DComTalk.com

Long story..short version. Just before I started with my current
employer, they moved office locations. They hired a local contractor
to install the electrical and low voltage cabling (voice and data).
When I saw the patch panel terminations, I got nervous. Each cable in
the patch panel has the jacket of the cable peeled back the full legnth
of the panel and the indivdual pairs are bound tightly together.
pictures

http://geocities.yahoo.com/bhokey/misc/DSC03790.JPG
http://geocities.yahoo.com/bhokey/misc/DSC03791.JPG
http://geocities.yahoo.com/bhokey/misc/DSC03792.JPG

Are network seems notorious slow to me. We have 2 bonded T1's to the
internet and all 70 users here access a web based app to do almost all
their work. Ping times are alway below 100ms and when I check our
internet utilization we occassional max out but not for longer than 10
seconds at a time maybe once or twice a day. The Cisco switchess dont
see any CRC's and all the machine NICS and ports on the switch are
forced to 100/full.

Could the patch panel termination be casuing this slow down.

sorry...pictures are here.

http://www.geocities.com/bhokey/misc/DSC03790.JPG
http://www.geocities.com/bhokey/misc/DSC03791.JPG
http://www.geocities.com/bhokey/misc/DSC03792.JPG

Brad <brad.hokanson@gmail.com> wrote:

Are network seems notorious slow to me. We have 2 bonded T1's
to the internet and all 70 users here access a web based app
to do almost all their work. Ping times are alway below 100ms
and when I check our internet utilization we occassional max
out but not for longer than 10 seconds at a time maybe once or
twice a day. The Cisco switchess dont see any CRC's and all
the machine NICS and ports on the switch are forced to 100/full.

Could the patch panel termination be casuing this slow down.

Unlikely. Although jackets should not be that stripped,
I haven't heard of it causing trouble so long as the pairs
are kept twisted.

I would expect to see CRC errors from bad cabling.
You can test with `ttcp`. There may be some software
issue, like an overloaded DNS or NetBIOS running.

-- Robert

Thanks Robert....thats' the response I was hoping to hear....I really
didnt want to have to rerun all the cables....but if it were the cause
of my issue, I would without hesitation.

Brad

Brad <brad.hokanson@gmail.com> wrote:

Thanks Robert....thats' the response I was hoping to hear....I
really didnt want to have to rerun all the cables....but if it
were the cause of my issue, I would without hesitation.

Oh, you should bring someone in with a certification scanner
long before doing anything as tough as recabling everything.

-- Robert

Brad wrote:

Long story..short version. Just before I started with my current
employer, they moved office locations. They hired a local contractor
to install the electrical and low voltage cabling (voice and data).
When I saw the patch panel terminations, I got nervous. Each cable in
the patch panel has the jacket of the cable peeled back the full legnth
of the panel and the indivdual pairs are bound tightly together.
pictures

http://geocities.yahoo.com/bhokey/misc/DSC03790.JPG
http://geocities.yahoo.com/bhokey/misc/DSC03791.JPG
http://geocities.yahoo.com/bhokey/misc/DSC03792.JPG

Are network seems notorious slow to me. We have 2 bonded T1's to the
internet and all 70 users here access a web based app to do almost all
their work. Ping times are alway below 100ms and when I check our
internet utilization we occassional max out but not for longer than 10
seconds at a time maybe once or twice a day. The Cisco switchess dont
see any CRC's and all the machine NICS and ports on the switch are
forced to 100/full.

Could the patch panel termination be casuing this slow down.

Still I would call up that electrical conmtractor and tell them to do

the job right. Stripping the claddding away like is wrong. Its an
extra layer of protection, and them taking it off is pretty dumb in my
opinion. god only knows how their electrical installation was done????

Joe Perkowski

Actually...God isnt the only one who knows....I know....and found out
the hard way. We have a standardized cublice structure of about 5-7
workstations in a cubicle area. He set up 2-3 cubicles (15-21
machines) on a single circuit. Its a joke. The problem is that it
has been almost a year since we moved to this building....(I identifed
the potential cabling issues 2nd day I was there) and management has
yet to act on it. I am almost sure it is too late.

BTW. I have access to a Fluke DSP 4000 to certify cable and an Optiview
II and every one I tested passed fine.

James Knott wrote:

Perkowski wrote:


Still I would call up that electrical conmtractor and tell them to do
the job right. Stripping the claddding away like is wrong. Its an
extra layer of protection, and them taking it off is pretty dumb in my
opinion. god only knows how their electrical installation was done????


Actually, stripping back the sheath is common in telecom work. If the
wiring is fairly dense, the sheath takes up a lot of space. On some of the
systems I've worked on there were 224 pairs terminated on a panel. You
simply don't have room for sheaths in that sort of install.


Ya I know, but, Im talking about is 24 port patch panels...I never known

installers to do that on panels like that.

Perkowski wrote:

Still I would call up that electrical conmtractor and tell them to do
the job right. Stripping the claddding away like is wrong. Its an
extra layer of protection, and them taking it off is pretty dumb in my
opinion. god only knows how their electrical installation was done????

Actually, stripping back the sheath is common in telecom work. If the
wiring is fairly dense, the sheath takes up a lot of space. On some of the
systems I've worked on there were 224 pairs terminated on a panel. You
simply don't have room for sheaths in that sort of install.

Perkowski wrote:

Ya I know, but, Im talking about is 24 port patch panels...I never known
installers to do that on panels like that.

But, as the pictures illustrate, so long as the twists are maintained,
the jacket, in this case not insulated should have little effect being
absent?

jon wrote:

Perkowski wrote:

Ya I know, but, Im talking about is 24 port patch panels...I never
known installers to do that on panels like that.


But, as the pictures illustrate, so long as the twists are maintained,
the jacket, in this case not insulated should have little effect being
absent?

Understood...I still wouldnt do a job like that....I think anyone who
leaves a panel like needs to have their heads examined...There is a
reason why cable manufacturers put cladding on the pairs you know!

jon wrote:

Perkowski wrote:

Ya I know, but, Im talking about is 24 port patch panels...I never
known installers to do that on panels like that.


But, as the pictures illustrate, so long as the twists are maintained,
the jacket, in this case not insulated should have little effect being
absent?

Yes, the missing jacket will defiantly have some kind of a effect on
cable performance. Will it be enough to cause Ethernet performance?
Maybe not. But will it effect Cat5e performance, absolutely. If the
components in this install are Cat5e, I would expect these cables to
fail the test.

Justin <jtoddDELETETHIS5dot1@yahoo.com> writes:

jon wrote:
Perkowski wrote:

Ya I know, but, Im talking about is 24 port patch panels...I never
known installers to do that on panels like that.
But, as the pictures illustrate, so long as the twists are
maintained, the jacket, in this case not insulated should have
little effect being absent?

Yes, the missing jacket will defiantly have some kind of a effect on
cable performance.

The jacket is unshielded twisted pair wiring surrounding the wire pairs
is mostly just a mechanical protection. The outer jacket protects
the insulation of the individual pairs form mechanical damage
that cna happen when pulling cable to it's place and when cable is
is installed. The outer jacket also keeps the pairs mechanically
ordered in the way they should (nicely evenly grouped together).

The outer jacket has early no effect on the signal traveling on
the cable! If you have just the original wire pairs from the
cable, just as tightly together as they ar enormally on the cable,
without the outer jackets, cable will perform still well for the
signals traveling though it. No problems. It might be even hard
to measure the difference between cable that has outer jacket
or not.

When talking about single cable away from other cables,
it should not matter much is the cable jacketed or not
for the signals traveing through it!
When you pack more cables together, the outer jackets
keep some separation between the pairs form different cables,
thus giving better isolation between the cables than just
putting all the wire pairs from different cables tightly
together.
Other thing the jacket adds an extra electrical insulation
layer. There is somewhat limited voltage that the cable
can withstand between the wires on the pairs and between
pairs. When the cable has this insulating outer jacket,
the cable can withstand more voltage between individual
pairs and outside of the cable than between the individual
wires inside the cable.

Things are somwehat different when talking about shielded
cables (those that have braid/foil shield on the jacket).
But for unshielded cables, the outer jacket is mainly just
mechanical protection (agains mecanical damage and to
keep pairs arranged) and extra insulation layer!

Will it be enough to cause Ethernet performance? Maybe not.

Removing the outer jacket does not cause problems to
Ethernet performance as long as you keep the wire pairs
arranged together as they were with the jacket.
And you don't put the wire pairs from different cables
too near to each other (for example keep some mechanical
separation instead of jackets).

If you put the wire pairs from different cables without
outer jackets together, then the performance of the
cable to cable crosstalk is poorer (more crosstalk) than
with the installtion where outer jackets are in place.
This can have effect or not for the system performance.
This depends on how close to the perfomance limits
the wiring already was.

But will it effect Cat5e performance, absolutely. If the
components in this install are Cat5e, I would expect these cables to
fail the test.

I think that doing the installation where the jackets were
removed for few centimeters distance and all wire pairs packed
togehter as shown on earlier posted link, the individual
cables might well pass the CAT5e performance tests when tested
individually (if not anymore CAT5e, then most propably still
CAT5). The perfomance when measuring single cable should have
not changed radically if no considerable mecahnical damage is
not done (wire pairs from one cable keep grouped/teisted together
as they were inside jackets, the cable fasteners not overtightened
etc..). The only neasurement where I would expect noticeable
change is cable to cable crosstalk figures. Those are poorer
than with the installation that keeps the jackets on the wire
pairs as close to the terminating point as possible.

--
Tomi Engdahl (http://www.iki.fi/then/)
Take a look at my electronics web links and documents at
http://www.epanorama.net/

Tomi Holger Engdahl wrote:

[SNIP]

But will it effect Cat5e performance, absolutely. If the
components in this install are Cat5e, I would expect these cables to
fail the test.


I think that doing the installation where the jackets were
removed for few centimeters distance and all wire pairs packed
togehter as shown on earlier posted link, the individual
cables might well pass the CAT5e performance tests when tested
individually (if not anymore CAT5e, then most propably still
CAT5). The perfomance when measuring single cable should have
not changed radically if no considerable mecahnical damage is
not done (wire pairs from one cable keep grouped/teisted together
as they were inside jackets, the cable fasteners not overtightened
etc..). The only neasurement where I would expect noticeable
change is cable to cable crosstalk figures. Those are poorer
than with the installation that keeps the jackets on the wire
pairs as close to the terminating point as possible.

I don't doubt that the cables shown in the pictures would pass Cat5.

I've tested some very sketchy cables and they passed Cat5. But that's
another story.
Yes, crosstalk is definatly where I have seen problems with this type of
install. And in the real world you're not going to see a single cable
install. Also, this single point would probably not be enough to fail a
Cat5e certification, but seeing this type of install would make me
question the rest of the installation and make me wonder at what other
point may there issues. So the cumulative effect would be a failed link.

Robert Redelmeier <redelm@ev1.net.invalid> wrote:

Brad <brad.hokanson@gmail.com> wrote:
Are network seems notorious slow to me. We have 2 bonded T1's
to the internet and all 70 users here access a web based app
to do almost all their work. Ping times are alway below 100ms
and when I check our internet utilization we occassional max
out but not for longer than 10 seconds at a time maybe once or
twice a day. The Cisco switchess dont see any CRC's and all
the machine NICS and ports on the switch are forced to 100/full.

Could the patch panel termination be casuing this slow down.

Unlikely. Although jackets should not be that stripped,

Without using the right tool, by the looks of things.

I haven't heard of it causing trouble so long as the pairs
are kept twisted.

And the cable ties are not too tight. Which is
another reason for leaving the jackets on.

--
Mark Evans
St. Peter's CofE Aided School
Phone: +44 1392 204764 X109
Fax: +44 1392 204763