Professional Documents
Culture Documents
0
Rev.A00
DOCSIS 3 Overview
DOCSIS 3 Benefits
Troubleshooting Summary
DOCSIS 3.0
Market Trends
Media Convergence
Source: Future services on HFC networks: 33th PIKE Conference, 14 October 2008, Zakopane, Poland
DOCSIS 3.0
Digital
Photos
Home
Networks
Data &
VoIP
Gaming
MP3
WMV
VOD
DVR/PVR
DVD
Blu-ray
You Tube
SDTV
HDTV
DOCSIS 3.0
Mobile
Video
iPod
Walkman
DOCSIS 3.0
Customer
Demand
Competitor
Offering
IPTV, Netflix,
Blockbuster, SIP
Video, Gaming, You
Tube (HD), Video
Phone (HD) ...
FTTx, GPON,
VDSL2, FiOS,
Wireless
Business
Services
IP Addresses
needed
T1/E1 solutions
HD Video
Conferencing
IPv6
Telcos are deploying VDSL2, GPON, FIOS and FTTx (USA & Europe)
Revolutionary steps need to decide if and when to implement a Next Generation HFC
network
DOCSIS 3.0
DOCSIS 3.0
DOCSIS Overview
DOCSIS 3.0 Benefits
DOCSIS Milestones
DOCSIS 1.0 (1999)
1st products certified (CableLabs started project in 1996)
Open standard for high-speed data over cable
Modest security, Best-effort service
DOCSIS 1.1 (2000)
Quality-of-Service (QoS) service flows
Baseline Privacy Interface (BPI+) Certificates
Improved privacy & encryption process
DOCSIS 2.0 (2002)
Improved throughput & robustness on Upstream
64/128 QAM modulation & higher symbol rates with FEC
Programmable interleaving to upstream channels
DOCSIS 3.0 (2006)
Channel bonding (4U/4D) for increased capacity
IPv6 support
Improved security (AES)
DOCSIS 3.0
Increased Upstream
bandwidth
IPv6 Support
Backwards compatibility
CMTS qualification
Modem certification
Additional network
security
DOCSIS 3.0
10
EuroDOCSIS
Version
Downstream
Upstream
1.x
2.0
3.0 (4 Channels)
3.0 (8 Channels)
Notes:
DOCSIS 3.0
11
Bonded
together for
higher
aggregate
speed and
capacity
DOCSIS 3.0
4D/4U =
Can be
150Mb/s
deployed
downstream
incrementally
120Mb/s
upstream
No upper
limit to # of
channels
HFC subsplit
effectively
limits #
upstream
channels
Existing
DOCSIS
modems
share
channels
with no
negative
impact
12
DOCSIS 3.0
13
14
RF
Bandwidth
Availability
DOCSIS 3.0
Headend
and Core
Network
Equipment
Preparation
Verify
QAM64
Upstream
Txmission
Verify
QAM256
Downstream
Txmission
DOCSIS 3.0
Modem
Emulation
IP/Ethernet
Testing
(Ping, FTP,
RFC2544,
Web)
15
Launch digital
only systems
Use unusable
old analog
broadcast
channels
Launch Digital
Simulcast and
migrate selected
analog channels
Use Switched
Digital Video to
reclaim
bandwidth
DOCSIS 3.0
Move test
carriers to
alternate
frequencies
DOCSIS 3.0
requires a minimum
of 4 to16 downstream
channels
16
Soon 1GHz
Reclaiming bandwidth:
Switched Digital Video
Test requirements:
Downstream expanding to 1GHz
MPEG 4 video
DOCSIS 3.0
17
Upstream Expansion
How much gain?
250Mb/s
500Mb/s
1000Mb/s
DOCSIS 3.0
18
Solutions are typically driven by specific technical, geographical or local market factors
Source: Michiel Peters, TNO - Benelux Chapter SCTE , 15 September 2008, Amsterdam
DOCSIS 3.0
19
DOCSIS 3.0
Plant Qualification & Test Methods
20
DOCSIS 3.0
21
Plant Qualification
Upstream Testing
Generate QAM signal in RP to verify attenuation, level, Tilt, MER and BER
Frequency response, Group delay, Constellation and Adaptive equalization
Check spectrum for ingress, noise, CPD and laser clipping
Check for modems transmitting excessive levels due to high value taps
Downstream Testing
Forward Sweep (Sweepless), frequency response, amplifier tilt
MER, CNR, Group Delay, Constellation, BER pre/post errors
MPEG-2 Video Signal Analysis
Useful Tips
Qualify the plant on a node by node basis
Cable drops should be Tri or Quad shielded
Check for leakage & improve thresholds (< 5uV/m is recommended)
Use the divide-and-conquer technique to locate problems
Avoid downstream/upstream frequencies near the band edges/roll off
Avoid downstream/upstream frequencies susceptible to ingress/interference
DOCSIS 3.0
22
DOCSIS 3.0
23
DOCSIS 3.0
24
DOCSIS 3.0
25
Level and
MER look
OK?
DOCSIS 3.0
26
Constellation diagram
DOCSIS 3.0
27
Constellation diagram
DOCSIS 3.0
28
Constellation
MER
64-QAM: 27 dB min
256-QAM: 31 dB min
BER
Pre/Post FEC
DOCSIS 3.0
29
Impairments
At room temperature, in 6 MHz bandwidth and 75 ohms circuit, the thermal noise is
approximately -60dBmV. After amplification, the noise level can get much higher.
All the other impairments are human made, they depend on the design, implementation
and operation of all the elements in the signal chain
DOCSIS 3.0
30
In-channel ingress
Data collisions
Laser clipping
DOCSIS 3.0
31
Input to Lasers 39 dB
Output of Nodes 37 dB
10 log
DOCSIS 3.0
32
Downstream Performance
Pre/Post FEC BER
DOCSIS 3.0
33
Notes on FEC
To have an accurate idea of the BER performance you need to know both pre
and post FEC bit error rate
Forward error correction (FEC) is a digital error checking system that sends
redundant information with the payload so the receiver can repair corrupted data
and eliminate the need to retransmit.
By using the same Reed Solomon decoder at the receiving end, bit errors can be
detected these are called Pre-FEC errors
Pre FEC BER is the error rate of the incoming signal prior to being corrected by
the FEC circuitry - a minimum of 1x10-7 is expected, but FEC may be able to
correct errors as high as 1x10-6.
Post FEC Bit errors are not acceptable and should be corrected
Both Pre and Post FEC BER need to be verified in order to determine if the FEC
circuitry is working to correct errors and if so how hard.
DOCSIS 3.0
34
DOCSIS 3.0
Quadrant 4
Quadrant 1
Quadrant 3
Quadrant 2
35
2
N 2
I j j
j 1
MER 10 log 10 N
2
2
I j Q j
j 1
Average error
power
Average symbol
power
A large cloud of
symbol points means
low MERthis is not
good!
A small cloud of
symbol points
means high MER
this is good!
Source: Hewlett-Packard
DOCSIS 3.0
36
DOCSIS 3.0
Modulation
Type
Std. Symbol
Rate (MHz)
Annex A
(8MHz)
QAM64
6.952
41.4
Annex A
(8MHz)
QAM256
6.952
55.2
Annex B
(6MHz)
QAM64
5.057
38
Annex B
(6MHz)
QAM256
5.361
43
37
Bandwidth
(MHz)
Modulation
type
1.0
3.2
QPSK
5.12
1.1
3.2
QPSK
QAM16
5.12
10.24
2.0
6.4
QAM16
QAM64
10.24
30.72
3.0
6.4
QAM64
QAM128
120
(4 channel bonding)
Standard symbol rate (bandwidth): 1.28 (1.6), 2.56 (3.2), 5.12 (6.4) MHz
DOCSIS 3.0
38
Constellation Display
DOCSIS 3.0
39
Decision Feedback Equalizer (DFE) - output is fed back to input, 108 taps long
DOCSIS 3.0
40
Impairment
Results
Tap Expert
DOCSIS 3.0
41
42
Group Delay
General Notes:
Amplitude and Group Delay responses help visualize the effects of filters, diplexers, traps, suck-outs in the
signal path, from (and including) the QAM modulator up to the point of test.
The frequency span of the calculated responses is directly related to sampling period of the Equalizer
Symbol period. For QAM-64, the span response is 5.05 MHz, while for QAM256 the span is 5.36 MHz
DOCSIS 3.0
43
Equalizer Stress
Noise Margin
DOCSIS 3.0
44
Linear Distortions
45
Operational RF Levels
DOCSIS 3.0
46
DOCSIS 3.0
47
DOCSIS 3.0
48
DOCSIS 3.0
49
50
DOCSIS 3.0
51
Copper (10/100/1000BaseT)
& Fiber (1000BaseX) based
Ethernet service should be
verified
DOCSIS 3.0
52
53
54
Signal
Level
Meter
DOCSIS 3.0
CM Analyzer
QAM-64
USG
Source
Digital
Spectrum
Analyzer
Group
delay &
frequency
response
tester
CX380
can do
it all
DOCSIS 3.0
Adaptive
equalizer
tester
Ethernet
Tester
CX350
can do
it all
55
RF Test Checklist
Constellation
display
Pre / Post
FEC BER
Linear
Distortions
Signal Level
problems
Transient
Impairments
Low MER or
CNR
Sweep
transmitter
interference
Adaptive
equalizer
graph
Analog
channel
signal level
Pre/Post-FEC
BER
Phase noise
Laser
clipping
In-channel
frequency
response
Digital
channel
power
Constellation
display
I-Q
imbalance
Loose
connections
In-channel
group delay
Upstream
transmit level
Upstream
packet loss
Coherent
interference
(ingress)
CPD
Constellation
display
Unequalized
Constellation
display
Gain
compression
Low MER or
CNR
MER
Unequalized
Equalizer
Graph
Microreflections
Laser
clipping
Sweep
transmitter
interference
DOCSIS 3.0
56
Troubleshooting
Integrated Up-converter
CMTS
To headend downstream
combiner
Attenuator
(if required)
DOCSIS 3.0
57
Troubleshooting
External Up-converter
CMTS
44 MHz IF input to
upconverter
(typ. +25 dBmV to +35
dBmV)
44 MHz downstream
IF output
(e.g., +42 dBmV +/-2 dB)
DOCSIS 3.0
Attenuator
RF upconverter
58
Check signal levels and BER at downstream laser input and node output
Bit errors present at downstream laser input but not at CMTS or up-converter
output may indicate sweep transmitter interference, loose connections or
combiner problems
Bit errors at node output but not at laser input are most likely caused by
downstream laser clipping
DOCSIS 3.0
59
Troubleshooting Tips
Residential wireless
networks may limit
DOCSIS 3.0
performance benefits
Routers, Switches,
and Ethernet cards
can limit bandwidth to
100Mbps or 10Mbps
PC performance can
effect or limit
throughput
Ensure Test
Equipment has
sufficient bandwidth
to perform high
throughput test
Hardware settings
can effect bandwidth
e.g. MTU size
DOCSIS 3.0
60
61
The phase and gain of both the I and Q carrier must be equal in order for the
constellation to be correct.
The gain difference between the 2 carriers should be less than 1.8% and the phase
difference should be less than 1 degree.
Phase Noise
DOCSIS 3.0
62
Hum
Same as hum on analog carriers, if the level is the same, its the system, if higher on the
digitals then its probably the QAM modulator
Echo Margin
A measurement in dB of how far the taps are from the template with the time equalizer
measurement.
DOCSIS 3.0
63
Group Delay
Different frequencies travel through the same medium at different speeds. So the lower
the lower frequencies of the same carrier arrive at the receiver at different timing than
the higher frequencies.
Frequency Response
Carrier Offset
DOCSIS 3.0
64
5 MHz
10 MHz
10 MHz
15 MHz
40 MHz
45 MHz
50 MHz
55 MHz
60 MHz
65 MHz
DOCSIS 3.0
15 MHz
20 MHz
25 MHz
35 MHz
10 MHz
15 MHz
20 MHz
30 MHz
5 MHz
20 MHz
25 MHz
HFC
(Filters,
Taps)
30 MHz
35 MHz
40 MHz
25 MHz
HFC
(Filters,
Taps)
45 MHz
50 MHz
55 MHz
60 MHz
65 MHz
Confidential & Proprietary Information of VeEX Inc.
30 MHz
35 MHz
40 MHz
45 MHz
50 MHz
55 MHz
60 MHz
65 MHz
65
Linear Distortions
In-Depth Understanding
ECHO MARGIN
The Coefficients of the Equalizer will also reveal the presence of an Echo, (a.k.a. micro-reflections). The Equalizer
will cancel such an echo, and in doing so, the equalizer coefficient which corresponds to the delay of the echo will
be much higher than the surrounding ones, it sticks out of the grass. The relative amplitude of this coefficient is
an indication of the seriousness of the echo, and its position gives the delay of the echo, hence its roundtrip
distance.
The Echo Margin is the smallest difference between any coefficients and a template defined by Cablelabs, as a
safety margin before getting too close to the cliff effect. It is normal to notice relatively high coefficients close
to the Reference as this corresponds to the filters in the modulator / demodulator pair and to the shape of QAM
signal.
EQUALIZER STRESS
The Equalizer Stress is derived from the Equalizer coefficients and indicate how much the Equalizer has to work to
cancel the Linear distortions, it is a global indicator of Linear distortions. The higher the figure, the less stress.
NOISE MARGIN
We all know that the lower the MER, the larger the probabilities of errors in transmission (Pre-FEC and then PostFEC); the MER degrades until errors are so numerous that adequate signal recovery is no more possible (cliff
effect). As Noise is a major contributor to the MER, we define Noise Margin as the amount of noise that can be
added to a signal (in other words, how much we can degrade MER) before get dangerously close to the cliff effect.
Noise is chosen because on the one hand it is always present, and on the other hand it is mathematically
tractable. Other impairments, such as an Interferer, are not easily factored into error probabilities.
DOCSIS 3.0
66
Linear Distortions
In-Depth Understanding
EQUALIZED MER vs. UN-EQUALIZED MER
The MER (Modulation Error Ratio) is the ratio of the QAM signal to Non-Linear distortions of the incoming QAM
signal. The MER should have included the Linear distortions to indicate the health of the signal; but the QAM
demodulator cannot operate properly without the Equalizer and the Equalizer uses the MER as a tool to
adaptively cancel the Linear distortions. Consequently it is convenient to distinguish the MER (non-linear
distortions only) from an Un-equalized MER (non-linear and linear distortions), the Un-equalized MER is
calculated from the MER and Equalizer Stress.
The Un-equalized MER is always worst than the MER. A small difference between the two indicates little Linear
distortions, a large difference shows that there are strong Linear distortions. Even if the Linear distortions are
cancelled by the Equalizer, we have to keep in mind that the Equalization is a dynamic process as it tracks Linear
distortions by trial and error even after converging. The larger the Linear distortions the larger the tracking
transients are, hence more probability of transmission error (pre-FEC or Post-FEC BER).
PHASE JITTER
Phase Jitter is caused by instability of the carrier of the QAM signal at the demodulator. This instability could be
found at the QAM modulator and up-converter or in the QAM receiver (Local Oscillators used in frequency
conversions). The phase jitter introduces a rotation of the constellation, where the symbols clusters elongate and
get closer to the symbols boundary. Eventually some symbols will cross the boundary and cause an error in
transmission. The QAM demodulator has a Phase lock loop to track phase variations of the carrier; it tracks easily
long term drift as well as some short terms variations (up to 10 or 30 kHz) but it cannot track very fast variations
above its loop response. So in a QAM demodulator, the wideband jitter is more damageable than short term
jitter.
DOCSIS 3.0
67
Recommended Reading
www.cable360.net/ct/operations/testing/15170.html
DOCSIS 3.0
68
Thank You.
Any questions?
DOCSIS 3.0
69