An Overview of the Electrical Validation of
`10BASE-T, 100BASE-TX, and 1000BASE-T
`Devices
`Application Note
`
`The number of devices that come
`with a built-in network interface card
`has risen steadily and will continue
`to rise as more and more digital
`entertainment devices with networking
`capabilities are designed and sold.
`Devices with network interface ports
`now range from personal computers
`to closed-circuit cameras. This is a far
`cry from the day when a 10-Mbit/s
`port could be found only on high-end
`servers and networking equipment.
`
`The technology used in these ports,
`commonly known as “LAN” or “NIC”
`ports, is usually one of the 10BASE-T,
`100BASE-TX, and 1000BASE-T
`standards or a combination of them.
`These standards transmit 10, 100
`or 1000 Mbit/s over UTP cable with
`an 8-pin RJ-45 connector. In this
`article, we will take a quick look
`at the electrical signals used in
`these technologies and how they
`can be probed for quick test and
`validation. This exploration will be
`useful for engineers involved in the
`electrical validation of the 10BASE-T,
`100BASE-TX, and 1000BASE-T
`implementations in their devices.
`
`10BASE-T
`The long-lived 10BASE-T standard
`has been around since 1990 and is
`showing no signs of going away,
`even though it is considered obsolete
`by many. It provides 10-Mbit/s data
`transmission over two pairs of a
`Category 3 or 5 cable, one pair for
`transmit and the other for receive.
`The other two pairs of the cable are
`unused.
`
`100BASE-TX
`100BASE-TX is the most widely
`used version of 100-Mbit/s Ethernet
`(also known as fast Ethernet) over
`UTP cable. It uses the same pairs as
`10BASE-T for transmit and receive but
`requires Category 5 or better cable.
`
`1000BASE-T
`1000BASE-T is the most common form
`of 1000-Mbit/s Ethernet (also known
`as Gigabit Ethernet) over UTP cable. It
`uses all four pairs of the UTP cable for
`both transmit and receive and requires
`Category 5e or better cable.
`
`Figure 1 and Table 1 below describe
`the pin assignment of the 8-pin RJ-45
`plug as used in a straight-through
`configuration.
`
`Figure 1. The 8-pin RJ-45 plug, also
`known as the 8P8C connector.
`
`Pin
`
`10BASE-T /
`100BASE-TX
`
`1000BASE-T
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`TD+
`
`TD–
`
`RD+
`
`Unused
`
`Unused
`
`RD–
`
`Unused
`
`Unused
`
`BI_DA+
`
`BI_DA–
`
`BI_DB+
`
`BI_DC+
`
`BI_DC–
`
`BI_DB–
`
`BI_DD+
`
`BI_DD–
`
`Table 1. The pin assignment for 10BASE-T,
`100BASE-TX, and 1000BASE-T on the
`8-pin RJ-45 plug in a straight-through
`configuration. TD/RD stands for transmit
`data/receive data. BI_Dx stands for
`bi-directional pair x.
`
`
`
`Page 1
`
` Dell Inc.
` Exhibit 1015
`
`

`
`Probing and Testing 10BASE-T Signals
`
`10BASE-T transmits a differential
`signal, and the most straightforward
`method to probe the signals is with
`the TD+ and TD– pins connected to
`a 100 Ω resistive load as shown in
`Figure 2a. In addition to the 100 Ω
`resistive load, the standard specifies
`two additional loads to be used for
`testing. These two additional loads
`are illustrated in Figure 3. Apart from
`the direct connection from the TD
`
`circuit to the load, the standard also
`describes the use of a “twisted-pair
`model” (also known as TPM in short).
`The TPM is an equivalent circuit that
`models the distortion introduced by
`a simplex link segment, and is made
`up of 4 segments of RLC circuitry
`not shown here. Tests for some of
`the 10BASE-T parameters are done
`iteratively with and without the TPM
`and on loads 1 and 2 including the
`
`100 Ω resistive load. This makes for a
`lot of tests!
`
`Let us take a look at the 10BASE-T
`waveforms. There are typically four
`different types of waveforms that
`need to be used for testing. All the
`waveforms in this article will be based
`on the circuit in Figure 2a with a
`100 Ω resistive load unless otherwise
`mentioned.
`
`+ –
`
`Vo
`
`Load
`
`+ –
`
`TD
`
`Figure 2a. The 10BASE-T TD circuit directly
`connected to the load. The output voltage Vo is
`measured across the load.
`
`+ –
`
`Vo
`
`Load
`
`Balun
`
`Twisted-Pair
`Model
`
`+ –
`
`TD
`
`Figure 2b. The 10BASE-T TD circuit connected to the load through the
`twisted-pair model (TPM).
`
`115Ω
`
`L=
`180µH
`
`76.8Ω
`
`L=
`220µH
`
`RP
`
`CP
`
`LS
`
`RS
`
`LOAD 1
`
`LOAD 2
`
`Figure 3. Loads 1 and 2, which are used to test 10BASE-T.
`
`2
`
`LS = L ± 1%
`RP ≥ 2 kΩ
`CP = 12 pF ± 20 %
`RS ≤ 0.5 Ω
`L DEFINITION
`
`
`
`Page 2
`
`

`
`Probing and Testing 10BASE-T Signals (continued)
`
`First is the LTP or link test pulse, also
`known as the NLP or normal link pulse.
`The LTP is the first signal transferred
`by the 10BASE-T transmitter and
`is used to indicate the presence of
`an active transmitter. If there is an
`active device at the end of the link, it
`responds with its own LTP. The LTP
`is also used in bursts to form data
`words where device capability data is
`exchanged during auto-negotiation. In
`all cases, the LTP has to fit within a
`defined template with all combinations
`of loads with and without the
`twisted-pair model.
`
`LTP without TPM
`
`LTP with TPM
`
`Figure 4. The link test pulse (LTP) waveform with and without the
`twisted-pair model.
`
`Figure 5. LTP signal with TPM in the LTP template.
`
`3
`
`
`
`Page 3
`
`

`
`Probing and Testing 10BASE-T Signals (continued)
`
`The next signal of interest is the
`TP _IDL signal. 1 OBASE-T data is
`transmitted in Manchester-encoded
`(transition indicates logical "1 ")
`data packets with a period of idle in
`between known as the interframe gap.
`The TP _IDL signal indicates the start
`of the idle period, and is therefore
`found at the end of each data packet.
`As with the LTP, the TP _IDL waveform
`also has to fit within a defined
`template with all combinations of
`loads with and without the
`twisted-pair model.
`
`...
`
`.
`
`4
`
`'
`
`~
`
`~
`
`..
`
`I ' '
`
`..
`
`.
`
`'
`
`•
`...
`
`'
`
`...
`
`~ .
`
`~
`
`-
`
`-
`
`'
`
`- -
`
`'
`
`'
`
`I
`
`o
`
`'
`
`'
`
`o
`
`'
`
`-
`
`..
`
`'
`
`'
`
`~
`
`-
`
`..
`
`.
`.. -
`
`-
`
`'
`
`'
`
`-
`
`...
`
`..
`
`'
`
`.•. ~±I ·'
`
`Figure 6.
`Manchester(cid:173)
`encoded random
`data packets.
`The waveform
`displayed in the
`lower half of the
`screen is the
`zoomed-in area
`contained in the
`white box on the
`waveform in the
`upper half.
`
`Figure 7. The
`TP _IDL is a
`positive-going
`pulse with a
`width of 300 ns
`or 350 ns
`depending on
`whether the
`last bit was
`one or zero
`respectively.
`
`Figure 8.
`The TP _IDL
`template test.
`
`Page 4
`
`

`
`Probing and Testing 10BASE-T Signals (continued)
`
`The signaling rate for 10BASE-T
`is nominally 10 MHz. An all-1’s
`Manchester-encoded signal will
`result in a 10-MHz waveform. This
`all-1’s waveform is used to test
`that all harmonics measured on
`the transmitting circuit are at least
`27 dB below the fundamental. This
`is easily achieved, as most modern
`digitizing oscilloscopes come with
`an FFT function. Through the use of
`FFTs made with the Hanning window
`function for frequency accuracy, it is
`easy to measure the magnitude of the
`spectrum at 10 MHz and its harmonics.
`
`Apart from the template tests and the
`test for harmonic content, the other
`parameters that can be tested are
`the peak differential output voltage
`and common-mode voltage. These
`tests are performed with random
`data signals, as shown in Figure 6,
`and are relatively straightforward
`measurements.
`
`Figure 9. All-1’s Manchester-encoded signal.
`
`Figure 10. The oscilloscope display is divided into
`two, with the trace on the upper portion displaying
`the all-‘1’s Manchester-encoded signal. The trace
`on the lower portion uses the oscilloscope FFT
`function to measure harmonic content of the all-1’s
`Manchester-encoded signal. This example shows
`a marker at the fundamental frequency of 10 MHz
`and another marker on the third harmonic (30 MHz).
`The magnitude of the third harmonic shown here is
`–28.45 dB from the fundamental.
`
`5
`
`
`
`Page 5
`
`

`
`Testing 100BASE-TX
`
`100BASE-TX uses a line encoding
`scheme known as MLT-3 where the
`data is alternated through three
`voltage levels with a transition
`indicating a logical 1. The MLT-3
`line coding enables the use of less
`bandwidth than would be required by a
`different coding scheme, such as NRZ,
`for the same data rate. 100BASE-TX
`is tested using an MLT-3 coded idle
`pattern. On most devices, setting the
`speed setting to 100 Mbit/s mode
`will automatically cause the device to
`output an idle pattern, part of which is
`shown in Figure 11. This same pattern
`is used for all 100BASE-TX tests.
`
`The 100BASE-TX standard provides
`the use of an eye pattern template
`that can be used to perform a quick
`check on the output of a 100BASE-TX
`transmitter. Note that the use of
`the eye template should not replace
`thorough testing of 100BASE-TX
`parameters, but it can provide a
`reasonably good indication of the
`performance of a particular transmitter.
`
`6
`
`Figure 11. The MLT-3 coded idle pattern from a 100BASE-TX
`transmitter.
`
`Figure 12. The eye pattern template of a 100BASE-TX signal. The
`oscilloscope used to create the eye-pattern is synchronized to the
`recovered clock from the transmit waveform.
`
`
`
`Page 6
`
`

`
`Testing 1000BASE-T
`
`1 OOOBASE-T transmits data over all
`four pairs of the twisted-pair and uses
`a five-level pulse amplitude modulation
`called PAM5 for data transmission.
`Performing tests on normal data
`transmission similar to 1 OBASE-T
`or 100BASE-TX is non-trivial. thus
`the standard defines the use of four
`different test modes for testing. These
`are named test modes 1 - 4 and are
`set by writing to bits 13 to 15 of the
`1 OOOBASE-T control register (register
`9.13:15). Tests are performed over all
`four pairs of the transmitter.
`
`Test Mode 1 causes the transmitter
`to send out all five signal levels of
`the PAM5 coding. which are the
`"+2". "-2". "+1," "-1" symbols
`alternating with 127 "0" symbols.
`This is then followed by long strings
`(128) of the "+2" and "-2" symbols
`repeated twice. and ends with 1024
`"0" symbols. Various points in the
`waveform are then labeled from A
`toM (I is skipped) to mark points of
`interest for testing. Points A. B. C. and
`D corresponds to the "+2." "- 2." "+1 ."
`and "- 1" symbols respectively.
`
`Three tests are done using Test
`Mode 1. First. the peak voltages at
`points A, B. C. and D are measured.
`The voltages at points A and B are also
`compared to ensure their amplitudes
`are within 1%. These measurements
`are relatively straightforward to
`make. and it involves zooming in to
`the point of interest and making the
`measurement.
`
`Figure 13. One cycle of a Test Mode 1 waveform with the test
`points A to M labeled.
`
`7
`
`Page 7
`
`

`
`Testing 1000BASE-T (continued)
`
`The next tests are the template tests.
`Points A, B, C, D, F, and H have to
`fit in defined templates after going
`through a 2-MHz high-pass filter
`and being normalized according to
`specific rules described in sub-clause
`40.6.1.2.3 of IEEE Standard 802.3-
`2005. These steps can be achieved
`using currently available digitizing
`oscilloscopes as shown in Figures 14a,
`14b and 14c below, using the template
`measurement for point A as an
`example.
`
`Apart from the template tests, there is
`also a droop test on the long strings
`of the "+2" and "–2" symbols. The
`voltage droop is measured from point
`F (minimum point at the start of "–2"
`symbol string) to point G (500 ns
`after point F) as well as from point H
`(maximum point of the waveform as
`indicated in Figure 3) to J (500 ns after
`point H).
`
`Figure 14a. The digitizing oscilloscope
`in this example allows the use
`of functions to be performed on
`waveforms. In this illustration, a
`high-pass filter with its lower cutoff
`at 2 MHz is applied on the Channel 1
`source using Function 2.
`
`Figure 14b. Function 3 then uses a magnify function to
`normalize the filtered waveform from Function 2. The
`point A waveform is normalized by dividing by the peak
`voltage value of the waveform at A.
`
`Figure 14c. The filtered and normalized point A waveform is then
`aligned to best fit to the template.
`
`8
`
`
`
`Page 8
`
`

`
`Testing 1000BASE-T (continued)
`
`Next, we will take a look at test modes
`2 and 3. These test signals consist of
`alternating "+2" and "–2" symbols
`timed to a 125 MHz timing clock called
`the TX_TCLK. Test Modes 2 and 3
`differ in the timing source used. Test
`Mode 2 is called Master mode, and
`uses the device’s own transmit clock,
`while Test Mode 3 is called Slave
`mode and uses a recovered clock from
`data transmitted by a link partner in
`Master mode. To be able to test the
`jitter of a 1000BASE-T device, the
`TX_TCLK of the device should be
`available for probing. The data sheet
`of the 1000BASE-T device should
`describe which pins should be
`probed to access the TX_TCLK. The
`waveforms for test modes 2 and 3
`are essentially the same; an example
`waveform and a TX_TCLK is displayed
`in Figure 13. Jitter testing will not be
`covered in detail here as it is beyond
`the scope of this article.
`
`Test Mode 2/3
`
`TX_TCLK
`
`Figure 15. Test mode 2/3 waveform along with a 125MHz TX_TCLK.
`
`9
`
`
`
`Page 9
`
`

`
`Testing 1000BASE-T (continued)
`
`Figure 16. The Test Mode 4 waveform.
`
`The last test mode that we will look
`at is Test Mode 4. In Test Mode 4,
`the device outputs a 2047 symbol
`pseudo random bit pattern that is
`PAM5 encoded and then further
`filtered through a partial response
`filter. The resulting output from the
`filter is a 17-level signal, which can
`be seen in Figure 14. The Test Mode 4
`waveform gives an idea of what the
`waveform from an actively transmitting
`1000BASE-T device looks like. The Test
`Mode 4 output is used as the source to
`measure peak transmitter distortion.
`
`The distortion measurement is
`not a straightforward oscilloscope
`measurement; it requires the use of
`post-processing to calculate distortion.
`Sample MATLAB® code is provided
`in the standard as a guidance to
`calculate distortion. In simple terms,
`what the code does is to sample each
`of the 2047 symbols at an arbitrary
`phase clocked from the TX_TCLK. The
`code then examines each sample for
`distortion and reports the highest value
`as the peak distortion. This usually
`involves obtaining a large record
`containing more than a few cycles
`of the Test Mode 4 waveform. This
`waveform record is then transferred
`to a personal computer and processed
`based on the guidance given in the
`MATLAB code.
`
`As in 10BASE-T where the standard
`calls for some tests to be run with a
`twisted-pair model, the 1000BASE-T
`standard also calls for the Test Mode 1
`and Test Mode 4 tests to be run in the
`presence of a disturbing signal. The
`disturbing signal is defined as a sine
`wave generator that simulates the
`potential interfering effect of another
`1000BASE-T transmitter. We will not
`cover testing with disturbing signal in
`detail in this article.
`
`10
`
`
`
`Page 10
`
`

`
`Return loss testing
`
`Return loss is a measurement of the
`sum of the reflected energy coming
`back from the receiving device to the
`transmitting device. Return loss is
`defined in the standards for 10BASE-
`T, 100BASE-TX, and 1000BASE-T. It is
`especially important for 1000BASE-T
`devices as these devices use bi-
`directional signaling. These means
`that the same pins which transmit
`data also receives data. If the
`receiving device has bad return loss,
`the information originating from the
`transmit side will be reflected back,
`causing problems for the transmit side
`as it is also acts as a receiver, listening
`for data from the transmitter at the
`other end. Therefore return loss testing
`is important to ensure interoperability
`with other devices.
`
`Return loss testing is performed using
`a vector network analyzer. The N5395B
`or N5395C Ethernet electrical probing
`fixture can be used to facilitate the
`return loss measurement. As the
`signals from the Ethernet device are
`differential, a balun transformer on
`the fixture performs the differential to
`single-ended impedance conversion
`to the network analyzer input. The
`N5395B or N5395C fixture also
`provides a return loss calibration
`fixture with RJ-45 short, open, and
`load connections to calibrate the
`vector network analyzer.
`
`The return loss can be measured on
`a single-port of the vector network
`analyzer using a forward reflection or
`S11 measurement in log magnitude.
`As this measurement is made in
`50 Ω through a balun, it has to be
`conversion mathematically to its
`equivalent in 100 ± 15 Ω. Therefore
`return loss is calculated at 85 Ω, 100
`Ω, and 115 Ω (111 Ω additionally for
`10BASE-T).
`
`To perform the conversion, we use the
`following equations:
`Return Loss in dB, RLdB = 20 log10 |Г0|
`
`From Zin, the reflection coefficients
`for the different impedances can be
`calculated.
`
`Zin - 85
`Г85Ω = ———
`Zin + 85
`
`Zin - 100
`Г100Ω = ———
`Zin + 100
`
`Zin - 115
`Г115Ω = ———
`Zin + 115
`
`And the resulting return loss in dB can
`be obtained by the following:
`RL85Ω = 20 log10 |Г85Ω|
`RL100Ω = 20 log10 |Г100Ω|
`RL115Ω = 20 log10 |Г115Ω|
`
`Where Г0 is the array of complex
`reflection coefficient values (vs
`frequency) of the measurement made
`on the VNA and is represented in
`terms of impedance by
`Zin - Z0
`Г0 = ———
`Zin + Z0
`Zin is the impedance of the DUT which
`is also a complex array (vs frequency),
`and Z0 is the standard reference
`impedance at which the measurement
`was made (real number).
`
`Since the measurement was made
`on the vector network analyzer at the
`standard reference of 50Ω, but first
`going through a 2:1 conversion through
`the balun, Z0=2*50=100.
`
`Solving for Zin,
`1 + Г0
`Zin = Z0 ————
`1 - Г0
`
`Figure 17. Return loss vs frequency plot of a 1000BASE-T device
`
`11
`
`
`
`Page 11
`
`

`
`Conclusion
`
`You now have an overview of the
`how the electrical validation of the
`popular 10BASE-T, 100BASE-TX, and
`1000BASE-T is done. The signals used
`in the transmission of Ethernet signals
`get more complex as the data rate
`increases exponentially from 10 to
`1000 Mbits/s. This trend will continue
`as designers try to transfer more data
`over the bandwidth-limited UTP cable
`using more complicated modulation
`schemes.
`
`The quality and signal integrity of your
`measurement tools play an important
`role in ensuring that you have the best
`representation of the signals you are
`measuring. Most of the measurements
`described here can be made manually
`on a modern digitizing real-time
`oscilloscope. The signals used in
`these illustrations were from an off-
`the-shelf network interface card and
`captured using an Agilent 80000 Series
`
`oscilloscope with an active differential
`probe. An Agilent N5395B or N5395C
`Ethernet electrical probing fixture was
`used to provide the probing circuits
`shown in Figures 2 and 3. There is also
`an automated test application available
`that can be used in conjunction with
`the N5395B or N5395C fixture. For
`more information regarding tools for
`Ethernet validation, visit
`www.agilent.com/find/n5392a.
`
`Related literature
`
`Publication title
`N5392A Ethernet Electrical Performance
`Validation and Compliance Software for
`Infiniium Oscilloscopes
`Infiniium DSO80000B Series Oscilloscopes and
`InfiniiMax Series Probes
`Infiniium 90000 Series Oscilloscopes
`
`Publication type
`Data Sheet
`
`Publication number
`5989-1527EN
`
`Data Sheet
`
`Data Sheet
`
`5989-4606EN
`
`5989-7819EN
`
`Product Web site
`
`For the most up-to-date and complete
`application and product information,
`please visit our product Web site at:
`www.agilent.com/find/n5392a
`
`12
`
`
`
`Page 12
`
`

`
`Agilent Technologies Oscilloscopes
`Multiple form factors from 20 MHz to >90 GHz | Industry leading specs | Powerful applications
`
`13
`
`
`
`Page 13
`
`

`
`8 Agilent Email Updates
`
`www.agilent.com /find/ emailu pdates
`Get the latest information on the
`products and applications you select.
`
`A~"
`,.~te
`www.axiestandard.org AdvancedT(cid:173)
`CA® Extensions for Instrumentation
`and Test (AXle) is
`an open standard that extends the
`AdvancedTCA® for general purpose
`and semiconductor test. Agilent
`is a founding member of the AXle
`consortium.
`
`Agilent
`Advantage
`Services
`
`Agilent Advantage Services is com(cid:173)
`mitted to your success throughout
`your equipment's lifetime. We share
`measurement and service expertise
`to help you create the products that
`change our world. To keep you com(cid:173)
`petitive. we continually invest in tools
`and processes that speed up calibra(cid:173)
`tion and repair. reduce your cost of
`ownership. and move us ahead of
`your development curve.
`
`www.agilentcom/lind/ advantageservices
`
`www.agilent.com/ quality
`
`/.)(1
`www.lxistandard.org
`LAN eXtensions for Instruments puts
`the power of Ethernet and the Web
`inside your test systems. Agilent is a
`founding member of the LXI consor(cid:173)
`
`tium. --·-r.A•
`
`http:/ /www.pxisa.org
`PCI eXtensions for Instrumentation
`(PXI) modular instrumentation
`delivers a rugged, PC-based high-per(cid:173)
`formance measurement and automa(cid:173)
`tion system.
`
`Agilent Channel Partners
`IIWIM!agilent.com/find/channelpartners
`Get the best of both worlds: Agilent's
`measurement expertise and product
`breadth, combined with channel
`partner convenience.
`
`MATLAB is a U.S. registered trademark of
`The MathWorks, Inc.
`
`www.agilent.com
`www.ag ilent.co m/find I n539 2a
`
`For more information on Agilent Tech(cid:173)
`nologies' products, applications or services,
`please contact your local Agilent office. The
`complete list is available at:
`www.a g ilent.com /find I contactus
`
`Americas
`Canada
`Brazil
`Mexico
`800
`United States
`
`(877) 894 4414
`(11) 4197 3500
`01800 5064
`
`(800) 829 4444
`
`Asia Pacific
`1 800 629 485
`Australia
`800 810 0189
`China
`800 938 693
`Hong Kong
`1 800 112 929
`India
`0120 (421) 345
`Japan
`080 769 0800
`Korea
`1 800 888 848
`Malaysia
`1 800 375 81 00
`Singapore
`0800 047 866
`Taiwan
`Other AP Countries (65) 375 81 00
`
`Europe & Middle East
`32 (0) 2 404 93 40
`Belgium
`Denmark
`45 70 13 15 15
`Finland
`358 (0) 10 855 2100
`France
`0825 010 700*
`'0.125 €/minute
`49 (0) 7031 464
`
`Germany
`6333
`Ireland
`Israel
`Italy
`Netherlands
`Spain
`Sweden
`United Kingdom
`
`1890 924 204
`972-3-9288-504/ 544
`39 02 92 60 8484
`31 (0) 20 547 2111
`34 (91) 631 3300
`0200-88 22 55
`44 (0) 118 9276201
`
`For other unlisted Countries:
`www.a g ilent.com /find I contactus
`Revised: October 14, 2010
`
`Product specifications and descriptions
`in this document subject to change
`without notice.
`
`© Agilent Technologies, Inc. 2011
`Printed in USA, January 13, 2011
`5989-7528EN
`
`·. ; .·
`··•··
`...•....
`.. ··:·· .. Agilent Technologies
`
`Page 14