Teledyne LeCroy QPHY-ENET - Manual (Page 59)

QPHY-ENET Software Option
QPHY-ENET Operator’s Manual Rev G
59
1000Base-T Master and Slave Jitter - with TX_TCLK
For devices exposing TX_TCLK, as required by 802.3-2005, jitter can be measured as specified in the 802.3-
2005 standard. This process involves three steps.
1. Measure Jtxout, defined as the peak-to-peak jitter on the MDI output signal (data) relative to the
TX_TCLK, while the DUT is in test mode 2 (Master timing mode) or test mode 3 (Slave timing mode). This
measurement is made on each of the four MDI pairs (both the master and the slave).
2. Measure both the unfiltered and filtered TX_TCLK relative to an “unjittered reference,” while the DUT is
configured for normal operation as a master (requiring linkage to the Link Partner using a short cable). In
our implementation, the “unjittered reference” is the mean frequency of each acquisition and phase-
aligned with the data on each acquisition.
The jitter is the difference between an actual observed zero crossing time and the corresponding “ideal”
crossing time.
3. Now, measure the unfiltered and filtered TX_TCLK jitter with the DUT operating as a SLAVE. The slave
jitter is defined with respect to the Master’s TX_TCLK. Therefore, both the DUT (slave) and Link Partner
(master) TX_TCLK must be acquired simultaneously. Also, the DUT must be connected to the Link
Partner through a Jitter Test Channel (which is a cable made of different impedance sections, defined in
the 802.3-2005 standard).
The following topics explain the three steps in greater detail.
Note: This process is documented and detailed in the “Test 40.1.5” section of the Gigabit Ethernet Consortium Clause 40 PMA Test Suite,
version 2.4. We’d like to thank the Gigabit Ethernet Consortium for agreeing to let us refer to this document. The Gigabit Ethernet Consortium
is based at the University of New Hampshire Interoperability Lab.
The document can be found at:
HU
ftp://ftp.iol.unh.edu/pub/ethernet/test_suites/CL40_PMA/PMA_Test_Suite_v2.4.pdf
U
Step 1: Master and Slave Jtxout Measurements
Jtxout in the master mode is measured with the transmitter under test and terminated into a 100 Ω resistive load.
The transmitter under test is set to transmit an alternating +0.5/-0.5 V pattern during configuration for test mode 2.
Connect the TX_TCLK and one pair of the MDI output to the oscilloscope, as shown in
X
Figure 47
X
(as follows).
Capture 100ms to 1s worth of edge data (requires multiple acquisitions). The peak-to-peak jitter of the output
signal (on the signal output lines) relative to the corresponding transition of the transmit clock (TX_TCLK) is
measured. Write this value down, and repeat the process for the other three pairs.
Jtxout in the slave mode is measured similarly, except the DUT is configured for test mode 3.
The result should be 8 values – one for each pair in master mode, and one for each pair in slave mode. These
values are combined with results from steps 2 and 3 (as follows). There are no pass/fail criteria on Jtxout values
themselves.
1. Attach an active or differential probe (Minimum Bandwidth 1.5GHz) to the TX_TCLK signal of the DUT.
On the oscilloscope’s Ethernet Tests menu, enter the channel to which this probe is connected as
“Master TX_TCLK.”
2. Apply power to the Device Under Test and set it to mode 2 (or mode 3). A clock signal should appear on
the channel to which the probe is attached.
3. Select 1000Base-T from the Select Standard field choices.
4. Select Master and Slave Jtxout from the Select Test field choices.
5. Connect one pair of the MDI output to the oscilloscope, using SMA cables. Use the menu to enter the
channel to which the SMA cables are attached: C2 for the Source for + Data field and C3 for the Source
for – Data field.
6. Click the Set Up and Start Test control to begin the test.
7. Parameter P1 is Jtxout, the p-p jitter on the MDI output pair relative to the TX_TCLK. Write down this
value for later use.