Selected Physical Layer topics
Distinguish
guided (wired) from unguided (wireless)
Three
major guided media for data comm.
1.
twisted
pair
2.
coaxial
3.
fiber
optic
Twisted Pair
-
copper
wires in bundle grouped in pairs
-
max
frequency about 5 MHz.
-
wires
in pairs twisted for uniform noise effect
(effect based on distance from noise source, so equal avg distance)
-
UTP
(unshielded) or STP (shielded)
-
UTP
most common, categories 1 thru 5 based on quality and max xmit rate
-
Cat
1: low rates,
Cat 3: up to 10 Mbps (standard telephone, 3 twists/ft),
Cat 5: up to 100 Mbps (tighter twists)
-
Signal
attenuates (degrades) with distance
-
Connectors:
4-8 pin RJ (standard telephone); RJ45 handles 4 pairs
Coaxial
-
a.k.a.
coax
-
solid
copper core, braided copper surround, separated by insulator
-
max
frequency 500 MHz
-
various
RG grades: higher the number the thinner (and more flexible) the cable (lower
rates, shorter distances)
-
Thicknet (.5" diameter), Thinnet(.25"), baseband (dedicated bandwidth), broadband
(cable TV - FDM), 10BaseX, are terms
relevant to applications of coax
(mostly ethernet) See Ethernet chapter.
-
Connectors:
various barrel network connectors (BNC): bayonet (1/2 turn with lock),
threaded, slip-on.
-
Use
of T-connector or vampire tap
Fiber Optic
-
inherently
digital (light pulse).
-
Fast
but expensive; immune to electrical interference
-
current
bandwidth limited by signal generator/detector technology, not medium
-
core
fiber is very thin (typical 50-100 micrometers), fragile and relatively
inflexible
-
each
core has very thin cladding. Its significance explained below.
-
various
insulating configurations depending, on application/environment
-
Connectors:
mostly barrel. Must be very precise.
-
Signal
generation/detection. Light source
either: LED (weak, unfocused, short
distance, cheaper) or laser (the
opposite). Use photodiodes for detectors.
-
light
transmission in fiber optic:
-
light
travels in straight line until density of medium changes.
-
Fiber
cladding is less dense than core.
-
What
happens when light beam in core strikes cladding (at some angle)?
-
Depends
on the angle. It may either be refracted into cladding (bad) or reflected back into core (good).
-
Reflected
beam will bounce in this fashion until destination
-
Multimode
vs. single mode fiber:
-
Multimode: numerous light beams
emitted at different angles from source.
Some travel straight thru, some bounce along. Those that bounce along arrive later! Received signal is a little out of focus.
-
Single mode: fiber is so thin that only
beams traveling straight thru can fit.
Received signal is well-focused.
Only laser can do this. Very
expensive.
Major
distinction: analog vs digital.
Analog information is continuous
and has infinite number of values (traditional clock with moving hands,
electrical current into your house)
Digital information is discrete and
has limited number of values (clock with digital display, 0 or 1 for binary)
Analog signals
Signal
may be periodic (continuously
repeated pattern -- like sine wave) or aperiodic.
Periodic
signal characterized by:
Amplitude : value at any instant of
time
Frequency : number of cycles per unit
time
OR
Period : amount of time to complete
one cycle ( reciprocal of freq )
Phase : position of waveform
relative to time zero.
Basic
period unit is second.
Basic
frequency unit is Hertz.
Basic
phase unit is degrees or radians.
Can
be represented by composition of sine waves.
Digital Signals
Also
characterized by Amplitude, Period and Phase.
Represented
by square waves.
Can
be approximated using analog waves.
Time and Frequency Domain
Time domain graphs Time on X-AXIS,
Amplitude on Y-AXIS
Frequency domain graphs Frequency on X-AXIS,
Amplitude on Y-AXIS
Harmonics
Frequency
graphed as Harmonic : multiple of
"fundamental frequency"
Harmonic
is discrete.
Frequency domain shows
harmonic components of complex analog signal
Complex
signal is composition of sine waves, each having different harmonic and
amplitude. This are discovered using Fourier Analysis.
As
number of harmonics increases, the
approximation of original signal improves.
As
number of harmonics decreases, it becomes more difficult to accurately
represent and recognize the signal.
In general, transmission
rate limited by medium.
Bandwidth is width of frequency range
(spectrum) available to coding
signals.
For
binary (2 level) analog signals, the following (approximate) relationship
holds.
BANDWIDTH
= TRANSMIT RATE * # HARMONICS
If bandwidth is fixed, then increase in xmit rate must be offset by decrease in # harmonics transmitted.
Bandwidth
of medium generally partitioned into fixed size subbands (channels), so fixed
bandwidth is rule rather than exception.
Usually,
frequencies at channel boundaries
cannot be used (guard).
Analog-to-Digital happens
when analog (POTS) signal received at
Central Office
§
Analog
wave (from POTS line) encoded by digital pulses
§
Multistep
process:
1.
Quantize
(sample signal, assign integer value to sample)
2.
encode
to binary (binary string that represents integer value)
3.
use
D/D encoding to transmit (above)
§
Used
to transmit analog telephone data on digital T-lines (e.g. T-1, T-2).
§
PCM
(Pulse Code Modulation)
§
Sample
analog signal at fixed intervals
(using PAM - pulse amplitude modulation)
§
Quantize
signal value into integer in range -127 to +127
§
Integer
encoded into 8-bit sign-and-magnitude number
(bit 7 is sign, bits 0-6 are absolute value)
§
8-bit
value encoded for digital transmission
§
Two
key factors:
§
Number
of bits in encoding
§
Sampling
rate (limits ability to decode signal)
§
Would
like to small number of bits per sample, small number of samples
§
Small
bits per sample limits precision of decoded signal
§
Nyquist:
sampling rate must be at least double the highest frequency.
§
What
is minimum sampling rate for voice-grade telephone comm?
§
PCM
standard is 8000 samples/sec, 8 bits per sample (64000 bps -- not 64k!)
Digital-to-Analog signalling
used by modems (receive digital signal, transmit analog)
§
Most
common analog medium is voice-grade telephone line.
§
Encode
by modulating a fixed carrier signal
§
Recall
that signal has 3 components: amplitude, frequency, phase
§
Carrier
has known amplitude, frequency, phase
§
Receiver
compares received signal to carrier -- differences reveal the code
§
Amplitude
Shift Keying (ASK)
§
Encode
by modifying amplitude of carrier
§
Two
amplitude values: one for 1 and one for 0
§
Can
generate by: adding unipolar digital signal to carrier
§
Susceptible
to interference (voltage spike affects amplitude)
§
One
bit per baud
§
With
clean line, can achieve baud rate equal to bandwidth, with carrier at midpoint
of spectrum.
§
Frequency
Shift Keying (FSK)
§
Encode
by shifting between two carrier frequencies (one for 1, one for 0)
§
Used
in early modems
§
Not
susceptible to interference
§
Can
generate by: applying unipolar digital signal to one carrier, inverse digital
to other, then summing the result
§
High
bandwidth requirements: baud rate plus difference between two carriers (e.g.,
two ASK spectra with small gap between)
§
Space
between two carriers must be: 2 * (half an ASK bandwidth)+gap = full ASK
bandwidth+gap = baud rate + gap.
§
Example: for 600 bps, half-duplex
(signal gets full bandwidth).
Carriers must be > 600 Hz apart -- assume 1000 Hz apart.
Lower half of lower carrier bandwidth is 300 Hz
Upper half of upper carrier bandwidth is 300 Hz.
Total is 300+1000+300 = 1600 Hz.
§
Phase
Shift Keying (PSK)
§
Encode
by modulating phase of carrier
(amplitude and freq not affected)
§
Not
susceptible to interference, low bandwidth requirements
§
For
2-PSK, define two phases, e.g. 0 and 180 degrees: one bit / baud
§
For
4-PSK, define four phases, e.g. 0,90,180,270 degrees: two bits / baud
§
For
8-PSK, define eight phases, e.g. 0,45,90,. . . degrees: three bits / baud
§
Often
shown as dots on a constellation diagram
(cartesean coordinate system, phase = angle of line from origin to dot,
amplitude = length of that line)
§
Quadrature
Amplitude Modulation (QAM)
§
Combination
of ASK and PSK
§
Value
determined by combined phase angle and amplitude
§
Always
more phase angles than amplitudes
(less susceptible to noice, easier to distinguish)
§
8-QAM:
2 amplitudes, 4 phases for each amplitude (3 bits/baud)
§
Possible
16-QAM: 2 amplitudes, 8 phases for each amplitude (4 bits/baud)
§
Possible
16-QAM: 4 amplitudes, 8 phases, some combinations not used
§
Given
2400 baud limit on phone line, what is required for 14.4Kbps, 28.8?
Modems
Modulator / Demodulator
See
notes on Digital-Analog Encoding (ASK, FSK, PSK, QAM)
Hayes
and Hayes-compatible
-
Refers
to modem with command interpreter
-
AT
(attention) commands: AT command [parm] command [parm] . . .
-
Used
for dialing and for other functions
Bell
Modem Standards
-
for
POTS (Plain Old Telephone Service)
-
operate
on 3000 Hz telephone channel
-
notice
development trends
-
103
: FSK, 300 baud/bps (full duplex)
-
202
: FSK, 1200 baud/bps (half duplex)
-
212
: FSK 300 baud/bps (103-compatible) plus 4-PSK 600 baud/1200 bps
-
201
: 4-PSK 1200 baud/2400 bps
-
208
: 8-PSK 1600 baud/4800 bps
-
209
: 16-QAM 2400 baud/9600 bps (3? amplitudes, 12 phases)
ITU-T
Modem Standards
-
The
V series
-
Many
are compatible with Bell (e.g. V.21 == Bell 103)
-
Selected
standards:
-
V.22bis
: 600 baud, 16-QAM for 2400bps plus 4-PSK for 1200 bps
-
V.32
: 32-QAM 2400 baud/9600 bps (trellis:
each P and A combo represents 5 bits: 4 data bits plus one redundant bit)
-
V.33
: 128-QAM 2400 baud/14400 bps (trellis with 7 total, 6 data bits)
-
V.34
: 4096-QAM 2400 baud/28800 bps (12 bits/baud), 2-3 times higher thruput with compression
V.90: 56K modems
·
In
Feb 1998, the ITU accepted K56flex as a new standard, now officially known as
V.90 or V.pcm.
·
V.90
is communication protocols that allows modems to communicate at speeds of up to
56K.
·
Because
of current FCC limitations, the maximum speed is 53K, not 56K.
·
56K
is only possible in downstream (when you're DOWNLOADING data) transfer;
upstream (when you are UPLOADING information) is still 33.6K maximum.
·
56K
is only possible if you're calling an ISP. Two 56K modems calling each other
will NOT be able to reach 56K due to the fact that 56K technology requires a
digital connection (like an ISP) at one end.
·
Reason
for assymetry? Here's the best explanation I can gather from consulting several
sources: Upstream signal is analog from you to Central Office, where codec does
A-D conversion into PCM encoded signal. The encoding will not be perfect due to
signal degradation between your house and CO, plus error due to A-D
quantization. Downstream signal is digital from source (ISP, which transmits
digital version of audio wave) to Central Office, where codec does D-A
conversion. Nothing is lost in the D-A
conversion; the signal is precise going into the analog line and can therefore
be transmitted more rapidly for given line.
·
Ability
to achieve 56K depends on the quality of your phone line and many other factors
(mostly beyond your control).
ADSL: Beyond modems
Assymetric
Digital Subsriber Line (ADSL)
·
Allows
downloads (from ISP to you) at over 1-8 Mbps, and uploads (from you to ISP) at
typical 64-640 Kbps. That's why its called assymetric!
·
Uses
existing copper wire. But you get the entire bandwidth, not just 4KHz. This is better than cable, where you have to
share the bandwidth with others in your neighborhood.
·
Bandwidth
is divided into 3 subbands: voice, upstream, downstream. So not only can you
use the phone at the same time, but you get full-duplex digital communication
to boot.
·
Higher
frequencies are more sensitive to noise than lower frequencies, so maximum
upload/download throughput will vary widely depending on distance from home to
Central Office.
·
Works
only if home is less than 3 miles from CO. Currently 80% of U.S. phones. If
distance is greater, line contains "load coil" devices that boost the
analog signal but also limit it to 4KHz bandwidth!
·
Uses
ordinary twisted pair, but requires special equipment and service.
·
See
www.adsl.com
Some Theoretical
Limits on XMIT RATE
Will use
voice grade telephone as example: typical usable bandwidth is 3000 Hz.
Nyquist theorem
(1924) provides theoretical upper bound on bits per second (BPS), assuming
noiseless wire.
D = 2 B log2
K
where:
D = max BPS rate
B = Baud rate (# signal changes per second - equivalent
to bandwidth)
K = number of signal levels
As noted above, baud rate is limited by wire
bandwidth. In the phone example, upper bound is about 3000 Hz.
For binary (2-level) signal and 3000 Hz
Baud, D = 2 * 3000 * 1 = 6000 !!!
So how can modems achieve high throughput
(e.g. 9600, 14.4K, 28.8K BPS)? By increasing the value of K!
Nyquist applies
only to noise-free. Extended by Shannon in 1948 to include noise.
D = H log2
(1+S/N)
where:
D = max BPS rate
H = available bandwidth
S/N = signal-to-noise ratio (thermal noise) on
medium.
Given H=3000, and S/N = 1000 (typical
analog phone line), D = 30,000 !!!
Regardless of number of signal levels. What is
required S/N to double the D with fixed H?
Digital-to-Digital is used
for purely digital communication (telco, LANs)
§
NRZ-L
§
Non-Return
to Zero
§
1
is represented by positive voltage
§
0
is represented by negative voltage
§
(or
vice versa)
§
RS-232
is example
§
Synchronization
properties?
§
NRZ-I
§
Non-Return
to Zero - Inverted
§
1
is represented by voltage transition (either direction)
§
0
is represented by no transition
§
good
synch on strings of 1's, poor synch on strings of 0's (less frequent)
§
RZ
§
Return
to Zero
§
Uses
3 levels: positive, negative, and 0 volts.
§
1
is represented by positive
§
0
is represented by negative
§
in
either case, signal goes to 0 volts midway through bit interval (synch)
§
Note:
requires double the bandwidth!
§
Manchester
§
Uses
2 levels: positive and negative voltage
§
1
is represented by negative-to-positive transition in mid-bit
§
0
is represented by positive-to-negative in mid-bit
§
if
string of 1's or 0's, also have transition at beginning of bit (to get back).
§
Used
in Ethernet
§
Differential
Manchester
§
1
is represented by no transition at beginning of bit
§
0
is represented by transition at beginning of bit (either direction)
§
There
is a bit transition in mid-bit in either case (for synch)
§
Used
in Token Ring
[notes
| CSC 465 | Peter
Sanderson | Computer Science | SMSU ]