1000BASE-T/100BASE-TX/10BASE-T Physical Layer Compliance Tests Manual
Intel Confidential
135
100Base-TX Test Procedure for the 82544 Chip
Appendix I: 100Base-TX Test Procedure
for the 82544 Chip
I.1
Introduction
A few of our prior test procedures will not work well with the 82544 chip. This document has been
written to enable 100Base-TX physical layer testing with the 82544.
To make 100Base-TX testing easier, most Intel 825xx physical layer chips allow the test operator to
disable scrambling and 4B/5B encoding, by setting the appropriate bits in the phy registers. By
disabling both scrambling and 4B/5B, and by transmitting certain fixed data patterns, it was easier
to repetitively generate the required waveforms for performing some of the 100Base-TX physical
layer tests. Continuous, repetitive waveforms are easier to trigger on, and they allow the operator to
use an oscilloscope’s averaging function while viewing the repetitively captured waveforms.
Note:
82544 does not have a test feature, to disable 4B/5B encoding. Consequently, continuously
transmitting the same pulse widths with the same width mid-level shoulders is virtually impossible.
This means that each captured pulse is more likely to have different amounts of “DC offset” (due to
baseline wander, which is related to the pulse-width and polarity of the preceding pulses, which can
not always be controlled with 82544.)
Although it is still possible to pulse-width trigger on 100Base-TX waveforms, which are
transmitted by the 82544, the constantly changing DC offsets cause a large amount of amplitude
jitter in the acquired waveform. Normally, averaging reduces the level of jitter that is visible to the
operator, and averaging makes it easier to accurately place an oscilloscope’s horizontal cursor on
key parts of the 100Base-TX pulse. Unfortunately, because the widths of the mid-level, MLT-3
“shoulders” cannot be controlled and are not continuously repetitive. Selecting the appropriate
location for the oscilloscope’s horizontal cursors, on the mid-level shoulder of an averaged
100Base-TX waveform is virtually impossible. Immediately to the left and right of the trigger
pulse, the waveform is continuously changing shape and level. On an averaged 82544 waveform,
there is no obvious place to place the cursor, when the operator is trying to locate the base of an
averaged 100Base-TX pulse.
Note:
It is still possible to perform 100Base-TX phy conformance tests, as long as oscilloscope averaging
is disabled. This document provides modified versions of some 100Base-TX phy conformance test
procedures, to make it easier for test operators to perform these tests, and to get consistent and
accurate test results.
This document provides alternate test procedures for 100Base-TX:
•
Differential Output Amplitude (ANSI TP-PMD specification 9.1.2.2) -
•
Rise and Fall Time (ANSI TP-PMD specification 9.1.6) -
•
Transmitter Duty Cycle Distortion (ANSI TP-PMD specification 9.1.8) -
I.2
Equipment Used for This Procedure
•
Test Equipment
