A-1005 Industrial Ethernet Cabling Standard
A-1005 Industrial Ethernet Cabling Standard
A-1005 Industrial Ethernet Cabling Standard
Contents
n Abstract: Paper Premise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .03
n Background: TIA-1005 Telecommunications
Infrastructure Standard for Industrial Premises . . . . . . . . . . . . . . . . . . . . . .03
n Technology Issues: IEEE 802.3u Considerations . . . . . . . . . . . . . . . . . . . . . . .04
n Testing Parameter and Procedures: Port Variability Test Methodology . . . . . . .06
n Test Setup: Data Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .07
n Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .08
n References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .08
This white paper was prepared by Anixter Inc., which is not affiliated with the Telecommunications Industry Association or the Electronic Industries Alliance. TIA is not responsible for the content of this publication. Some material
in this publication is reproduced from standards publications that are copyrighted by the Telecommunications Industry Association.
3
Abstract: Paper Premise The biggest change in the TIA-1005 standard over the TIA-568-B documents
is the incorporation of a 2-pair cable for installation into the newly defined
The Anixter Infrastructure Solutions Lab wanted to determine what effect the automation island. However, using only 2-pairs will limit the Ethernet
new TIA-1005 industrial cabling infrastructure standard would have on the capability to that of only 10 or 100 Mbps, or only to the IEEE 10BASE-T[4]
data throughput performance of real Ethernet data packets running between and 100BASE-T[5] -PHY[6] definitions. Gigabit Ethernet must run on a 4-pair
SmartBits test cards and various manufacturers’ 10/100 Ethernet switches in cabling system from end to end. To understand the reasons for the cabling
a real-world simulation. The test included five (5) different IP20-rated switches designs and automation island designations, the MICE [3] concept is introduced
and three (3) different enterprise rack-mounted switches using various cabling and integrated into the design and performance requirements.
channels made from both Category 5e and Category 6 cabling components
and connector pairs that are allowable under the standard. The premise also Annex C of the new TIA-1005 standard shows the allowed connectivity
asserts that the effect of the cabling channel interference will also vary from and cabling situation for connecting devices with network interface cards
port to port and switch to switch because of the variable transmitter and (NICs) on one end of a machine to the IP67-rated switch on the other end
receiver functionality. using M12 connections and traversing multiple environmental zones that
require adaption to different connectivity. The test included a standard
Background: TIA-1005 Telecommunications four (4) connector channel configuration and a more complex six (6)
performance from port to port, which can cause possible errors in Source: Anixter Infrastructure Solutions Lab
higher temperature situations.
• Network interface cards typically are not as robust as switch ports Figure 3 – The MLT-3 encoded signal as shown produces four separate
and very few have “CRC” error trapping or indication. “symbols” or information bits for every megahertz of bandwidth used.
• Category 5e cabling systems will probably not run error free in The fourth symbol is obtained from the “direction” change that the voltage
channels where there are more than four (4) connections and takes from its previous state.
temperatures exceed 20ºC (68ºF).
In a previous Lab procedure, three (3) instances of Ethernet packet
Technology Issues: IEEE 802.3u [7] Considerations transmission were used[9]. Each instance used exactly the same data
patterns and was precisely timed so the analog signal traced on the digital
As background for understanding the principles in this throughput study,
the makeup of the signal needs to be defined. The basic premise behind the oscilloscope (DSO) as an exact overlay. The three (3) instances were the
encoding scheme for 100 Mbps Ethernet over copper was to use a four (4) transport over cable channels of Category 5, 5e and 6 performance.
symbol process and a frequency scrambler to keep from having a “killer Figure 2 shows the difference in signal edge clarity and strength for
packet” as was the case in 10BASE-T. By constantly moving the center each type of cable channel.
frequency up and down the spectrum, it was thought that no constant
data pattern could cause an increase in error rate. Figures 2 and 3 show
the binary and multisymbol schemes used for Fast Ethernet transmission
over copper cabling. +
Volts
Source: Anixter Infrastructure Solutions Lab Figure 4 – Actual data signal over three different cable channels of
Category 5, 5e and 6 (overlaid).
Figure 2[8] – Uses 4B-5B encoder and 25 MHz clock for 125 M baud and
100 Mbps. Fast Ethernet or 100BASE-TX uses a more complex encoding Figure 4a, on the next page, zooms into the three (3) center time pulses and
technique called MLT-3, which is similar to Manchester in its encoding shows clearly the distortion on signal caused by lesser performing cabling
process, but uses a three-state alternating wave to digitally encode the channels. The Category 6 wave shape is clear and sharp with no jitter, and the
binary information. An example of an MLT-3 encoded signal is shown attenuation is far less then that of the Category 5 and 5e cabling. The Category
in Figure 3. 5 channel shows the actual edge breakdown in the pulse as well as major
insertion loss deviation (ILD) effects. The original Fast Ethernet specification
requires only Category 5 cabling performance, which has no return loss
requirements and, therefore, is usually erratic in its bit error rate testing (BERT)
performance. The equalization design of the receiver does not allow for ILD[10].
5
0
Volts
160 170 180 190 200 210 220 230 240 250 Source: Anixter Infrastructure Solutions Lab
Time (Milliseconds) Source: LeCroy 1999
Figure 4a – Zoom of Overlaid Signals Figure 6 – MLT-3 MASK
In order to verify the transmitters of the 10/100 Ethernet switch are outputting Figure 7 is a representation of how the transmitted signal from a port on the
Industrial Ethernet switch is modified by the actual cabling channel between
-0.5
-1
Figure 5 – InterOperability Lab MASK Spec Figure 7 – MLT-3 MASKs of IP20 Switch – 5,700 Sweeps
Figure 6 is the MASK that was used on the DSO in Anixter’s Lab for the evaluation The other major component of the transmission path is the receiver, which
on the IP20 switch port transmitters. in most installations is a NIC or chip set as part of a computer or intelligent
device used to output data over Ethernet. In a factory, this could be a PLC or
smart I/O device. The NIC is essentially a single-port transceiver as are the
individual ports of a switch. Figure 8, on the following page, shows a previous
study of 17 different brands of NICs where eight (8) of them failed the rise-time
test out of the box and all of them had different transmitter signatures even
though all were compliant with the IEEE specifications.
6
Products. Technology. Services. Delivered Globally.
1.2 Figure 9 show the diagram of the test bed configurations used for all
1 of the testing.
10ft - RJ45
Patch Cable
0.8 RJ45 Jack
4-pair
C-1
0.6
le
Volts
i r ab
C-3 RJ45 Patch Panel
5
Pa RJ4
pa C
4- tch
Cat6
0.4
ft
10
0.2 260 ft
Industrial Ehternet Cable C-2
IP30 Switch w/8 10/100 Ports 4-pair
0
38ft Cable
4-pair
-0.2
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 10ft - RJ45 C-4
Patch Cable
Nanoseconds Source: LeCroy 1999 4-pair
Figure 8 – 8/17 NIC cards sampled had rise times > 5 nanoseconds Test setup for Category 5e and 6 channels (30°C and 40°C equivalent
Because over-the-counter Ethernet electronics can vary from manufacturer to of 341 ft. and 355 ft.) and five Category 6 connectors.
manufacturer, this NIC test, which was done nearly six years ago when third- 40ºC
20ft - RJ45
Cable
4-pair
generation 10/100 network interface cards were hitting the market, shows how
RJ45 Plug
different the output characteristics can be on a new product. Add a few years of 7ft - RJ45
Patch Cable
To Jack
6ft - RJ45 C-3
4-pair
service at high temperature or temperature cycling in a normal on/off mode and C-2
Cable
4-pair
RJ45 Cat
the voltages will drift even more than shown here. Hardware qualification tests Jack
C-1
10ft - RJ45
Patch Cable
4-pair
can only be performed on the transmitters, not the receivers. The only way to Industrial Ehternet C-3
4- h Ca J45
ir le
RJ45 Plug
tc - R
pa b
IP30 Switch C-5 RJ45 Patch Panel
Pa 0ft
W/8 10/100 Ports To Jack
test a receiver is to use it. That is why SmartBits and the entire group of RFC Cat6
1
throughput, back to back, latency and BERT performance suites were developed. 260ft
Cable C-4
4-pair
4-pair
• Set up cabling test beds to show three (3) possible factory
Ethernet conditions: Figure 9 – Test setup for TIA standard Category 5e and 6 channels
– TIA standards-based channel at 20ºC (68ºF) (20°C equivalent of 328 ft.) and four connectors.
– More typical factory temperature of 30ºC (86ºF)
In order to verify that the test-bed channels built for this test were compliant
– Worst-case temperature of 40ºC (104ºF)
with the TIA-568-B.2 standard and the new TIA-1005 performance requirements
• Set up both Category 5e and Category 6 channels with both
for Category 5e, all configurations were tested using a Fluke DTX-1800
four (4) and five (5) connectors.
• Use SmartBits SB200[12] as source of 100BASE-T Ethernet packets. automated field test device, which is rated for Category 5 through Category 6A
and up to 500 MHz resolution. Figures 9 and 10 show the worst-case results of
• Use lab-grade RAS[13] or network sniffer for error trapping.
these test, and all passed the channel requirements.
• Test all switch ports on all test beds, averaging the error counts
of three (3) consecutive burst trials. Both channels passed certification
60
Return Loss (RL) Return Loss (RL)
60
This is an actual picture (Picture 1) of the test setup in the Anixter
Infrastructure Solutions Lab including the Category 6 at 40ºC (104ºF) channel.
dB
dB
SmartBits 200 20 20
0 0
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
Frequency Frequency
Cat 5e Channel - 40ºC Cat 6 Channel - 40ºC
Source: Anixter Infrastructure Solutions Lab
Both channels passed certification The output of the received signal was trapped by the RAS and displayed through
120
Return Loss (RL)
120
Return Loss (RL) its user interface. Picture 3 shows this interface.
80 80
dB
dB
20 30
0 0
0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350
Frequency Frequency
Cat 5e Channel - 40ºC Cat 6 Channel - 40ºC
Source: Anixter Infrastructure Solutions Lab
Conclusions References
1. The average CRC error rates are dramatically affected by the electrical [1] ANSI/TIA-1005, formally SP-3-4822-REV H, Telecommunications Infrastructure Standard
performance design of the cabling channel components, with the For Industrial Premises.
[2] ANSI/TIA/EIA-568-B.2-2–Commercial Building Telecommunications Cabling Standard.
Category 6 channels consistently close to error-free transmissions Balanced Twisted Pair Cabling Components. Transmission performance specifications
at these elevated temperatures. for 4-pair Category 5e and Category 6 cabling.
[3] MICE – Environmental rating system from ISO/IEC 24702. CLC EN 50173-1. The suffixes
2. All eight (8) of the 10/100 Ethernet switches had different transmission
for the four primary environmental criteria are either 1, 2 or 3.
signatures from each port; the assumed cause is the combination of differing [4] 10BASE-T: A 10BASE-T node (such as a PC) that transmits on pins 1 and 2 and receives
chip suppliers and the passive components that supply the actual output on pins 3 and 6 to a network device is most often on a straight-through cable in the MDI
signal to the RJ45 jack. wiring pattern where RX goes to RX and TX goes to TX. A 10BASE-T transmitter sends
2 differential voltages, +2.5 V or -2.5 V.
3 During the running of the tests, it was observed that when the background [5] 100BASE-T: A 100BASE-TX transmitter sends 3 differential voltages, +1 V, 0 V, or -1 V.
data patterns were change to some other sequence, such as 808080s, the [6] PHY: connects a link layer device (often called a MAC) to a physical medium such as an
average CRC error rates also seemed to change. This, however, is a subject optical fiber or copper cable. A PHY typically includes a PCS and a PMD layer. The PCS
for further study. encodes and decodes the data that is transmitted and received. The purpose of the
encoding is to make it easier for the receiver to recover the signal.
4. As a follow up to the RAS testing, a validation test of one of the ports was [7] IEEE 802.3u: 100BASE-T is any of several Fast Ethernet standards for twisted-pair cables,
performed between two of the SmartBits ML-7710 cards. The RAS is standard including 100BASE-TX (100 Mbps over 2-pair Category 5 or better cable), 100BASE-T4
(100 Mbps over four-pair Category 3 or better cable, defunct), 100BASE-T2 (100 Mbps
NIC based and more typical of the real-world testing for the paper, as
over two-pair Category 3 or better cable, also defunct). The segment length for a
apposed to the ML-7710 cards, which are precision-test devices with 100BASE-T cable is limited to 100 meters (328 feet) (as with 10BASE-T and Gigabit
high-quality components and tend to have reliable receivers and transmitters. Ethernet). All are or were standards under IEEE 802.3 (approved 1995).
The result of this test also showed that the bit error rates were higher with [8] MLT-3: 100BASE-TX introduces an additional, medium dependent sublayer, which employs
the Category 5e channel then with the Category 6 channel. MLT-3 as a final encoding of the data stream before transmission, resulting in a maximum
fundamental frequency of 31.25 MHz. The procedure is borrowed from the ANSI X3.263
FDDI specifications, with minor discrepancies.
[9] Bill Richardson, LeCroy R&D, Vigilent, 1997.
[10] ILD: has basically two components one which essentially represents an offset of the
summed up operational attenuation of the components, and a second component, which
is oscillatory in nature over frequency. “Look into the systematic behaviour of insertion
loss deviation in data grade channels, its partial measurement and characterization
as a specification requirement.” Jörg-Hein (Jo) Walling, IEEE Proceedings, Science,
Measurement and Technology, November 2003, Volume 150, Issue 6, p. 279–288.
[11] UNH-IOL: The University of New Hampshire, Interoperability Lab. “Fast Ethernet
Consortium, Clause 25, Physical Medium Dependent (PMD), Test Suite 3.4.”
[12] SB200: Spirent/SmartBits Test Equipment with ML-7710 Interface Cards for
10/100 Ethernet.
Figure 13 – SmartBits CRC Error Testing, Category 6 and 5e at 40°C [13] RAS: Digitech/LeCroy Remote Access System.
Aerospace Hardware • Electrical and Electronic Wire & Cable • Enterprise Cabling & Security Solutions • Fasteners
9F0001X0 © 2009 Anixter Inc. • 10/09
Anixter Inc. • 1.800.ANIXTER • anixter.com • World Headquarters: 2301 Patriot Boulevard, Glenview, IL 60026-8020 • 224.521.8000
Anixter is a leading global supplier of communications and security products, electrical and electronic wire and cable, fasteners and other small components. We help our customers specify solutions
and make informed purchasing decisions around technology, applications and relevant standards. Throughout the world, we provide innovative supply chain management services to reduce our customers’ total cost
of production and implementation. A NYSE listed company, Anixter, with its subsidiaries, serves companies in more than 52 countries around the world. Anixter’s total revenue exceeded $6.1 billion in 2008.
Anixter Inc. does not manufacture the items described in this publication. All applicable warranties are provided by the manufacturers. Purchasers are requested to determine directly from the manufacturers the applicable product warranties and limitations.
Data and suggestions made in the publication are not to be construed as recommendations or authorizations to use any products in violation of any government law or regulation relating to any material or its use.