Beruflich Dokumente
Kultur Dokumente
Channel
(copper or fiber)
OE
TIA
-18 dBm
10 A
LA
10 mV p-p
EQ
CDR
DMUX
400 mV p-p
4-foot cable
Copper Cable
( )
H e L
15-foot cable
waveform
eye diagram
V2
V1
V3
+0.5
-0.5
C1
1s
1V1
V1
bandpass filter
1 V2
V2
V1
V3
+0.5
-0.5
1s
V1
C1
1V1
V1
V2
(1 ) V
V3
Feedforward Path
Vout
Vcontrol
f (Hz)
10
teq = 45ps
PW = 108ps
0.3
teq = 60ps
PW = 100ps
-0.3
2.4
teq = 75ps
PW = 86ps
2.5
2.6
2.7
2.8
t (ns)
Simulations indicate that ISI correlates strongly with FFE transition time teq.
Optimum teq is observed to be 60 ps.
Nonlinearities affect pulse shape, but not location of zero crossings.
11
Slicer
Exhibits controlled
transition time
12
Feedback Path
13
V+
VS
VVS
Transient Characteristic:
ISS
CSS
V+ V
V+ V
(b)
(a)
VS
(b)
(a)
14
Integrator
A0
1
H(s) =
1+ s int A0 s int
A0 = gm 1 ro1 || ro2
int
CL
=
gm 1
15
Detector + Integrator
From
FFE
tFFE
From
Slicer
tslicer= 60ps
slope
detector
slope
detector
FFE transition
Time tFFE
Vcontrol (mV)
90ps
60
40
75ps
20
60ps
_
+
Vcontrol
EECS 270C / Winter 2014
-20
-40
45ps
-60
15ps
10
20
30
40
50
t (ns)
Prof. M. Green / UC Irvine
16
System Analysis
tslicer
detector
Kd
Vcontrol
integrator
feedforward
equalizer
H(s)
teq
Keq
detector
Kd
t eq
t slicer
K d K eq H (s )
1 + K d K eq H (s )
1
H (s )
s int
EECS 270C / Winter 2014
t eq
t slicer
int = 75ns
1
1+ s
int
K d K eq
adapt =
int
= 20ns
Kd Keq
17
Measurement Setup
EQ inputs
EQ outputs
18
EQ output
4-foot
RU256 cable
(-5 dB atten. @ 5 GHz)
15-foot
RU256 cable
(-15 dB atten. @ 5 GHz)
19
Supply voltage
3.3 V
Power Dissipation
350 mW
(155 mW not including output driver)
Die Size
0.81mm X 0.87mm
Output Swing
Random Jitter
20
21
Pre-Cursor/Post-Cursor ISI
0
T
cursor
Output pulse:
pre-cursor ISI
post-cursor ISI
0
T
22
d-1
d0
output pulse:
# d &
X% 1 (
$ d0 '
Result with 0
pre-cursor ISI:
23
Din (t T )
Time domain:
a1
a1 =
_
+
Dou t (t)
1+ a12
Frequency domain:
H(s) = 1 a1 e sT
H( j ) = 1 a1 e jT
d1
d0
1 a1
= 1 a1 cos T j sin T
24
Din (t)
a0
a1
a2
an
FFE can cancel both pre- and post-cursor distortion.
Dou t (t)
25
R
_
R
Vout +
Din+ (k)
Din (k)
V0
Din+ (k 1)
Din (k 1)
V1
Din+ (k 2)
V2
Din (k 2)
ISS
26
27
Adaptation (1)
Assume original sequence Din(k) is known.
Define error signal e(k) as:
^
e(k) D^out (k) Dout (k) where Dout(k) is an appropriately delayed version of Din(k).
de2
Steepest Descent Algorithm: Set ai =
dai
e2
a2
step size
a1
optimum
setting
28
Adaptation (2)
Least mean-square (LMS) algorithm:
Set ai =
d e 2 (k)
dai
] = 2 D^
d e 2 (k)
dai
out
Dout
dDout
dai
= 2 e(k) Din (k i )
Analog version of LMS:
ai (t) =
e(t) D
in
(t iT )dt
ai = 2 e(k) Din (k i )
both signals are
available on chip.
29
Adaptation (3)
Types of adaptation:
1. Training Sequence
A training sequence with known properties is sent through the channel +
equalizer. The equalizer output is compared to the original sequence and an
error signal is generated.
2. Blind Adapation
Adaptation is continually performed while system is running. Only limited
properties of the signal are known. An error signal must somehow be
generated without having the original sequence.
30
Adaptation (4)
Generation of error signal:
Din
FFE
Dout
Dout
e
Slicer restores logic levels and opens eye vertically.
Bit sequences at slicer input & input
are identical.
Slicer has no effect on placement of zero crossing.
Slicer can be realized using CML buffers with sufficient gain and speed.
EECS 270C / Winter 2014
31
FFE structure:
a0
a1
a2
an
Dou t (t)
Noise applied to FFE input will be retained (perhaps filtered) at the output.
Din (t)
DFE structure:
Dou t (t)
+
- -
bm
b2
T
EECS 270C / Winter 2014
b1
T
Prof. M. Green / UC Irvine
T
32
Dou t (t)
bm
b2
b1
33
Dou t (t)
+
- -
post-cursor
distortion
1-tap example:
2/3
Din (k)
bm
b2
Dout (k)
(desired)
consistent
with
b1
1/3
Dout (k 1)
Din (k)
1
Dout (k 1)
3
b1 =
2/3
1
3
34
FFE + DFE
Din (t)
a0
T
a1
T
a2
an
Dou t (t)
+
-
bm
b2
b1
35
Vin
PGA
VA
ADC
Dout [1:n]
ADC requires strict control
over its input amplitude VA.
VC
AGC
Automatic Gain Control
Programmable Gain Amplifier (PGA):
36
PGA Design
1. Differential Pair:
Iout-
2. Source Degeneration:
Iout-
Iout+
Vin+
Iout+
Vin+
Vin-
Vin2RS
ISS
+
V_out
Rf
+
VC_
For biasing in weak inversion:
ISS
2nVT
% V V (
= ID0 exp' C T *
& nVT )
Iout =
Iout = gmVin
ISS
gm
Vin
1+ gmRS
Vin
RS
Vout =
Rf
Vin
RS
37
Vout =
Realization of RS:
Rf
Vin
RS
RS1 = 420
RS1 + RS2 = 529
2 dB steps
= 4200
38
Av = gmR
= N nCox
W
IO R
L
39
40
Simulation Results
T/H differential output for fin = 1.5 GHz and fs=10 GS/sec
41
High-speed Comparator
High-Level Clocking:
Improves isolation between the input
and output, reducing kickback from
output.
Cascoding of the clock switches
reduces the Miller effect of the input
transistors.
Reduced headroom
42
43
T/H output
Metastable event
Ct
Ct
+
v1
Ct
+
v2
dvd
v
= d + gmvd
dt
R
% t (
vd (t) = vd (0) exp' *
& m )
m
vd (t)
gm
Ct
1
1
1
gmR
gm = 4.2 mS
R = 400
Ct = 36 fF
m = 30 ps
45
11
10
01
00
-Vdec
0
3 Vin (analog)
2
VLSB
2
=
AcompVLSB
+Vdec
2
2Vdec
VLSB
Vdec = minimum detectable logic level
= minimum input at t = 0 so that output
level is Vdec at t = T/2
46
Recall:
$ t '
vd (t) = vd (0) exp& )
% m (
m = 23 ps
m = 30 ps
vd (Tc 2) Vdec
'T *
Vdec = exp) c ,
( 2 m +
' T *
= Vdec exp) c ,
( 2 m +
2
AcompVLSB
& T )
2Vdec
exp( c +
AcompVLSB
' 2 m *
47
Latch output
48