Gary,
About 50% of our 100BASE-Tx cards require an additional 4-line common-mode choke
between the transformer-filter and the RJ-45 connector, to meet Radiated
Emissions limits with enough margin to satisfy our electromagnetic compatibility
(EMC) folks. So we snuggle all three components as close together as our
manufacturing guidlines will allow, usually within 0.025" of one another
pad-to-pad or pad-to-body. Since the transformers and filters in the
transformer-filter are symmetrical between the + and - sides, and the transmit
and receive sides often are the same, we shuffle the pin numbers in the
common-mode choke and transformer-filter so that the traces run almost parallel
(no crossovers) in the critical area between the RJ-45 connector and the
transformer filter:
RJ-45
+--------+ common- transformer-
! ! mode filter
! O ! choke +------+
! O ! +------+ /-! !
! O------! !--/ ! !
! O ! ! ! /-! !
! O ! /-! !--/ ! !
! O-----/ ! ! ! !
! O------! !--\ ! !
! O-----\ ! ! \-! !
! ! \-! !--\ ! !
+--------+ +------+ \-! !
+------+
This forces vias and crossovers in the RXD+, RXD-, TXD+, and TXD- lines to the
zone between the transformer-filter and the PHY or MAC/PHY, where they are much
less critical. We also try to get the PHY or MAC/PHY as close as we can to the
transformer-filter, no more than 1 inch away and preferable within 0.5 inch,
with the trace lengths closely matched.
Depending on how bold the engineer is, on our first engineering cards we will
short the pads of the common-mode choke together with traces or with 0-ohm
surface-mount resistors paralleling the windings of the common-mode choke. I
personally prefer the resistors, because if we ever discover that we need the
common-mode choke, all I have to do is no pop the four resistors and put the
common-mode choke in their place. I've considered using a 0-ohm 4-resistor
resistor network, but didn't feel comfortable with the coupling that might cause
between the receive and transmit signals on pins 2 and 3. Since my most recent
card had over 1000 components crammed onto a 12.5" x 7.9" card, including a
350MHz X86 processor, six large ball-grid array (BGA) parts, and provisions for
512MB of SDRAM, I wasn't willing to take any risks that I didn't have to. (We
started production two months ago using the second spin of my card...)
John Barnes Advisory Engineer
Lexmark International
-------------------------------------------
This message is from the IEEE EMC Society Product Safety
Technical Committee emc-pstc discussion list.
Visit our web site at: http://www.ewh.ieee.org/soc/emcs/pstc/
To cancel your subscription, send mail to:
majord...@ieee.org
with the single line:
unsubscribe emc-pstc
For help, send mail to the list administrators:
Michael Garretson: pstc_ad...@garretson.org
Dave Heald davehe...@mediaone.net
For policy questions, send mail to:
Richard Nute: ri...@ieee.org
Jim Bacher: j.bac...@ieee.org
All emc-pstc postings are archived and searchable on the web at:
No longer online until our new server is brought online and the old
messages are imported into the new server.
