10GEPON Burst Receiver
Download
Report
Transcript 10GEPON Burst Receiver
10GEPON Burst Receiver
Ad-hoc
Goals
• The goals of this ad-hoc are to resolve the following issues:
- TIA Gain
• Is different gain for 1G and 10G required?
– 10GEPON Burst Receiver architecture
• AC Coupled? or DC Coupled?
– Guard Time
• Is different Guard Time between Bursts of different data rate required?
Outline
•
•
•
•
Optical Receiver – Basic
TIA Parameters
LIA - AC vs. DC Coupled
Preamble Length
Optical Receiver Architecture
V(t)
V(t)
t
t
Clock
TIA
LIA
Data
i(t)
AGC
t
CDR
Offset
Compensation
10GEPON TIA
TIA Parameters
•
The primary function of the TIA is to convert the small current, from the
photodiode, into a voltage while adding as little noise as possible to the output
signal
APD TIA
optimized for 10G
Source 1:
1.25Gbps 8B10B Coding
1.25Gbps LIA
Source 2:
10.3125Gbps 64B66B Coding
?
•
10.3125Gbps LIA
TIA is characterized by several parameters:
– Transimpedance Gain
– Input Referred Noise
– Bandwidth
RFB
-A
Iin
CT
• One of the main “problems” of the TIA is the trade-off between Gain,
Noise and Bandwidth
Vout
TIA – Bandwidth, Gain, and Noise
•
All 3 parameters of the TIA are
function of the RFB
•
Bandwidth: The Transimpedance gain
is equal to the RFB, while the
Bandwidth is determined by the RC
time constant.
f
3dB
A 1
2 R FBC T
•
Gain: Transimpedance gain of the TIA
is the ratio of the output voltage to the
input current.
A
R FB
Z T A 1
R FB C T
1 j A 1
•
Noise: Noise contribution of the TIA is
characterized by the input referred
noise current
V n,out
I n,in
ZT
–
The input referred noise current is
related to the output noise voltage by
the following equation:
2
2
2
TIA Gain Burst Parameter - Different for
1G and 10G?
• In order to optimize the performance of the TIA in both 1G and 10G
we need to support Variable TIA Gain
• Gain can be varied between 1G and 10G bursts by changing the
feedback resistor
• In order to analyze the impact on TIA performance in 1G, we need
to calculate the SNR of the TIA
TIA – Input Referred Noise
• The ultimate limitation on the Optical Receiver sensitivity is the
Noise
• The noise includes the Photodiode’s Shot Noise and the noise
added by the TIA
• The major noise sources are the Feedback Resistor and Voltage
Amplifier
di²Rf
Rf
di²AMP
di²eq,in
CDiode
gm
Cin
Rout
Cout
Preamble length – function of maximum CID
(Consecutive Identical Digits)
•
The data signal has a sequence of consecutive high and low bits in the
middle of the sequence
•
The DC level during the consecutive bits begins to droop
•
Long sequence of consecutive bits can significantly change the DC level
of the data and the optimum threshold voltage
•
A poor low frequency cut-off vertically closes the eye diagram and can
reduce the sensitivity of the system
•
In order to achieve a lower low frequency cut-off, we need to extend the
number of preamble bits
•
For example, in GPON, the CID is 72 bits
TIA – AGC Loop Timing
Peak Detector output
Noise
idiode
TIA_Gain
Amplified Noise
TIA_Output
• TIA AGC initialization time parameter needs to be much longer than
the CID bit time
• During “0” CID, the AGC loop should remain constant and not
“running” to infinite gain
• In between Bursts, the AGC needs long preamble, greater than
AGC_τ to enable AGC to “learn” new peak value
TIA – AGC Response Delay
V
1.2
1.0
0.8
0.6
0.4
0.2
0
-120
-80
-40
0
40
τagc
80
120
t
• Practical AGC has delayed response to signal-level
change
10GEPON LIA
The Problem – AC or DC Coupled
APD TIA
optimized for 10G
Source 1:
1.25Gbps 8B10B Coding
1.25Gbps LIA
Source 2:
10.3125Gbps 64B66B Coding
?
10.3125Gbps LIA
?
10G LIA – AC Coupled
• 10G LIA is simulated by the following
Transfer Function
• The lower cut-off pole (f1) determine
the CID length, while the higher cutoff pole (f2) determine the Bandwidth
• For the lower frequency, 3MHz was
simulated
• For the higher frequency, 7GHz was
simulated
• In order to maintains minimum DC
droop from the baseline, we need at
least factor 4 over the τ
3dB
3MHz
7GHz
s
H ( s)
2 f 1
1
s
s
1
1
2 f 1
2 f
2
10G LIA – DC Coupled
• In DC-coupled the RC (τ) is determined by
internal capacitors
• One capacitor for “fast” acquisition during
preamble - and the second for CID support
• During reception of the preamble, the threshold
acquisition done by the “high” frequency cut-off
pole, then switching to the “low” frequency cutoff pole to support CID
Transfer Function
3dB
Vout
Vin
S
AS 1
WL
S
S
1
AK
S 1
WL WH
A: forward gain.
K: feedback gain.
W L: low pass pole frequency (in feedback loop).
W H: high pass pole frequency (in forward path).
f1 = 3MHz
•
•
f2 = 80MHz
f1 and f2 determine how much DC droop
we are allowed from the baseline
During preamble we use f1_high then after
“short” time we switch to f1_low to support
CID
Preamble Length - A Formula
• Assuming that 4 time constants is needed
Npreamble
4 Ncid
ln(1 X )
Where:
N
N
CID
= Number of CID
preamble
= Required Preamble length [bits]
X
= Deviation of baseline permitted during CID
An example:
CID = 64
X = 0.1 (10%)
N=
-4 * 64
= 2430bits
ln(1 - 0.1)
For 5% droop
N=
-4 * 64
= 5000bits
ln(1 - 0.05)
Just to the LIA
Open Questions
• AC or DC Coupled?
• In AC Coupled
– What should be the Maximum overhead?
• In DC coupled
– What should be the Minimum overhead?
• Different Guard time between different Bursts?
Back up