Fluke DSX-600 CableAnalyzer User manual

Copper Cabling
Troubleshooting Handbook

3
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
Why the need for advanced diagnostics? . . . . . . . . . . . . . . . . . 4
Troubleshooting Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Before you begin testing, you should verify the basics . . . . . . . . . . . 5
Link models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Permanent Link . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8
The automated DSX-600 diagnostics. . . . . . . . . . . . . . . . . . . . .9
Shield Integrity Test . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
HDTDR diagnostics. . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
HDTDX diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Causes of Cabling Faults . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Wiremap. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Length. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Delay/Skew . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Insertion Loss (Attenuation) . . . . . . . . . . . . . . . . . . . . . . . 19
NEXT and PSNEXT . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Return Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
ACR-F and PS ACR-F (older names: ELFEXT and PSELFEXT) . . . . . . . . 22
Resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Architects, Consultants and Designers pages . . . . . . . . . . . . . . . 23
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
CableAnalyzer Comparison Guide . . . . . . . . . . . . . . . . . . . . . 26

4
“Certication” is the process of comparing the transmission performance of an
installed cabling system to a standard using a dened method of measuring the
performance. Certication of the cabling system demonstrates component quality
and installation workmanship. It is typically a requirement to obtain the cabling
manufacturer’s warranty. Certication demands that the cabling links yield a “Passing”
result. Technicians must diagnose the failing links and after a corrective action has
been implemented they must retest to ensure the link meets the required transmission
performance. The total time to certify an installation does not only include making the
certication measurements but also documentation and troubleshooting.
Today’s cabling installation professionals must therefore know how to troubleshoot and
diagnose high-performance cabling systems.
Why the need for advanced diagnostics?
As newer high-performance cabling systems have been developed and deployed, every
aspect of the installation demands a higher level of skill and more attention to detail.
New test parameters have been added. The links must be tested using one of two link
models, the Permanent Link or the Channel and the links are tested and evaluated
over a greater frequency range and with more data points. The components with which
these links are constructed need to deliver better performance and the quality of the
workmanship during installation must rise accordingly.
Due to the increased complexity of these cabling systems, determining the cause of
failure and quickly restoring suitable performance has become a challenging task.
This handbook guides you through the troubleshooting of advanced structured cabling
systems using the Fluke Networks DSX-600 CableAnalyzer™ so you can increase your
productivity and deliver better value to your organization.
Most common causes of failures in twisted pair cabling:
Introduction

5
1. Installation errors: failure to make proper connections while maintaining the wire
pairs and the twist rate in each pair (always keep the “original twist” in each wire
pair as much as practical)
2.Connectors that do not meet the required transmission quality
3.Incorrect tester set up
4.Defects or damage in the installed cable
5.Bad patch cords*
* Patch cords rate high on the list when we are discussing operating networks.
Certication is often executed using the Permanent Link Model since the actual patch
cords used by the operating network are not yet installed or available.
Before you begin testing, you should verify the basics:
• Has the correct test standard been selected? The certication test is executed as
an automated test or an “Autotest”. The test standard you select for an autotest
determines the link model (Permanent Link or Channel), the test parameters to
be measured and the frequency range over which the tests are executed and the
Pass/Fail criteria for each test.
• Has the correct link model been selected?
• Are you using the appropriate test adapter with a plug that matches the jack in the
telecommunication outlet (TO) or the patch panel?
• Has the test reference been set in the last 30 days? It is recommended to set the
reference on a regular basis and at a time that is easy to remember. (Such as at the
beginning of each testing session)
• Are you using the most current version of the tester software?
• Is the NVP set correctly for the cable under test? NVP plays a role when the tester
reports length or distance to a defect
• Is the tester within its operating temperature range and in calibration?
Remember that your Fluke Networks DSX-600 CableAnalyzer™ is a very accurate
instrument that measures small noise disturbances in cables. These instruments are
calibrated in the factory before shipping and this calibration should be veried every
12 months in an authorized Service Center. If the tester has been stored in a colder
or warmer place than the one in which you are working (i.e. overnight in a vehicle),
allow the unit to warm up to its steady-state operating temperature before setting
a reference or executing any measurements. This may take 10 to 15 minutes or more
depending on the temperature differential.
Troubleshooting Basics

6
The maximum allowed length of the Permanent Link is 90 meters. This link is most
often used by installers to prove the xed portion of the cabling system (excluding
user cords) meets the requirements of ANSI/TIA/EIA-568 or ISO/IEC 11801. The cabling
vendor offering the warranty will require this. If you see an installer making these
measurements using Channel adapters, you may wish to stop them and ask why they are
not making Permanent Link measurements.
Figure 1
To obtain meaningful results, it is essential to select the appropriate autotest and link
model.
Permanent Link
The Permanent Link performance is dened in such a way that after adding good patch
cords to a passing link, the channel performance is automatically met. By good patch
cords, we mean patch cords that pass the same class or category rating as the link or a
higher level of performance.
For this reason, it is recommended that new cabling installations are certied using
the Permanent Link model and test standard. Patch cords and equipment cords may be
changed many times during the life of the permanent link.
Link models

7
The Permanent Link shown in Figure 1 includes an optional Consolidation Point. Other
Permanent Link models may have a second patch panel instead of the TO which is
typically seen in a Data Center. The Permanent Link measurement includes the RJ45
plug at the end of the test lead, but does not include the cable portion of the test lead
or connection to the test equipment (See Figure 1). In the case of the DSX Permanent
Link adapters, the RJ45 plug is guaranteed to meet the NEXT, FEXT and Return Loss
requirements of the earlier mentioned standards. The cable portion of the test lead is
removed from the measurement using calibration constants stored in the adapter.
Another Permanent Link conguration that’s being seen more often nowadays is the
Modied Permanent Link. This is not recognized in TIA or ISO/IEC, but is supported in
ANSI/BICSI-004. The Modied Permanent Link is used to asses a horizontal link that
results in a direct connection to equipment like a PoE (Power over Ethernet) Wireless
Access point or CCTV camera. To assess this type of link the TIA Permanent Link test limit
must be used. The main tester must be connected via a Permanent Link adapter to the
patch panel. A Channel adapter on the remote unit at the remote end must be used to
connect the RJ45 plug of the direct connection.

8
Figure 2
Channel
Channel measurements are typically performed when restoring a service or to verify the
existing cabling for application support. The maximum length allowed for a Channel
is 100 meters. It is uncommon to perform Channel tests at the conclusion of a new
installation, since the patch cords that belong to each link are rarely available at that
time. Correct Channel measurements must cancel the effects of the mated connection
in the tester’s channel adapters as seen in Figure 2, which also shows the optional
consolidation point we saw in the Permanent Link. The same Data Center and Direct
equipment links also exist for the Channel.

9
When an Autotest fails or delivers a “marginal” pass result, the DSX-600 can use its
diagnostic capabilities to nd the cause of the marginal failures.
First, let us review what a marginal test is. The margin of a test is the difference between
the measured value and the applicable Pass/Fail limit value. The margin is positive when
the test passes, negative when the measurement fails and zero when the measured value
is equal to the limit value. A greater margin indicates that the result is further away
from the limit line. A larger positive margin therefore indicates a better test result. A
very small margin means that the test result is close to the limit value. A test result is
called marginal when its margin is less than the accuracy specication of the instrument
for the test parameter. For example, in gure 3, the Return Loss accuracy is +/- 1.79
dB @ 22.5 MHz and the worst case margin of a link at 22.5MHz was 0.4 dB. This Return
Loss result at 22.5 MHz is considered very close to the limit and is called a marginal test
result. A marginal PASS is indicated by PASS*. In this situation, the tester can be set
up to perform HDTDR and HDTDX tests showing diagnostics information to help identify
and point out what may have caused this marginal result. This information provides
the opportunity to quickly locate the problem, correct it, and deliver a link with good
performance. For more information, see the section that discusses HDTDR and HDTDX
diagnostics.
Figure 3
Shows result for worst pair where
the screen has been zoomed to show
the result in more detail. For this TIA
Cat 5e Permanent. Link Return Loss
accuracy is +/- 1.79 dB @ 22.5 MHz.
The margin was 0.4 dB so the result
has fallen within the accuracy band
and therefore is marked PASS*
The automated DSX-600 diagnostics

10
Figure 4
Shows a wire map failure on wires 6
and 7. Both are open at 1.4 m from
the near end and 4.7 m from the
far end, suggesting the cable has
been severed or not connected at
the same points. Inspect the link at
these points to identify the issue.
If the cabling link fails to pass the wire map test - the test which veries that all 8
wires connect the right pins at either end of the cable - the tester halts and presents a
window saying ‘Warning Bad Wire Map”. If you press OK, then the wire map diagram is
shown with the option to turn ‘SCAN ON’ which may help nd intermittent connection,
or ‘FIX LATER’ which enables you to save the results then go to then next link to be
tested. Alternatively, you can ‘CONTINUE’ which continues the test with the bad wire
map, this may allow you to nd other issues to help you troubleshoot the problem in the
link. Figure 4 shows such a wire map failure. The DSX-600 main unit is always assumed
at the left hand side of these screens, i.e. the near end. The wire map diagram shows
Wire 6 of pair 3/6 and Wire 7 of pair 7/8, both are open at 1.4 Meters at the near end
and 4.7 Meters from the far end. The software halts and asks the operator whether or
not to continue the test. It often makes more sense to tackle the wire map error before
continuing the test. The open wire causes the results for some test parameters to fail. In
this case we continued and the DSX shows a FAIL. At this point we could inspect the link
at 1.4M from the near end to discover the cause of the failure. We could also select the
DIAGNOSTICS tab which enables us to see the results of the HDTDR or HDTDX results.

11
Shield Integrity Test
Wire Map testers use DC Resistance measurements to ensure the shield is correctly
terminated and presented in a cabling system. In Data Centers, where the shield is
typically grounded at both ends, any wire map tester will report the shield as connected,
even if one of the ends is not. That’s because the shield test is a simple DC Resistance
measurement. If the shield is open, it will nd the remote unit using the building ground
(earth) or the shields of other cables. See Figure 5. This can result in a false PASS and
poor Alien Crosstalk performance.
Figure 5
Alternate electrical paths, such as through the building ground (earth) can
cause some testers incorrectly report the shield is properly connected

12
Advanced troubleshooting diagnostics
The DSX-600 diagnostics discussed above represent a distillation of a more complex
analysis of the test results data. In this section we will go into more detail about the
diagnostic analysis of link failures. Knowledge of the advanced diagnostics techniques
may help you to distinguish cases in which the automated diagnostics fall short. The
basis of the tester’s ability to report the distance to a location along the link-under-
test where crosstalk is excessive or where Return Loss is excessive, is the conversion
of the test results data collected in the frequency domain to the time domain. The
DSX-600 executes this conversion using unique and patented digital signal processing
techniques. The data in the time domain is in turn converted to a prole of the measured
disturbance along the length of the link.
Figure 6
The DSX-600 CableAnalyzer can make
Shield integrity measurements using
TDR techniques that can prevent false
shield test passes.
Using TDR techniques, the DSX-600 can prevent false passes when checking shield
continuity in a data center with common grounds.
See Figure 6. (S = Shield).
The automated link diagnostics save time and costs over trial and error techniques that
usually include the re-terminating of wiring and/or replacing of connecting hardware
in several locations in order to get a failing link to pass. The section on advanced
troubleshooting techniques will show you how you can obtain and interpret the
underlying diagnostic information generated by the tester’s analysis algorithms. To nd
out more about these diagnostic capabilities let’s have a look at the HDTDR and HDTDX
capabilities of the DSX-600.

13
Figure 7
Return loss failure with worst pair
7/8 reporting -2.0 dB margin @
4.9 MHz
The two parameters that provide the time domain information are HDTDX (High
Denition Time Domain Crosstalk) and HDTDR (High Denition Time Domain
Reectometry). As the name indicates, the HDTDX parameter shows the prole of
crosstalk happening along the link-under-test while HDTDR shows signal reections
along the link. Impedance changes cause signal reections that contribute to the
measured Return Loss value. If these reections become too big and the total amount of
energy reected exceeds the maximum amount allowed, the Return Loss test fails.
HDTDR diagnostics
In gure 7, we see a Return Loss failure at the remote end, so how do we go about
identifying the problem? First we identify the worst pair for return loss. In this example
the worst pair is pair 7/8. The worst pair result is 19 dB - that’s -2.0 dB Margin at 4.9
MHz. But if we examine the graph in more detail, we see that all the failures are below
50 MHz. This often suggests a cable issue.

14
In gure 9, we see another Return Loss failure, but this time from the main unit or
near end. If we examine the graph we nd that all four pairs fail and the failures are
especially bad at the lower frequencies. The worst Return Loss for all four pairs is at 5.4
dB with a really bad margin of -12.4 dB at 42MHz. This often suggests water in the cable.
Figure 9
Return Loss failures on all pairs at
the low frequencies.
Figure 8
HTDR showing Return loss issues
along the whole cable
Now let’s run the HDTDR diagnostics, the results of which are shown in gure 8. First
let’s nd and view the worst pair which was pair 7,8. Now let’s pinch and zoom on the
vertical axis to see the 1% horizontal grid lines in greater detail. Events should not
exceed 0.8% so we have drawn red lines on the diagram to make it easier to see which
and how many events pass the 0.8% lines. No more than four events maximum should
pass over the red lines. In the case we have many along the whole length of the cable
going in the positive and negative direction. This would suggest we have a cable issue
as we have many reective events all along the cable. This would mean we would have to
replace the cable.

15
Figure 10
The HTDR analysis shows Return Loss at
the beginning of the cable for all four
pairs, usually a characteristic of water
in the cable.
HDTDX diagnostics
In gure 11, we see a NEXT (near end crosstalk) failure. Once again we identify the worst
case pair. In this case is 3,6-4,5. If we examine the NEXT graph we see the worst case
NEXT is 30.0 dB with -5.5 dB margin at 245.5 MHz. So let’s do the HDTDX diagnostic test
and view the worst pair 3,6-4,5 and seen in gure 12.
So let’s make a HDTDR diagnostic measurement on this link to see if it can tell us where
the water is. We can see the results in gure 10. The results show most of the Return
Loss is happening within the rst six meters of the cable. The rst event where the
water begins is normally in the region of 6% to 15%. In this case, we would have to
replace the cable and solve the reason for the water getting in it.
Figure 11
NEXT worst pair shows a NEXT failure of
-5.5 dB margin @ 245.5MHz on
Pair 3,6 -4,5.

16
Figure 12
HDTDX showing connectors at the
start and middle of the link.
On analyzing the HDTDX trace we see two pulses, one at the beginning 0.0 meters, one
in the middle at 6.0 meters and the end of the cable is at 11 meters, this showing us
the connectors in the link. Connectors should ideally be no worse than 30% on links
less than 15 m (50 ft.). We have drawn a horizontal line on the HDTDX graph at 30% so
it’s easier to see where 30% has been exceeded. The rst connection clearly exceeds
30%. The pulse in the middle of the graph also exceeds 30%, so clearly shows another
connector in the middle of the link at 6.0 meters. This is not ideal! In addition, we see
some aberrations between the rst and second pulse. This indicates NEXT along the
cable, maybe the cable has been damaged. Events in the cable should ideally not exceed
10%. If they do, there should be no more than four of them. Any which way, it looks like
we should recommend replacing this whole link to eradicate the middle connector and
maybe improve the NEXT because of the potentially damaged cable.
In the next example, shown in gure 13, we see another NEXT failure. Once again
we look for the worst case NEXT margin failure pair. In this example it’s pair 3,6-4,5.
Typically it’s always 3,6-4,5 that exhibits the worst case NEXT, by the nature of the
construction of the cable.
Figure 13
Next failure at -6.0 dB Margin
@ 267.0 MHz on Pair 3,6 -4,5

17
This time we have a NEXT of 28.9 dB with a margin of -6.0 dB at 267.0 MHz. So to nd
out where this is happening along the cable, we need to run the HDTDX diagnostic test.
We see the results for this on gure 14. Again we view the worst pair 3,6-4,5 and we see
the far end connector far exceeds the ideal of 30%, it’s nearly 60%. So, we could re-
terminate the connector which may allow the link to pass. However, if we further analyze
the HDTDX trace, we also see that there are many aberrations along the cable that exceed
our ideal 10%. So the likely recommendation would be to change the cable which would
also mean remaking the far-end termination. This should now result in a NEXT PASS.
Figure 14
HDTDX shows crosstalk all along the cable

18
For each of the required TIA and ISO structured cabling measurements, you will nd
troubleshooting tips to help quickly pinpoint the cause of failures when they occur. In
some instances, you will nd suggested reasons why the measurement does not fail in
cases you would expect to see a failure.
Wiremap
Causes of cabling faults
Test Result Possible Cause of Result
Open • Wires broken by stress at connections
• Cables routed to wrong connection
• Wire is not punched down properly and does not make
contact in the IDC
• Damaged connector
• Cuts or breaks in cable
• Wires connected to wrong pins at connector or punch
block
• Application-specic cable (e.g. Ethernet using 1,2/3,6
only)
Short • Improper connector termination
• Damaged connector
• Conductive material stuck between pins at a connection
• Damage to cable
• Application specic cable (e.g. factory automation)
Reversed pair • Wires connected to wrong pins at connector or punch
block
Crossed pair • Wires connected to wrong pins at connector or punch
block
• Mix of 568A and 568B wiring standards (1,2 and 3,6
crossed)
• Crossover cables used (1,2 and 3,6 crossed)
Split pair • Wires connected to wrong pins at connector or punch
block

19
Test Result Possible Cause of Result
Length exceeds limits • Cable is too long – check for coiled service loops and
remove in this case
• NVP (nominal Velocity of Propogation) is set incorrectly
Length reported is
shorter than known
length
• Intermediate break in the cable
• NVP is set incorrectly
One or more pairs
signicantly shorter
• Damage to cable
• Bad connection
Length
Test Result Possible Cause of Result
Exceeds limits • Cable is too long – Propagation Delay
• Cable uses different insulation materials on different
pairs – Delay Skew
Delay/Skew
Test Result Possible Cause of Result
Exceeds limits • Excessive length
• Non-twisted or poor quality patch cables
• High impedance connections – Use time domain
techniques to troubleshoot
• Inappropriate cable category – e.g. Cat 3 in a Cat 5e
applications
• Incorrect autotest selected for cabling under test
Insertion Loss (Attenuation)
Note: Standard practices dictate that the cabling length be dened by the length of the shortest pair. NVP
varies per pair, meaning each pair could be reported with a different length. These two conditions can result in
a cable with three of four pairs over the length limit, yet the link result is a PASS (e.g. a channel with 101, 99,
103, 102 meters for the four pairs). A PASS in this case is the correct interpretation.

20
Test Result Possible Cause of Result
FAIL or PASS* • Poor twisting at connection points
• Poorly matched plug and jack (Category 6/Class E
applications)
• Incorrect link adapter (Cat 5 adapter for Cat 6 links)
• Poor quality patch cords
• Bad connectors
• Bad cable
• Split pairs
• Inappropriate use of couplers
• Excessive compression caused by plastic cable ties
• Excessive noise source adjacent to measurement
Unexpected PASS • Knots or kinks do not always cause NEXT failures,
especially on good cable and far removed from the ends
of the link
• Incorrect autotest selected (e.g. “Bad” Cat 6 link
tested to Cat 5 limits)
• “Fails” at low frequency on NEXT graph but passes
overall. When using the ISO/IEC standards, the so
called 4dB rule states all NEXT results measured while
insertion loss <4dB cannot fail
NEXT and PSNEXT