Magnetics for 10GBase-T

advertisement
Magnetics for 10GBase-T
Thuyen Dinh
LAN Products Division
IEEE802.3 Plenary Meeting
San Francisco
July 20-25, 2003
Pulse
T. Dinh
Topics of discussion
●
●
●
●
●
●
Design example 1
Test set ups
Test results
Design example 2
Test results
Summary
Pulse
T. Dinh
10GBase-T magnetics design
example1
●
●
Transformer:
– 1:1.4 turns ratio for 50Ω:100Ω impedance
matching. No center taps on transformer.
– Insulation to meet 1500 VAC
– Lower inductance than traditional 10/100
magnetics (about 100 uH)
– Not packaged (wires soldered to test fixture
pads)
Common-mode choke
– Typical wide-band common-mode noise
suppression material
Pulse
T. Dinh
Test set up, example 1
●
Back-to-back test set up in Insertion Loss (S21) to
minimize number of interface levels
NETWORK ANALYZER
PORT A
PORT B
1:1.4
1.4:1
INSERTION LOSS
TEST SET UP
Pulse
T. Dinh
Test set up, example 1 (continued)
●
Return loss was tested from the driver side to
eliminate need for wide-band 50Ω:100Ω balun
NETWORK ANALYZER
PORT A
100
1:1.4
RETURN LOSS
TEST SET UP
Pulse
T. Dinh
Test results, example 1
10GBase-T Magnetics Performance Curves
1
10
100
1000
400
-5
350
-10
300
-15
250
-20
200
-25
150
-30
100
-35
50
-40
0
Delay (ns)
Loss (dB)
0.1
0
Frequency (MHz)
S21 (back-to-back)
S11 (back-to-back)
Pulse
S11 (50-Ohm side)
Delay (Back-to-back)
T. Dinh
Test results, example 1 (continued)
●
●
Insertion Loss
– 6-dB bandwidth (back-to-back) spans from
around below 100 kHz to above 600 MHz
– Delay stays relatively flat from 2 MHz and
beyond
Return Loss
– Better than 6 dB from 100 kHz to 500 MHz
– Better than 10 dB from 200 kHz to 375 MHz
Pulse
T. Dinh
10GBase-T magnetics design
example2
●
●
Transformer:
– 1:1 turns ratio for 100Ω:100Ω impedance
matching.
– Insulation to meet 1500 VAC
– Lower inductance than traditional 10/100
magnetics (about 80 uH)
– Has a center tap on each side
– In a surface-mount plastic package
No common-mode choke
Pulse
T. Dinh
Test set up, example 2
●
Use non-isolated 50:100-Ohm baluns to interface
with network analyzer
NETWORK ANALYZER
PORT A
PORT B
UUT
50:100
Ohm balun
50:100
Ohm balun
INSERTION LOSS
TEST SET UP
Pulse
T. Dinh
Test set up, example 2 (continued)
●
Use a non-isolated 50:100-Ohm balun to
interface with network analyzer
NETWORK ANALYZER
PORT A
UUT
100
50:100
Ohm balun
RETURN LOSS
TEST SET UP
Pulse
T. Dinh
Test results, example 2
Wide-band transformer, 1:1
Frequency (MHz)
0.0
100.0
200.0
300.0
400.0
500.0
600.0
700.0
800.0
900.0
1000.0
0.0
-5.0
Loss (dB)
-10.0
-15.0
-20.0
-25.0
-30.0
Return Loss (S11)
Pulse
Insertion Loss (S21)
T. Dinh
Test results, example 2 (continued)
●
●
Insertion Loss
– 3-dB bandwidth spans from around below 200
kHz to about 430 MHz
Return Loss
– Better than 10 dB from 1 MHz to 1 GHz
– Better than 15 dB from 3 MHz to above 600 MHz
Pulse
T. Dinh
Summary
●
●
●
Performance at high frequencies are very
sensitive to package size and whether center taps
on the transformer are needed.
Bandwidth (3dB) from less than 200 kHz to more
than 400 MHz is feasible.
Since the numbers of turns will be small, odd
turns ratios such as 1:1.4, 1:1.15, etc. may be
difficult to implement, especially if center taps
are needed.
Pulse
T. Dinh
Download