cable so they have to
be matched in length - one pair cannot be significantly longer because
the number of twists are different for instance.

--
mailto : news {at} admac {dot} myzen {dot} co {dot} uk

On 03/05/2020 21:53:20, alan_m wrote:
On 03/05/2020 21:37, Jimk wrote:
We are now.
Thx

"Brian Gaff \(Sofa 2\)" Wrote in message:
Some network hardware not as tolerant perhaps, though the whole idea
is that
the cable can almost be any bits of wire I assumed. Are we sure the two
joined cables are not faulty.
Â* Brian

Higher speed use more than one pair of wires in a cable so they have to
be matched in length - one pair cannot be significantly longer because
the number of twists are different for instance.

I thought each pair was treated separately, and the 2b bytes on each
lane concatenated. Can you provide a link suggesting I am wrong?

That is aademic in this case, since these are supposedly actual cables made
for the job. I think that you really do have to have a large mismatch in
length and capacitance to make it really misbehave though I don't use the
faster speeds here myself.

Brian

--
----- --
This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...

Blind user, so no pictures please
Note this Signature is meaningless.!
"Fredxx" wrote in message
...
On 03/05/2020 21:53:20, alan_m wrote:
On 03/05/2020 21:37, Jimk wrote:
We are now.
Thx

"Brian Gaff \(Sofa 2\)" Wrote in message:
Some network hardware not as tolerant perhaps, though the whole idea is
that
the cable can almost be any bits of wire I assumed. Are we sure the two
joined cables are not faulty.
Brian

Higher speed use more than one pair of wires in a cable so they have to
be matched in length - one pair cannot be significantly longer because
the number of twists are different for instance.

I thought each pair was treated separately, and the 2b bytes on each lane
concatenated. Can you provide a link suggesting I am wrong?

Jimk wrote:
Wrote in message:
On Sunday, 3 May 2020 15:41:00 UTC+1, JimK wrote:
I bit the bullet & ran some network cables revently
to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?
TIA
Check that all eight conductors are present and connected. Some
patch cables only have two pairs fitted and will only operate at
100Mbit/s. Similarly, a broken conductor may cause the connection
to operate at 100Mbit/s.
John


Any way to check that without taking them out again?

You can buy a simple tester for a fairly modest amount.

The basic ones typically have a detachable dongle you stick at the remote
end and you plug the tester in at the other end. They check all 8
conductors are connected and not crossed over etc.

(You can get smart ones which check bandwidth etc but they cost serious
money- not the kind of thing a hobbyist would normally have.)





--
https://www.unitedway.org/our-impact...an-trafficking

On Sun, 3 May 2020 21:16:00 +0100 (GMT+01:00), Jimk
wrote:

snip

Turns out dodgy cable connector - sorted now.

Gdgd.

Do you intend replacing both leads with a single one at some time?

See how we go.

Hmmmm. ;-(


Personally I prefer to have a continuous piece of cable from end to
end (for the reason you have discovered).

I've got 3 more connectors in stock...

Yeahbut, when they go intermittent they (or the reason you are getting
'network issues') could be a PITA.

Cheers, T i m

On Sun, 3 May 2020 21:53:20 +0100, alan_m
wrote:

snip

Higher speed use more than one pair of wires in a cable so they have to
be matched in length - one pair cannot be significantly longer because
the number of twists are different for instance.

Wouldn't *every* (Cat5e Ethernet cable in this case) have different
lengths per per because each pair *is* twisted differently?

eg, Wouldn't the pair with the most twists be longer than one with
fewer?

Cheers, T i m

On Monday, 4 May 2020 10:22:38 UTC+1, T i m wrote:
On Sun, 3 May 2020 21:53:20 +0100, alan_m
wrote:

snip

Higher speed use more than one pair of wires in a cable so they have to
be matched in length - one pair cannot be significantly longer because
the number of twists are different for instance.

Wouldn't *every* (Cat5e Ethernet cable in this case) have different
lengths per per because each pair *is* twisted differently?

eg, Wouldn't the pair with the most twists be longer than one with
fewer?

Yes.

John

T i m Wrote in message:
On Sun, 3 May 2020 21:16:00 +0100 (GMT+01:00), Jimk
wrote:

snip

Turns out dodgy cable connector - sorted now.

Gdgd.

Do you intend replacing both leads with a single one at some time?

See how we go.

Hmmmm. ;-(


Personally I prefer to have a continuous piece of cable from end to
end (for the reason you have discovered).

I've got 3 more connectors in stock...

Yeahbut, when they go intermittent they (or the reason you are getting
'network issues') could be a PITA.

We will see.

The only difference I can see between what I now have & what you
are pushing, is one cable connector...

--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

On 03/05/2020 19:02, Jimk wrote:
John Rumm Wrote in message:
On 03/05/2020 15:40, Jimk wrote:
I bit the bullet & ran some network cables revently
to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?

My guess was that you that the AV300 devices were getting real world
throughput of 200Mbps or better, and now you have gone to ethernet, but
for whatever reason it has fallen back to 100Mbps operation.

This would normally be as a result of a caballing issue - either not all
4 pairs properly connected, or, the pairs have been split incorrectly.

Start with a visual inspection of the connectors to verify the proper
wire ordering. Compare with:

https://en.wikipedia.org/wiki/TIA/EIA-568#Wiring

(the B variant is the more common I find)

If all look good, then buzz out each connection with a multimeter.

Note also that most network switches will give and indication of the
speed of the device conected to any port - either with a separate
"gigabit" LED, or changing the colour of the "connected" LED (typically
from amber for 100Mbps, to green for gigabit)



Now that would have slowed me down a bit!

Sorted - dodgy cable connector.

Cheers


But the blinky binky leds should tell you what speed is running anyway.

My Netgear GS105 four port unmanaged switch has two green leds
per port :-

Left hand = 100M
Right hand = 10M
Both = Gigabit

Ditto everything else connected.

Only my samsung 22H5600 smart tv runs at 100M (according
to the GS105 leds), everything else is Gigabit

On 04/05/2020 10:58, wrote:
On Monday, 4 May 2020 10:22:38 UTC+1, T i m wrote:
On Sun, 3 May 2020 21:53:20 +0100, alan_m
wrote:

snip

Higher speed use more than one pair of wires in a cable so they have to
be matched in length - one pair cannot be significantly longer because
the number of twists are different for instance.

Wouldn't *every* (Cat5e Ethernet cable in this case) have different
lengths per per because each pair *is* twisted differently?

eg, Wouldn't the pair with the most twists be longer than one with
fewer?

Yes.

John


CAT5 is generally tough as old boots. I bought a 30 metre (might
have been 20) patch cable and some old CaT5e, about 50 metres, that
had been stripped out of an office and it looked like someone had
driven over it in parts. Only cost £3 though for both

The patch cable worked at gigabit (according to the indicator
leds), do I chopped it up and crimped my own connectors.

Ditto the solid core CAT5E which is now fed around my house. All
seems fine according to the leds.

Only CAT6 is a pain to install in a domestic situation because
of the strict bending limitations which *theoretically* make
installation into a domestic wall box impossible.

Jimk wrote:

T i m Wrote in message:
On Sun, 3 May 2020 21:16:00 +0100 (GMT+01:00), Jimk
wrote:

snip

Turns out dodgy cable connector - sorted now.

Gdgd.

Do you intend replacing both leads with a single one at some time?

See how we go.

Hmmmm. ;-(


Personally I prefer to have a continuous piece of cable from end to
end (for the reason you have discovered).

I've got 3 more connectors in stock...

Yeahbut, when they go intermittent they (or the reason you are getting
'network issues') could be a PITA.

We will see.

The only difference I can see between what I now have & what you
are pushing, is one cable connector...

And 2 x 8 contact pairs that are unattended and not wiped against each
other quite frequently by movement of the apparatus or when they happen
to be playing up. An unnecessary fault waiting to happen. Pre-wired
cat6 patch leads at least up to 30m long are freely available for not
enormously much.

--

Roger Hayter

On Mon, 4 May 2020 13:26:39 +0100 (GMT+01:00), Jimk
wrote:

snip

I've got 3 more connectors in stock...

Yeahbut, when they go intermittent they (or the reason you are getting
'network issues') could be a PITA.

We will see.

You will indeed. ;-)

The only difference I can see between what I now have & what you
are pushing,

'What I'm pushing'? ICGAF if your network has issues as it's not mine.
You were also the one asking for help and presumably, haven't been
installing networks as a part of your living for 30+ years?

is one cable connector...

Exactly. !00% too many 'cable connectors' in any network segment.

They might be considered acceptable as a temporary measure, as part of
a test or to get you out of trouble, but not something I believe 'most
people' (that know what they are doing) would want to leave there if
they didn't have to (for technical [1] and reliability issues).

This would also apply to using Powerline or WiFi instead of Ethernet
cable.

If the CCTV feed was important and you missed something crucial
because of a bad (and undesirable in the first place) connector, you
too might see the issue?

Cheers, T i m

[1] In any Ethernet cable length (or twisted pair cable in general) I
you should have the *minimum* of cable that is untwisted as possible,
as seen when it's going though a plug and socket. When you join two
cables together as you have you have at least *doubled* the quantity
of the amount of parallel cable in the link. I'm not saying that will
stop it working, just that it's sub-optimal.

On 04/05/2020 14:43, Roger Hayter wrote:
Jimk wrote:

T i m Wrote in message:
On Sun, 3 May 2020 21:16:00 +0100 (GMT+01:00), Jimk
wrote:

snip

Turns out dodgy cable connector - sorted now.

Gdgd.

Do you intend replacing both leads with a single one at some time?

See how we go.

Hmmmm. ;-(


Personally I prefer to have a continuous piece of cable from end to
end (for the reason you have discovered).

I've got 3 more connectors in stock...

Yeahbut, when they go intermittent they (or the reason you are getting
'network issues') could be a PITA.

We will see.

The only difference I can see between what I now have & what you
are pushing, is one cable connector...

And 2 x 8 contact pairs that are unattended and not wiped against each
other quite frequently by movement of the apparatus or when they happen
to be playing up. An unnecessary fault waiting to happen. Pre-wired
cat6 patch leads at least up to 30m long are freely available for not
enormously much.


Why would anyone need CAT6 in a domestic environment ?. Complete
overkill. CAT5E will do everything you need and is much more
flexible.

On Mon, 4 May 2020 14:43:18 +0100, (Roger Hayter)
wrote:

snip

The only difference I can see between what I now have & what you
are pushing, is one cable connector...

And 2 x 8 contact pairs that are unattended and not wiped against each
other quite frequently by movement of the apparatus or when they happen
to be playing up. An unnecessary fault waiting to happen. Pre-wired
cat6 patch leads at least up to 30m long are freely available for not
enormously much.

I think Jim thinks he's in the b-i-y group (bodge it yourself) and of
course, knows better than those of us who have still got the T shirt.
;-)

Cheers, T i m

(Roger Hayter) Wrote in message:
Jimk wrote:

T i m Wrote in message:
On Sun, 3 May 2020 21:16:00 +0100 (GMT+01:00), Jimk
wrote:

snip

Turns out dodgy cable connector - sorted now.

Gdgd.

Do you intend replacing both leads with a single one at some time?

See how we go.

Hmmmm. ;-(


Personally I prefer to have a continuous piece of cable from end to
end (for the reason you have discovered).

I've got 3 more connectors in stock...

Yeahbut, when they go intermittent they (or the reason you are getting
'network issues') could be a PITA.

We will see.

The only difference I can see between what I now have & what you
are pushing, is one cable connector...

And 2 x 8 contact pairs that are unattended and not wiped against each
other quite frequently by movement of the apparatus or when they happen
to be playing up. An unnecessary fault waiting to happen.

How long have I got then?
Less than the other cable ends that don't move either?


--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

T i m Wrote in message:
On Mon, 4 May 2020 13:26:39 +0100 (GMT+01:00), Jimk
wrote:

snip

I've got 3 more connectors in stock...

Yeahbut, when they go intermittent they (or the reason you are getting
'network issues') could be a PITA.

We will see.

You will indeed. ;-)

The only difference I can see between what I now have & what you
are pushing,

'What I'm pushing'? ICGAF if your network has issues as it's not mine.
You were also the one asking for help and presumably, haven't been
installing networks as a part of your living for 30+ years?

is one cable connector...

Exactly. !00% too many 'cable connectors' in any network segment.

They might be considered acceptable as a temporary measure, as part of
a test or to get you out of trouble, but not something I believe 'most
people' (that know what they are doing) would want to leave there if
they didn't have to (for technical [1] and reliability issues).

This would also apply to using Powerline or WiFi instead of Ethernet
cable.

If the CCTV feed was important and you missed something crucial
because of a bad (and undesirable in the first place) connector, you
too might see the issue?

Cheers, T i m

[1] In any Ethernet cable length (or twisted pair cable in general) I
you should have the *minimum* of cable that is untwisted as possible,
as seen when it's going though a plug and socket. When you join two
cables together as you have you have at least *doubled* the quantity
of the amount of parallel cable in the link. I'm not saying that will
stop it working, just that it's sub-optimal.


I'll do an experiment then!

Do calm down dear...
--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

On Mon, 4 May 2020 15:14:18 +0100 (GMT+01:00), Jimk
wrote:

snip

[1] In any Ethernet cable length (or twisted pair cable in general) I
you should have the *minimum* of cable that is untwisted as possible,
as seen when it's going though a plug and socket. When you join two
cables together as you have you have at least *doubled* the quantity
of the amount of parallel cable in the link. I'm not saying that will
stop it working, just that it's sub-optimal.


I'll do an experiment then!

Good for you.

Do calm down dear...

I'll calm down if you promise not to waste anyone else's time here
when your experiment turns out as some of us have predicted?

Cheers, T i m

Jimk wrote:

(Roger Hayter) Wrote in message:
Jimk wrote:

T i m Wrote in message:
On Sun, 3 May 2020 21:16:00 +0100 (GMT+01:00), Jimk
wrote:

snip

Turns out dodgy cable connector - sorted now.

Gdgd.

Do you intend replacing both leads with a single one at some time?

See how we go.

Hmmmm. ;-(


Personally I prefer to have a continuous piece of cable from end to
end (for the reason you have discovered).

I've got 3 more connectors in stock...

Yeahbut, when they go intermittent they (or the reason you are getting
'network issues') could be a PITA.

We will see.

The only difference I can see between what I now have & what you
are pushing, is one cable connector...

And 2 x 8 contact pairs that are unattended and not wiped against each
other quite frequently by movement of the apparatus or when they happen
to be playing up. An unnecessary fault waiting to happen.

How long have I got then?
Less than the other cable ends that don't move either?

Welll if you get problems with the link the first thing you're going to
do is waggle the accessible RJ45s, isn't it. Apart from any other
disturbance dusting or vacuuming etc etc.

--

Roger Hayter

(Roger Hayter) Wrote in message:
Jimk wrote:

(Roger Hayter) Wrote in message:
Jimk wrote:

T i m Wrote in message:
On Sun, 3 May 2020 21:16:00 +0100 (GMT+01:00), Jimk
wrote:

snip

Turns out dodgy cable connector - sorted now.

Gdgd.

Do you intend replacing both leads with a single one at some time?

See how we go.

Hmmmm. ;-(


Personally I prefer to have a continuous piece of cable from end to
end (for the reason you have discovered).

I've got 3 more connectors in stock...

Yeahbut, when they go intermittent they (or the reason you are getting
'network issues') could be a PITA.

We will see.

The only difference I can see between what I now have & what you
are pushing, is one cable connector...

And 2 x 8 contact pairs that are unattended and not wiped against each
other quite frequently by movement of the apparatus or when they happen
to be playing up. An unnecessary fault waiting to happen.

How long have I got then?
Less than the other cable ends that don't move either?

Welll if you get problems with the link the first thing you're going to
do is waggle the accessible RJ45s, isn't it. Apart from any other
disturbance dusting or vacuuming etc etc.


Shouldn't be a problem - as I predict.
--
Jimk


----Android NewsGroup Reader----
http://usenet.sinaapp.com/

Brian Reay wrote:
Jimk wrote:
Wrote in message:
On Sunday, 3 May 2020 15:41:00 UTC+1, JimK wrote:
I bit the bullet & ran some network cables revently
to get away
from using power line plugs.
I had some long patch cables in stock so threaded those through.

One through a gigabit switch, serves 2 raspberry pis running cctv .

When connected by dlink300av powerline adaptors the data
throughput seems higher than the new cable, with smoother
playback of recorded events & no lag.

Switch to the cable & it's very different & slower.... I was
hoping it would be at least as good if not quicker?

(The cable is in reality two CAT5 patch cables joined with a
straight through connector. Initially with a moderately priced
cat6 labelled connector it didn't work, with a cheapo but still
called cat6 connector it does, sort of, see above).

Next steps to resolve please?
TIA
Check that all eight conductors are present and connected. Some
patch cables only have two pairs fitted and will only operate at
100Mbit/s. Similarly, a broken conductor may cause the connection
to operate at 100Mbit/s.
John

Any way to check that without taking them out again?

You can buy a simple tester for a fairly modest amount.

The basic ones typically have a detachable dongle you stick at the remote
end and you plug the tester in at the other end. They check all 8
conductors are connected and not crossed over etc.

(You can get smart ones which check bandwidth etc but they cost serious
money- not the kind of thing a hobbyist would normally have.)

A very few computers, had a built-in tester. It used TDR
and could detect shorts and opens. And also give a
distance to the fault. Some Marvell GbE network chips
had this. The machine I got it on, had a dirty RJ45, with
solder flux deposits on one wire of the connector,
and I got to use the test feature right away. It could
tell me "one pair is open", because the pin could not
make contact.

To use the feature, you connect the computer to your
router, with a cable. Turn off the power to the router
(so it cannot interfere with the TDR operation). Then
run the test, and it'll tell you the status of the pairs
and the wire.

https://web.archive.org/web/20030320...hite_Paper.pdf

Paul