Re: 2 Ethernet cabling question

[Gene Heskett]

On Sunday, December 26, 2010 11:22:47 am James McKenzie did opine:

>  On 12/25/10 9:23 PM, Tim wrote:
>  > On Sat, 2010-12-25 at 12:51 -0800, Joe Zeff wrote:
>  >> Part of the trouble-shooting was making sure there weren't any
>  >> cabling issues, so the client brought out an electrician.  Not only
>  >> weren't the colors on the cable standard, they were different at
>  >> each end! Our only guess was that there was a splice down in the
>  >> pipe because whoever'd run the cable had run out of one batch and
>  >> simply spliced on another.
>  > 
>  > Our phone lines do that, thanks to a plumber putting his oxy-acetelene
>  > torch through the wires a bit near a pipe he cut through.
>  
>  If he was using oxy-acetylene, he did not know what he was doing or he
>  was working with welded iron pipe.  Most plumbers use propane or
>  oxy-propane as it cannot overheat copper piping.
>  
>  >    He could have
>  > 
>  > lifted the wires away, but he didn't.  And some of the pairs were so
>  > badly corroded, that it was impossible to re-splice them, no matter
>  > how far back we stripped the insulation.
>  
>  Replace the ENTIRE run.  NEVER splice, and I mean NEVER splice, LAN
>  cabling.  You might have crossover problems and you might not notice the
>  loss in bandwidth until you hit 20MB+ connection speeds and then things
>  get very 'interesting'.  Yes, it is a PITA to do so, but in the long run
>  you will be happier.
>  
>  (BTW, who ran your cables, they should not be near plumbing anyway.)
>  
>  James McKenzie 
It must be time for my annual diatribe about broadband cabling.  It may not 
look like it, but the characteristic impedance of a twisted pair is every 
bit as important for a piece of cat5/6 as it is for a length of coaxial 
cable.

Any disturbance in the spacing of the wires such as created by a splice, 
represents a change in impedance at the point of the splice, which in turn 
will cause part of the signal to be reflected back toward the source.  This 
_includes_ a cable with proper plugs on its end, being plugged into a 
splice block such as sold at the Shack.  The plug itself is a horribly bad 
bump when looked at with a time domain reflectometer, and we get by with it 
only because it is normally less than 1 cm from the signal transmitter, or 
the receiver on the other end.  Even that in a gigahertz circuit is 
important enough I have had to cut the plug off and redo the installation 
of another just because the wire was stripped back too far, leaving 1/2" of 
it loose out the back of the plug.

All of this stuff IS a transmission line, and must be treated as such if 
you want it to work at all, let alone well.

Anyone who considers splicing this stuff, really should have his punch down 
tool confiscated and be kept away from anyplace that allows him access to 
this cabling.  They have no concept of a wires impedance or propagation 
velocity.  And the fancier the diploma on their office wall, the harder it 
is to impress upon them that they do not know it all.

That said, I am constantly amazed that a 120 foot run of now well aged 
Belden brand bright blue cat5, with about 45 feet it it swinging in the 
wind between my back porch roof and the peak of the roof of a shop 
building, is now in its 7th year out in the elements without a dropped bit 
that I have been made aware of.  When I put it in, it was for bragging 
rights that I had the internet on my small milling machines control 
computer, fully expecting it to be destroyed before spring rolled around 
again. 

If it dies yet today, it doesn't owe me a thing and I will gladly string 
another piece.  But I won't splice it ever, its all one run from the switch 
about 6 feet away, to a small 4 port switch out in the shop building, which 
I put in so I could run another 100 feet back across the back yard to the 
garage my alu siding won't let a wireless connection work in.  Yeah, you 
could call my place wired...

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Dreams are free, but there's a small charge for alterations.

2 Ethernet cabling question

[Tim]

On Mon, 2010-12-27 at 21:13 -0600, Mikkel wrote:
>  One thing to be careful of is the shield - make sure it is only
>  grounded on one end. You are better off not grounding it, then
>  grounding it at both ends. 
Well, unless they buy the right connectors, they won't be able to easily
ground it, anyway.  Most connections sold are for UTP (un-shielded
twisted pair).


-- 
[tim@localhost ~]$ uname -r
2.6.27.25-78.2.56.fc9.i686

Don't send private replies to my address, the mailbox is ignored.  I
read messages from the public lists.

2 Ethernet cabling question

[Mikkel]

On 12/27/2010 01:40 AM, Tim wrote:
>  
>  As I recall, CAT3 only goes as far as 10 mbit per second.  And depending
>  on how bad your gear is at automatic speed negotiation, it may keep
>  trying to go full pelt, instead of sticking to 10 mbits.
>   
I believe it is only rated for 10 mbits. But for a short run, you
may get 100 mbits performance out of it. In this case, it wouldn't
hurt to try, and watch the stats from the connection. This is where
the difference between theory and practice come in. The specs say
that you can not have a run more then 100 meters long. While I do
not recommend it, I have had exceeded that a couple of times when
there was no choice, and no place to put a repeater in the run. (We
told the costumer what he needed, and that he was going to had
trouble from time to time with this setup... This is what he wanted.)

For a short run, say 50', the CAT-3 will meet the specs for a 328'
(100m) run of CAT-5. For that matter, some of the CAT-3 sold is the
80's actually meets CAT-5 standards. One thing to be careful of is
the shield - make sure it is only grounded on one end. You are
better off not grounding it, then grounding it at both ends.

Mikkel
-- 

  Do not meddle in the affairs of dragons,
for thou art crunchy and taste good with Ketchup!

2 Ethernet cabling question

[Tim]

Peter Boy:
> > Currently I use a wireless bridge, but the connection
> > between the devices often breaks and it will take up to
> > 15 minutes until they manage to reconnect. 

On Mon, 2010-12-27 at 02:07 -0800, S Mathias wrote:
>  have you ever tried on other frequency - channel? 

I'd second that suggestion.  

My wireless access point has an automatic channel selection option
that's supposed to automatically pick the best channel to use.  I don't
know how, but I'd suspect scan for busy channels, and avoid them.
However, it always picked the same channel, every time, the same one as
a neighbour's.

So, I manually forced a particular channel on it, instead, and that
improved things, no end.

This is a setting on the access point.  Generally, you don't need to do
anything on the client.  I think you'd only need to do that if you had
multiple access points, using the same SSI, as different nodes for the
same network.

-- 
[tim@localhost ~]$ uname -r
2.6.27.25-78.2.56.fc9.i686

Don't send private replies to my address, the mailbox is ignored.  I
read messages from the public lists.

2 Ethernet cabling question

[S Mathias]

have you ever tried on other frequency - channel?

--- On Mon, 12/27/10, Peter Boy <pboy(a)barkhof.uni-bremen.de&gt; wrote:

>  From: Peter Boy <pboy(a)barkhof.uni-bremen.de&gt;
>  Subject: Re: 2 Ethernet cabling question
>  To: "Community support for Fedora users" <users(a)lists.fedoraproject.org&gt;
>  Date: Monday, December 27, 2010, 2:09 AM
>  Hi,
>  
>  there seem to be cable experts here so I beg your pardon
>  for jumping in
>  and would like to ask a question I'm considering for some
>  time: 
>  
>  Do I have a chance to successfully reuse an existing Cat 3
>  cable (8
>  wires, installed about 1980 for the inhouse telefon system)
>  to bridge a
>  gap of about 15 meter between the new location of my DSL
>  router and the
>  central switch of my regular Cat 5 cabling (i.e.connect the
>  100 mbit tx
>  ports)? And what have I to expect if I give it a try?
>  
>  Currently I use a wireless bridge, but the connection
>  between the
>  devices often breaks and it will take up to 15 minutes
>  until they manage
>  to reconnect.
>  
>  The 8 wires a twisted pairwise and the cable is shielded by
>  some sort of
>  aluminium jacket. I don't need the full 100 mbit, the
>  connection is a 50
>  mbit VDSL, so half the usual 100 mbit would be great.
>  
>  I found there exists a 100 mbit (tx) T4 / T2 specification
>  specifically
>  for cat 3 cables, but none of the devices used in Europe
>  support it.
>  
>  
>  Any hints appreciated.
>  
>  Peter
>  
>  -- 
>  users mailing list
>  users(a)lists.fedoraproject.org
>  To unsubscribe or change subscription options:
>  https://admin.fedoraproject.org/mailman/listinfo/users
>  Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
>

2 Ethernet cabling question

[Tim]

On Mon, 2010-12-27 at 03:09 +0100, Peter Boy wrote:
>  Do I have a chance to successfully reuse an existing Cat 3 cable (8
>  wires, installed about 1980 for the inhouse telefon system) 
I'd be very careful about trying to re-use phone wiring, no matter what
category wiring it is.  Unless you know that the wiring is ONE cable,
with no branches to other wiring, there's a good chance that it's part
of a run that goes between more than just what you think the beginning
and end of it is.
>  
>  Currently I use a wireless bridge, but the connection between the
>  devices often breaks and it will take up to 15 minutes until they
>  manage to reconnect. 
You might try relocating it, so it can transmit through a clearer path,
rather than through obstacles.  Even if that means buying a few metres
of patch leads so you can mount bridges in different places.

But I've nearly always found cabling more reliable that wireless, so I'd
buy cabling to do a direct patch between where you want.  About the only
time I find wireless to be an advantage is between different buildings,
where you get serious mains wiring voltage differences between them.
>  
>  The 8 wires a twisted pairwise and the cable is shielded by some sort
>  of aluminium jacket. 
STP - shielded twisted pair

>  I don't need the full 100 mbit, the connection is a 50
>  mbit VDSL, so half the usual 100 mbit would be great. 
As I recall, CAT3 only goes as far as 10 mbit per second.  And depending
on how bad your gear is at automatic speed negotiation, it may keep
trying to go full pelt, instead of sticking to 10 mbits.

-- 
[tim@localhost ~]$ uname -r
2.6.27.25-78.2.56.fc9.i686

Don't send private replies to my address, the mailbox is ignored.  I
read messages from the public lists.

2 Ethernet cabling question

[Peter Boy]

Hi,

there seem to be cable experts here so I beg your pardon for jumping in
and would like to ask a question I'm considering for some time: 

Do I have a chance to successfully reuse an existing Cat 3 cable (8
wires, installed about 1980 for the inhouse telefon system) to bridge a
gap of about 15 meter between the new location of my DSL router and the
central switch of my regular Cat 5 cabling (i.e.connect the 100 mbit tx
ports)? And what have I to expect if I give it a try?

Currently I use a wireless bridge, but the connection between the
devices often breaks and it will take up to 15 minutes until they manage
to reconnect.

The 8 wires a twisted pairwise and the cable is shielded by some sort of
aluminium jacket. I don't need the full 100 mbit, the connection is a 50
mbit VDSL, so half the usual 100 mbit would be great.

I found there exists a 100 mbit (tx) T4 / T2 specification specifically
for cat 3 cables, but none of the devices used in Europe support it.


Any hints appreciated.

Peter