SDA QAM Overview

1
(No Transcript)
2
Overview

• What is QAM?
• Why Use QAM?
• Quadrature Amplitude Modulation
• Bits and Symbols
• QAM Encoding and Implementation
• QAM Measurement
• What Constellations Tell Us
• Modulation Error Ratio (MER)
• BER
• FEC

3
Why Go Digital?

• Cable and Terrestrial TV signals are going
  digital
• Digital Cable - Now Terrestrial Xmit - 2006
• Standard Definition TV (SDTV)
• High Definition TV (HDTV)
• Better Picture and Sound Quality
• Cable Modems transmit and receive digital data
• Digital signals can be less susceptible to noise
• Data Compression, error detection and correction
  is done with digital data
• Datacasting easily multiplexed into digital
  signal
• Higher Data Security

4
Analog vs. Digital

• Analog signal components are visibly discernable
  using a spectrum analyzer

• Digitally modulated signals only show a
  haystack on a spectrum analyzer regardless of
  modulation or content (more tools needed)

5
Digital TV Waterfall Graph
6
Effect of Noise on Analog Systems
(Gradually poorer C/N)
7
Effect of Noise on Digital Systems
(Gradually poorer MER)
Noise has very little affect on digital systems
until the system fails completely
8
Modulation formats in Cable
9
What is QAM?

• Quadrature Amplitude Modulation pronounced as
  kwam)
• Modulation Scheme where Phase and Amplitude are
  modulated to represent data
• Similar to QPSK which is robust and has been used
  for years (QPSK is the same as 4QAM)
• By providing different levels of amplitude and
  phase modulation, groups of bits can be
  represented as a symbol.
• Additional levels of modulation provide higher
  data capacity (16QAM, 64QAM, 256QAM, 1024QAM)

10
Why Use QAM?

• QAM is the standard for DOCSIS and DVB-C
• Improves spectral efficiency thereby providing
  more channels within a limited bandwidth
• 64 QAM can transmit 27Mbps or the equivalent of 6
  to 10 analog channels or 1 HDTV signal over one
  6MHz bandwidth
• 256 QAM can transmit 38.8 Mbps or the equivalent
  of 11 to 20 analog channels or 2 HDTV signals
  over one 6MHz bandwidth
• An SD signal requires 2 to 3.5Mbps (depending on
  quality) and an HD signal requires 19.2 Mbps.
• New compression techniques can provide up to 3 HD
  signals on a 256 QAM carrier

11
Data over Cable
11100100100
Mod.
Demod.
11100100100
12
Quadrature Amplitude Modulation

• Both I and Q are at the same frequency but
  amplitude and phase are modulated.
• I Incidental or in-phase Axis
• Q Quadrature Axis (90 degrees to I)
• Modulated Amplitude Levels
• Four different levels for 64 QAM
• Eight different levels for 256 QAM
• I and Q can be in phase (I 0 degrees, Q 90
  degrees) or out of phase (I 180 degrees, Q 270
  degrees)

13
Quadrature Amplitude Modulation
RF-Out 64-QAM
14
64 QAM Waveforms

• I and Q are in phase or 180 degrees out of phase
• I and Q are four discrete independent levels

15
Quadrature Modulation

• Simply measuring the carrier level relative to
  the noise level does not take into account any
  phase noise that may also be present on the signal

Carrier Amplitude Modulation
Carrier Amplitude Modulation
Analog Video AM Modulation
Carrier Phase Shift
QAM Modulation
16
QAM
17
Bits and Symbols

• A Symbol is a waveform that represents one or
  more bits
• Data is encoded into symbols for transmission
• Symbol Rate Bit Rate/Number of Bits per Symbol
• Assume a 8 bit sampler at 10kHz (voice) -Bit
  rate is 80Kbps

18
Forward Error Correction (FEC)

• Adds redundant information to the data stream
• Trade-off of data size vs error correction
• Trellis Encoding
• Randomization
• Interleaving
• Reed Solomon

19
FEC Made Easy
Alternates odd even, sum is 100
20
How FEC works

• Video Stream 1011100010110100
• Stream with FEC 1011100010010100111111000

After Transmission with bit error
21
Digital Modulation Stream
Reed-Solomon Encoder
Reed-Solomon Coding provides block encoding and
decoding to correct up to three symbols within an
RS block
Interleaver
Interleaving evenly disperses the symbols,
protecting against a burst of symbol errors from
being sent to the RS decoder
Digital Modulation Stream
Randomizer
Randomizes the data on the channel to allow
effective QAM demodulation synchronization
Trellis Encoder
Trellis Coding provides convolutional encoding
and with the possibility of using soft decision
trellis decoding of random channel errors
Modulation
22
QAM Measurements

• Spectrum Digital Average Power Level
• MER
• BER
• Constellation Display
• QAM Ingress
• Group Delay
• In-Channel Frequency Response
• Equalizer Stress
• Sweep

23
Digital Average Power Level Measurements

• Digital Average Power Measurements and
  Measurement Bandwidth
• The spectrum analyzer view is an excellent tool
  to see discreet RF-carriers.
• Caution is needed when viewing digital modulated
  signals (noise mountain). The signals level is
  depended from the selected measurement bandwidth
  (resolution bandwidth). At a RBW 300 kHz, a
  64QAM - 6 MHz wide digital signal reads in the
  spectrum analyzer trace 3 dB to low.
• The Average Power principle takes little slices
  from the integrated RF-energy, summing them
  together to one total power reading in the
  LEVEL-mode.

Summing slices of the total integrated energy
Analog and digital (broadcast) signal. The delta
in level should be 10 dB.
24
Spectrum analyzers can cause confusion

• The spectrum analyzers different
  resolution-bandwidth filter give different
  results for power level measurements.

25
Level meters that use correction factors can be
inaccurate Averaging over time. Unreliable
method, not according to the standard
t
26
Level measurements on digital video channels

• Average Power Level according to standards
• Scanning the level envelope of the channel using
  a 280 kHz IF-filter and summing the values of all
  samples.
• Can be used on all digital channels QPSK, QAM,
  8-VSB

gt 10 dB
27
Make sure you setup the right measurement
bandwidth
28
Modulation Error Ratio (MER)

• Analogous to S/N or C/N
• A measure of how tightly symbols are recorded
  with respect to desired symbol location
• MER(dB) 20 x log RMS error magnitude
  average symbol magnitude
• Good MER
• 64 QAM 23 dB MER
• 256 QAM 29 dB MER

Average symbol magnitude
RMS error magnitude
29
MER

• Modulation Error Ratio (MER) in digital systems
  is similar to S/N or C/N used in analog systems
• MER determines how much margin the system has
  before failure
• Analog systems that have a poor C/N show up as a
  snowy picture
• A poor MER is not noticeable on the picture right
  up to the point of system failure - Cliff
  Effect
• Cant use the TV as a piece of test equipment
  anymore

30
Effect of Noise on Analog Systems
(Gradually poorer C/N)
31
MER? Modulation Error Ratio (dB)(EVM? Error
Vector Magnitude) ()
Amplitude and phase error

• Equivalent to analog C/N
• The bigger the number the closer to the target.
• Field test 32 - 35dB.
• Set top boxes 28dB.
• Headend gt 40dB.
• Bad MER Bad BER

32
What is a Good MER?

• A 64-QAM signal requires better than 23 dB MER at
  the set top box or CM to operate
• A 256-QAM requires better than 28 dB MER at the
  set top box or CM to operate
• A 1024-QAM signal requires better than 33 dB MER
  at the set top box or CM to operate)
• To allow for degradation a margin (or headroom)
  of at least 3 to 4dB is preferred

33
Error Vector Magnitude
34
Error Vector Magnitude (EVM)

• EVM is defined as follows
• Expressed in percentage

Error Magnitude
Ideal Symbol
Max Symbol Magnitude
35
BER Introduction

• Bit Error Rate is a major indicator of system
  health
• As data is transmitted some of the bits may not
  be received correctly
• The more bits that are incorrect, the more the
  signal will be affected
• Its important to know what portion of the bits
  are in error
• Need to know how much margin the system has
  before failure
• The harder FEC is working, the closer the system
  is to failure (The Cliff)

36
BER

• Good signal BER 10-10
• Threshold for visible degradation BER 10-6
• FEC can improve BER from 10-4 to 10-10
• BER before FEC correctable uncorrectable
  errors
• BER after FEC uncorrectable errors
• Bit Error Tester (BERT)
• Inject known signal

37
BER Example

• A 256QAM channel transmits at a symbol rate of 5M
  symbols per second
• Bit rate 8 bits per symbol X 5M symbol per
  second 40M bits per second
• Error Incident Bit rate X BER Errors Per
  Second

38
Pre and Post FEC BER

• FEC - Corrected Errors Estimated uncorrected
  Errors
• Pre FEC corrected uncorrected errors
• Post FEC uncorrected errors
• Pre and Post FEC BER indicate how hard the FEC is
  working to correct errors

39
Bit Error Rateprovides benefit for commissioning

• Number of bad bits for every good bit.
• Forward Error Correction when working will output
  gt10-11
• 1 error in 100 billion bits
• 1 error every 42 minutes
• MPEG-2 likes good BER
• FEC will work to about 10-4
• 1 error in 10000 bits
• 1 error every 276 uses
• FEC causes Cliff Effect

40
FEC causes Cliff Effect

• A small variation in MER (/- 1 dB) will cause a
  large variation in BER measurement.
• Using BER for trouble-shooting and fault location
  is not repeatable and very inaccurate.

41
C/N vs. BER vs. MER
42
Constellations
43
Constellation Basics

• The constellation display shows both I and Q
• A symbol is the smallest piece of information
  transmitted - plotted as a point representing a
  digital bit(s)
• It is the digital equivalent of a Vectorscope
  display
• Useful for determining modulation problems
• Amplitude Imbalance
• Quadrature Error
• Phase Error
• Modulation Error Ratio

44
Quadrature Amplitude Modulation
RF-Out 64-QAM
45
Typical Constellations
Decision Boundary
16 QAM
64 QAM
256 QAM
46
Constellations, Symbols, and Digital Bits

• Each dot on constellation represents a unique
  symbol
• Each unique symbol represents unique digital bits
• Digital data is parsed into data lengths that
  encode the symbol waveform.

16 QAM
47
Gain Compression

• If the outer dots are pulled into the center
  while the middle ones are not affected, the
  signal has gain compression
• Gain compression can be caused by IF and RF
  amplifiers and filters, up/down converters and IF
  equalizers

Outer edges pulled in
48
System Noise

• A constellation displaying significant noise
• Dots are spread out indicating high noise and
  most likely significant errors
• An error occurs when a dot is plotted across a
  boundary and is placed in the wrong location
• Meter will not lock if too much noise present

Dots are spread out showing error
49
Phase Noise

• Display appears to rotate at the extremes
• HE down/up converters can cause phase noise
• Random phase errors cause decreased transmission
  margin
• Caused by transmitter symbol clock jitter
• Bad LO in meter can cause phase noise
  Constellation

50
Coherent Interference

• If the accumulation looks like a donut, the
  problem is coherent interference
• CTB, CSO, spurs and ingress
• Sometimes only a couple dots will be misplaced
• This is usually laser clipping or sweep
  interference

51
Ingress Under the Carrier

• Interference will cause poor MER
• Noise
• Discreet Signal
• Ingress
• Bad Modulator
• CSO/CTB (TV)
• CSO/CTB Digital

52
QAM Ingress (Ingress Under the Carrier)

• Meter knows how much error is in signal from
  measuring Constellation points
• Meter uses this error to plot Ingress Under the
  Carrier

53
CSO and CTB under QAM 256 carrier

• Using ingress under the carrier, the SDA can
  uncover CSO and CTB that are not visible using
  standard spectrum analysis.

54
CSO under QAM 256 carrier
55
CTB under QAM 256 carrier
56
Group Delay

• Definition Group delay is the measure of the
  slope of the phase shift with frequency.
• Effects If there are group delay variations in
  the network, then signals of one frequency can
  make it through the network faster than signals
  at another frequency.
• For analog signals this typically can cause
  misregistration of the chrominance to luminance
  since the chrominance subcarrier is 3.58MHz
  higher than the luminance carrier. The visible
  effect is that the colors are not within the
  outline of the subject.

57
Group Delay

• For digital signals the effect can lead to QAM
  symbol misinterpretation. The net effect is that
  short duration pulses that are input into the
  network will exit the network having a longer
  duration. This spreading leaves energy from one
  pulse in the time slot of other pulses. This
  causes the BER to degrade.
• For downstream carriers, the DOCSIS 1.0 spec
  requires the group delay ripple to be less than
  75nS.
• Bad filters are a typical cause of group delay

58
In-Channel Frequency Response

• In-Channel Frequency response is amplitude
  ripple. This means that signals at one frequency
  are attenuated relative to signals at another
  frequency.
• For downstream digital carriers DOCSIS 1.0
  specifies a max ripple of 0.5dB in 6MHz. DOCSIS
  1.1 has relaxed this specification to 3.0dB in
  6MHz.

59
Equalizer Stress

• Digital demodulation receivers utilize adaptive
  equalizers to negate the effects of signals
  arriving other than the desired signal.
• Signals can arrive ahead of or after the desired
  signal. In a cable system, the majority of
  signals are reflections and micro-reflections
  that arrive after the desired signal.
• Cable modems and digital set top boxes must be
  able to handle pre and post signals at levels
  defined by DVB standards. If the equalizer is
  pushed beyond those limits, errors will occur.
• By using the Velocity of Propagation, the
  distance to the source of the reflection can
  sometimes be located. If the reflections occur
  before the next upstream amplifier, they are
  simply amplified and passed downstream thereby
  eliminating the ability to perform fault
  detection based on reflection time.
• Equalizer stress is used more as a figure of
  merit for the margin available to the set top box
  or cable modem.

60
Equalizer Stress
Signal arriving about 0.8usec before desired
carrier
Signal arriving about 2usec after desired carrier
61
What faults cause CATV signals to fail ?(80-90
of the time, the same faults)

• Success rate of finding and fixing the following
  problems using
• Signal Levels
• TILT
• Gain / Loss
• Suck-outs (notches)
• C/N
• HUM
• CTB/CSO Intermodulation
• CPD - Forward and Reverse
• Reverse Ingress
• BER / MER
• Reflections / Standing waves

Source Research 11/97-2/98 Market survey with
200 US and European CATV operators
62
Sweep is the best way to prepare the network for
256 QAM

• Standing waves, suck-outs, intermodulation
  distortion and non-linear performance effect
  digital performance

Bad Forward Sweep Trace
63
Reflections causes by bad terminations
f
Bad Forward Sweep Trace - Standing waves

• Reflections or standing waves caused by any
  defective, miss-matching devise
• Damages cable, connectors ground block,
  splitters, etc.
• A sweep signal is transmitted by the SDA 5500
  over coaxial cable (the medium). A portion of the
  transmitted sweep signal on the cable will be
  reflected back to the transmitter if the load is
  not a perfect 75Ohm impedance match. The
  reflected energy will be the same frequency as
  the incident (sweep) signal but different in
  phase. The resulting signal (incident
  reflected) will appear as standing waves on a
  frequency sweep (see figure). The reflection is
  such that the peaks of the individual cycles can
  be translated to distance to the fault (impedance
  mismatch) through the following equation
• D 491Vop/f
• Where Ddistance to fault,
  Vopvelocity of propagation of the cable, and f
  frequency of 1 cycle of the standing wave.

64
Suck-outs

• Bad taps or connectors are mostly causing a
  suck-out (notch) in frequency response.
• It generates individual channel errors, Sweep is
  a very efficient way to locate bad taps or
  connectors. Scanning the channels works too, but
  the error is less apparent.
• Causes are
• Humidity problems
• Small RF leaks to mass.
• Bad mounted connectors

Bad Forward Sweep Trace - Suck-out
Bad Level SCAN-Trace Trace - Suck-out
65
Terms

• QAM - Quadrature Amplitude Modulation
• Symbols - Collection of Bits
• Symbol Rate - Transmission Speed
• I Q - Components of QAM data
• Constellation - Graph of QAM Data
• MER - Modulation Error Ratio
• BER - Bit Error Rate
• FEC - Forward Error Correction

66
QAM Data Capacity (Annex B, 6MHz)
67
(No Transcript)