The Permanent Link

Document Sample
The Permanent Link Powered By Docstoc
					            The Change to Permanent Link in field testing.
                      What does this mean?

The Telecommunications Industry Association (TIA) is in the process of revising the main
standards document (TIA-568-A) that describes the design, installation and performance
requirements for telecommunications cabling systems for commercial buildings. The new
document or revision B of this standard will be designated TIA-568-B and will very likely be
approved in the early spring of 2001. This document will incorporate all of the
Telecommunications Systems Bulletins, referred to as TSB, that have been adopted since the
approval of TIA-568-A.
From the viewpoint of field test requirements, a very important change has been introduced that
will be incorporated in TIA-568-B. The TIA has adopted a new link configuration to be used to
verify the performance of the installed twisted-pair cabling links in the field. This new
configuration is called the permanent link; this configuration had been adopted earlier in the
international standards (ISO 11801 and other regional derivative standards documents). This
article describes the permanent link configuration, explains the reasons why this new concept
has been adopted and highlights the benefits of the permanent link test to both the installer and
end-user. The permanent link measurement furthermore has a significant impact on the
operations and performance requirements of field test equipment.

Channel, basic link and permanent link defined
The TIA and ISO standards call the completed cabling link over which the active network
equipment must communicate the channel. Figure 1a depicts the channel in its most generic
definition. This end-to-end link includes the equipment patch cords to connect the active
network devices in the telecommunications or equipment room, the cords to connect the
network devices in the office or work areas as well as the patch cords in an optional patch
Many people have correctly emphasized that the operation of the network relies on the
performance of the channel – the complete end-to-end link. The installation contractor for the
cabling system seldom if ever takes responsibility of the completed channel with all of its
equipment cords and patch cables to remain in place. Some percentage of the links installed
may have been planned for future expansion and may not be completed into channels any time
soon. It is generally recognized that the equipment cords and the patch cables are installed after
the fixed portion of the installation has been completed and tested. In practical use throughout
the life of the network cabling system equipment cords and patch cords may be exchanged
many times. Therefore, it is mandatory to have a means to certify that the fixed cabling
infrastructure meets a level of performance such that the completion of the channel with good
equipment and patch cords reliably delivers the desired channel performance. This is precisely
the reason why the industry standards defined performance criteria for the fixed cabling
infrastructure, which is typically the link configuration to be completed by the installation
contractors. The TIA standards initially adopted the basic link configuration for this purpose
while the ISO standards adopted the permanent link.

The Permanent Link                                                             Hugo Draye
                                       Page 1 of 8                          Fluke networks
Basic Link
Figure 1b depicts the basic link configuration. The basic link configuration is made up of a
maximum of 90 meters (328 ft) of uninterrupted solid-copper twisted-pair cable with a
termination connection on each end and with at each end a patch cord to connect the field
tester. The maximum length for each of the tester patch cables is 2 meter. The performance
limits for the basic link are based on this maximum length configuration and include the
anticipated contributions made by the two tester patch cables.

Permanent Link
Figure 1c illustrates the permanent link configuration. The permanent link is defined to meet the
same goal as the basic link, namely to verify the performance of the fixed portion of the installed
cabling. The permanent link configuration defines however that the contributions made by the
patch cables used to access the link-under-test are to be excluded from the measurement
results. The test limits for the permanent link are therefore different from those for the basic link
in the amount that is projected to be contributed by the tester patch cables. The goal of defining
the permanent link was to define a test configuration, which characterizes the permanently
installed, fixed portion of the cabling as accurately as possible.

   Place holder for schematic figure of channel, basic link and permanent link
   Figure 1 - Schematic representation of the different link models

The difference between these two test configurations may upon first inspection seem trivial but
the advent of new test parameters – primarily return loss as we shall discuss later – and of
higher performance requirements have accentuated the difficulties experienced in practice with
the basic link model. All the standards are now again converging after the predicted changes in
the TIA standards. The performance parameters for the basic link are no longer specified in
draft TIA/EIA-568-B for category 3 and 5e as well as in the draft addendum for category 6. The
consolidation point – marked CP in Figure 1a depicting the generic channel – had never been
incorporated in the basic link test configuration. As described earlier, the basic link assumes an
uninterrupted cabling link from one termination to the other. The permanent link model
incorporates the consolidation point. A consolidation point is often used in open office cabling
systems and is considered an element of the fixed wiring plant.

The effect of patch cords on the measurement results

Table 1 provides an overview of the test parameters defined for the high performance cabling
links that must meet either Cat 5e or the proposed Cat 6 performance levels. The table
distinguishes between the parameters that are directly measured and those that are derived by

The Permanent Link                                                                  Hugo Draye
                                          Page 2 of 8                            Fluke networks
calculation or combination of measured parameters. We will discuss the effect of the patch
cords on the parameters that are directly measured and the ways in which the test equipment
can adjust or adapt these measurements to comply with the permanent link model.

Table 1 - List of test parameters

Test Parameters                         Measured       Calculated
Wire map                                    
Propagation delay                           
Delay Skew                                                  
Length                                                      
Attenuation                                 
Near-End Crosstalk (NEXT)                   
Power Sum NEXT (PSNEXT)                                     
Attenuation to Crosstalk Ratio (ACR)                        
Power Sum ACR (PSACR)                                       
Far-End Crosstalk (FEXT)                    
Equal Level FEXT (ELFEXT)                                   
PSELFEXT                                                    
Return Loss (RL)                            

Wire Map
It is obvious that the wire pairs in the patch cords must be connected correctly to provide signal
continuity and to be able to verify any other transmission requirements. This test is not affected
by the transition from basic link to permanent link.

Propagation delay (length)
Length is derived from the propagation delay measurement – the measurement of the time
required for the signal to travel the length of the link. Length (distance) is calculated from this
time measurement by multiplying time and speed. The Nominal Velocity of Propagation (NVP)
value of the cable characterizes the speed. The current TIA standards (TSB-67) define the
maximum length of the basic link as 94 m plus an allowance of 10 percent for uncertainty
(variability) in the value of NVP. The nominal 94 m value includes the maximum length of 90 m
of the actual fixed link and the maximum 4 m accounting for the tester patch cords. When
selecting the permanent link configuration, the tester can make the simple correction for the
known length (propagation delay) of the tester patch cables that are shipped with the test tool.

The test limit formulas for the basic link in TSB-67 include the attenuation expected from the
tester patch cables. Attenuation is a linear function of length and a correction for the attenuation
contributed by the two 2 meter patch cables should nearly be perfect. This parameter will not
cause any difficulty in measuring the permanent link configuration.

The Permanent Link                                                               Hugo Draye
                                         Page 3 of 8                          Fluke networks
Two sets of parameters measure the crosstalk performance of a link: Near-End Crosstalk
(NEXT) between all of he pair combinations from both ends of the link and Far-End Crosstalk or
FEXT (also to be measured between all of the pair combinations from both ends of the link).
NEXT and FEXT occur simultaneously. NEXT is the crosstalk signal that “returns” on the
affected wire pair to the transmitter side while the FEXT component of the crosstalk on the
affected wire pair travels in the same direction as the transmitted signal to the far end of the link.
Figure 2 demonstrates the affect of NEXT on wire pair 2 from a signal traveling along wire pair
1. In the same figure, the
affect of the FEXT                                           Transmitted Signal
coupling between wire
pairs 3 and 4 is depicted.                 1                                                         1
As Figure 1c depicts, the
connections at the end of         NEXT 2                                                             2
the installed link are part
                                           3                                                         3
of the permanent link
model. The difference                      4                                                         4 FEXT
between the crosstalk
measurements of the
basic link model and the        Figure 2 – Pairs 1&2 demonstrate NEXT coupling; Pairs 3&4
permanent link model            demonstrate FEXT coupling in a twisted-pair link
amounts to the crosstalk
contribution of the patch cordage itself. The connections at the end of the tester patch cords
with the link-under-test are to be included in the results defined by either model. As with the
channel model, the end connections in the field tester units are not included in the basic link
model. These end connections are of course also not included in the permanent link model.
Current field tester designs use test link adapters that connect with the field test units by means
of a “high performance” connector. This connector establishes the reference plain for the “base
line accuracy” of the tester and typically offers very good crosstalk performance. How can we
limit and control the crosstalk contribution of the patch cord itself? Fluke Networks has
constructed these basic link test adapters using individually shielded twisted-pair patch cordage.
This cable construction is also referred to as “pair in metal foil” (PiMF) or SSTP (Shielded
shielded twisted pair). It is also the cable type that has been defined for the ISO Class F (Cat 7)
links. The SSTP cable construction delivers extremely good crosstalk performance. The
crosstalk of a short patch cable constructed with SSTP can clearly be ignored for a very minor
penalty in the accuracy of the NEXT and FEXT parameters of a permanent link. The variability
of the mated 8-pin modular connections between patch cable and link-under-test totally
overshadows the minute amount of crosstalk in 2 meters of SSTP cable. In summary, using
high performance SSTP test cords provides the desired accuracy for the permanent link NEXT
and FEXT tests.

Return Loss
All of the recent test standards require that Return Loss be measured and tested. It is now
generally recognized that Return Loss is a very important transmission characteristic. Return
Loss (RL) is a measure of the reflected energy – the echo signal – on each wire pair of a link.
This echo is caused by changes in impedance along the cabling link. Any change in impedance

The Permanent Link                                                             Hugo Draye
                                        Page 4 of 8                         Fluke networks
along the wire pair causes a reflection of some amount of signal energy back into the direction
of the transmitter.
Return Loss represents the total signal echo on a wire pair. As a noise source, Return Loss is
measured and evaluated to assure that the reflected signal energy is sufficiently small in
reference to the transmitted signal such that the reliability of the transmission is not in jeopardy.
Return Loss is an important characteristic for any transmission line because it may be
responsible for a significant noise component that hinders the ability of the receiver when the
data is extracted from the signal. Return Loss is especially critical for network applications that
utilize the wire pairs in a full duplex mode – simultaneous transmission in both directions over a
wire pair. Gigabit Ethernet or 1000BASE-T uses full duplex transmission over all wire pairs
simultaneously. All of the newer standards (TIA Cat 5e, ISO Class D 1999, TIA Cat 6 and ISO
Class E) certify the cabling link for full duplex transmissions.
Impedance changes or impedance anomalies occur for several reasons:
(1) imperfections within a twisted-pair cable (characterized by the structural return loss value of
    the cable itself)
(2) mismatches between cable of different construction (solid core copper for the horizontal run
    versus the stranded construction of patch cables and equipment cords) and, potentially,
    different models and brands
(3) mismatches between cable and connecting hardware
(4) untwisting of wire pairs at the connector termination even if the twist in the pairs is properly
    maintained; untwisting has a very significant impact on the Return Loss performance
(5) separating the wire pairs; the impedance of a wire pair in the cable is affected by the
    proximity of the other wire pairs and metallic conductors such as shields
(6) bundling of cables should be performed with devices that do not exert too much force on the
    cables in the bundle. Simple visual inspection can foretell problems especially when physical
    deformation can be observed.
Patch cords have a very significant influence on the Return Loss characteristics of a link.
Impedance changes or anomalies near the end of the link – near the transmitter source make a
significant contribution to the reflected energy measured as Return Loss. Signals traveling along
a transmission link are always subject to attenuation and, as mentioned earlier, attenuation is a
linear function of the distance the signal has traveled. That is to say, attenuation remains
constant along the link and the same amount of attenuation occurs every foot (or unit of length)
along the way. The signal will have endured twice the amount of attenuation at 20 feet
compared to the attenuation at the 10 foot mark.
If the signal encounters an impedance anomaly that is located only a few feet from the tester,
the signal has only undergone a minor amount of attenuation – it is still very strong. A small
reflection, say 3% of this signal bounces back and in turn arrives with very little attenuation at
the transmitter end. A reflection of this magnitude is almost certain to fail the return loss
measurement. On the other hand, if a similar anomaly is located 100 feet from the tester, the
signal arriving at the anomaly has already undergone a significant amount of attenuation. A 3%
reflection of this much weaker signal may not amount to anything measurable when it travels
the distance of 100 feet back to the tester (transmitter end).
The return loss characteristics of the patch cordage make a critical contribution to the measured
RL value for the link-under-test. The fact that patch cords have a significant impact on the return
loss of a link furthermore emphasizes the importance of using high quality patch cables when
operating the channel.

The Permanent Link                                                                Hugo Draye
                                         Page 5 of 8                           Fluke networks
                                                             This effect of patch cables on the
 A place holder for graphs                                   return loss measurement is the most
                                                             significant reason to adopt the
 Figure 3 - Return Loss degradation due to coiling and       permanent link model for testing the
 uncoiling of SSTP cable                                     installed cabling links because this
                                                             model prescribes that any contribution
                                                             made by the tester patch cables be
                                                             excluded. This fact is further
                                                             emphasized by the very undesirable
                                                             characteristic that the return loss
                                                             performance of any twisted-pair tester
                                                             patch cord continuously degrades due
                                                             to handling, coiling, storing and
                                                             uncoiling. No treatment or cure can
                                                             recover this degradation in
                                                             performance; it is a one way slope. A
                                                             laboratory experiment exemplifies this
degradation. Figure 3 shows the effect of coiling and uncoiling a tester interface adapter
constructed with quality SSTP flexible cable. The blue test results line in this figure shows that
the initial RL performance of the tester patch cord is better than 12 dB above the test limits for
the TIA Cat 6 basic link. This is a very good patch cord. After coiling and uncoiling this patch
cord 50 times the RL performance is depicted by the green line in Figure 3. The yellow line
shows the effect of yet another 50 times of coiling and uncoiling. After a total of 200
coiling/uncoiling operations the RL performance degraded to the performance depicted by the
red line. The “margin” of this patch cable over the basic link test limits is now reduced to 6 dB.
The accuracy of the return loss measurement is in serious jeopardy when the patch cable
“through” which the tester evaluates the return loss performance of the fixed installation is itself
a major source of significant reflections. Based on the appearance of many tester patch cords
that have been in the field for some time, we can attest to the fact that this laboratory
experiment treated the patch cables more gently than they are treated in the real world.
In summary, even very well constructed twisted-pair patch cables degrade over time with the
expected handling as test cords and aging patch cords seriously affect the accuracy of the
return loss measurements. The measurement accuracy is no longer a function of the accuracy
of the main tester units but is clearly affected and limited by the performance of the tester patch
cables. This is the primary reason for the adoption of the permanent link model. Clearly, the
challenge then is to device an access mechanism that contributes little or no return loss itself.
The next section describes a practical implementation to measure the permanent link model
with greatly enhanced accuracy.

The Permanent Link                                                               Hugo Draye
                                        Page 6 of 8                           Fluke networks
The solution
For practical field test operations some amount of flexible cable is needed to connect the test
tools to the link-under-test. For each of test parameters, the contribution made by the tester
                                                    access cords must be small and predictable
  Place holder for photo                            such that the measurement algorithms in the
                                                    test tools can implement an accurate
                                                    compensation. Return loss adds one very
                                                    critical aspect to these requirements namely
                                                    that of stability over time.
                                                     Fluke Networks is introducing an interface
  Figure 4 - The Permanent Link Interface            adapter that is constructed with a totally
  Adapter for the Fluke DSP-4X00                     different type of cabling as depicted in Figure
  CableAnalyzers                                     4. This flexible interface cable assembly is no
                                                     longer constructed with twisted-pair cabling
                                                     but with a proprietary cable design developed
                                                     to deliver the highest performance for testing
                                                     applications. This cable type very accurately
meets the nominal 100 Ω impedance requirements of the twisted-pair cabling system and the
network physical layer specifications. It actually meets this impedance better (within much
tighter tolerances) than most twisted-pair cables. But more importantly, this cable maintains its
impedance characteristics under normal handling procedure much better than any twisted-pair
cable. These properties deliver both a very high return loss performance as well as extremely
good stability over time. The user can now measure the permanent link as defined by the
revised industry standards with the highest level of accuracy, repeatability and confidence.
As the photo in Figure 4 details, an attachment or personality module is connected to this
permanent link interface adapter. The personality module contains the plug to connect with the
end of the installed link. This allows the user to replace the access plug with which the adapter
is plugged into the jack at the end of the fixed cabling link. Cat 6 plugs and jacks from a random
selection of connector manufacturers are not yet fully interchangeable. Several Cat 6
personality modules will be available to ensure that the NEXT and FEXT characteristics of the
connections between the adapter and the link-under-test are measured with maximum
Furthermore, in case the plug wears out, the adapter tip can be replaced easily and with modest
cost. It must be noted however, that users of field test equipment believe that the plug of an
adapter is the critical part subject to wear and tear. The reality instead is that bending and
twisting the twisted-pair patch cable is a far more serious but invisible mechanism of
degradation. This is caused by the reduced return loss performance in twisted-pair adapter
cables as was demonstrated with the data in Figure 3 from our laboratory experiment.
The personality modules can be replaced with special (optional) modules or artifacts that allow
the user to calibrate the test tool with its interface adapter. This calibration step extends the
base line accuracy of the tester to the location where the permanent link starts. As this interface
cable may have degraded even in minor ways over time, a simple calibration procedures
restores the accuracy with which the permanent link is measured. The field tester base line
accuracy or the accuracy at the high accuracy measurement port – the connector on the tester
itself – is no longer the relevant accuracy specification. As we have discussed in this paper, the
actual measurement results if taken with poor tester patch cables are inaccurate and corrupted.
Such test results could very easily produce a „fail‟ for an installed link that is perfectly fine. We
have certainly encountered this scenario for return loss measurements.

The Permanent Link                                                                Hugo Draye
                                         Page 7 of 8                           Fluke networks
The acceptance by the standards of the permanent link as the test configuration to be used to
verify the performance of twisted-pair cabling “as installed” forces field tester manufacturers to
design new test solutions. The test solutions that properly implement this test methodology
inherently yield much more accurate measurement and test results for the installed cabling
links, especially for return loss. In addition, to maintain a high degree of measurement accuracy
over time, a calibration method must be readily available that includes a verification of the
performance of the test tool with the interface adapter. It is highly desirable for the
manufacturers of field test tools to provide methods and artifacts that allow the user to verify the
performance status of the interface adapters and automatically make adjustments in the
measurement algorithms to maintain a high degree on measurement integrity.
The adoption of the permanent link test model and the implementation of a compliant field test
solution will provide end-users and installers with much more consistent and reliable test results.

The Permanent Link                                                               Hugo Draye
                                        Page 8 of 8                           Fluke networks

Shared By: