Phase-Locked Loop (PLL)

EE174 – SJSU
Tan Nguyen

1
 OBJECTIVES

• Introduction to Phase-locked loop (PLL)

• Historical Background
• Basic PLL System
• Phase Detector (PD)
• Voltage Controlled Oscillator (VCO)
• Loop Filter (LF)
• PLL Applications

2
 Introduction to Phase-locked Loop (PLL)
• A Phase-Locked Loop (PLL) is a negative feedback system consists of a phase
detector, a low pass filter and a voltage controlled oscillator (VCO) within its loop.
Its purpose is to synchronize an output signal with a reference or input signal in
frequency as well as in phase.
• In the synchronized or “locked” state, the phase error between the oscillator’s
output signal and the reference signal is zero, or it remains constant.
• If a phase error builds up, a control mechanism acts on the oscillator to reduce
the phase error to a minimum so that the phase of the output signal is actually
locked to the phase of the reference signal. This is why it is called a PLL.
• The majority of PLL applications fall into four main categories:
• Frequency synthesis (Most widely used so PLL is also referred as frequency
synthesizer).
• Frequency (FM) and phase (PM) modulation and demodulation.
• Data and carrier recovery.
• Tracking filters.
• Classification of PLLs: Analog or Linear PLL (LPLL), Digital PLL (DPLL) is Analog
PLL with digital phase detector, All-Digital PLL (ADPLL) is a digital loop in two
senses: all digital components and all digital (discrete-time) signals.
3
How Are PLLs Used?

4
 Brief Phase-Locked Loop (PLL) History

1932: Invention of “coherent communication” using vacuum tube, (deBellescize)

1943: Horizontal and vertical sweep synchronization in television (Wendt and
Faraday)
1954: Color television (Richman)
1965: PLL on integrated circuit
1970: Classical digital PLL
1972: All-digital PLL
PLLs today: in every cell phone, TV, radio, pager, computer, …
Clock and Data Recovery
Frequency Synthesis
Clock Generation
Clock-skew minimization
Duty-cycle enhancement

1.people.ee.duke.edu/~mbrooke/defense/Borte.ppt 5
 Basic PLL System

The basic PLL block diagram consists of three components connected in a feedback loop :
• A Phase Detector (PD) or Phase Frequency Detector (PFD)
• produces a signal V𝝓 proportional to the phase difference between the fin and fosc signal.
• A Loop Filter (LF)
• filters output voltage Vout that controls the frequency of the VCO.
• A Voltage-Controlled Oscillator (VCO)
• Vout at the input of the VCO determines the frequency fosc of the periodic signal Vosc at the output of the VCO

A basic property of the PLL atemps to maintain the frequency lock fosc= fin between Vosc and Vin even if the frequency
fin of the incoming signal varies in time.
Assume the PLL is in the locked condition, and the frequency fin of the incoming signal increases slightly. The phase
difference between the VCO signal and the incoming signal will begin to increase in time. As a result, the filtered
output voltage Vout increases  the VCO output frequency fosc increases until it matches fin, thus keeping the PLL in
the locked condition. 6
 Stages of PLL Operations
There are three stages of PLL operations:
• Free Running Stage: When no input is applied at the
phase detector, PLL out put frequency is fosc = fo
where fo free running frequency of the VCO.
• Capture Stage: When an input is applied at the phase
detector and due to feedback mechanism PLL tries to
track the output with respect to the input.
• Phase Locked Stage: Due to feedback mechanism,
the frequency comparison stops when fosc = fin.

Lock Range and Capture Range of PLL:

Lock Range of the PLL: The range of frequencies where the locked PLL remains in the locked: fmin ≤ fin ≤ fmax
The lock range is wider than the capture range.

• If the PLL is initially locked, and if fmax < fin < fmin  the PLL becomes unlocked (fin ≠ fosc).
• When the PLL is unlocked, the fosc= fo where fo is called the center frequency, or the free-running frequency of the VCO.

Capture Range of the PLL: The lock can be established again if the incoming signal frequency fin gets close enough to fo.
The range of frequencies such that the initially unlocked PLL becomes locked: fo- fc ≤ fin ≤ fo+ fc
Sometimes a frequency detector is added to the phase detector to assist in initial acquisition of lock. 7
 Linear (Analog) Phase Detector
cos(ωint + φin)
An analog multiplier mixer can be used as a phase detector which compares
the phase at each input and generates an error signal, Vφ(t), proportional to
the phase difference between the two inputs. cos(ωosct + φosc)
Recall that the mixer takes the product of two inputs. Vφ(t) = ½{cos(φ0sc – φin)
Vφ(t) = cos(ωosct + φosc) cos(ωint + φin) + cos[(2ωosct)+ (φosc+φin)]}
= (1/2) {cos[(ωosct – ωint )+ (φosc – φin)] + cos[(ωosct + ωint )+ (φosc + φin)]} High frequency
component to be removed
When loop is locked (ωosc = ωin)  we have an output proportional to the by low-pass filter
cosine of the phase difference and one output at twice the input frequency .
Vφ(t) = (1/2) {cos(φosc – φin) + cos[(2ωosct )+ (φosc + φin)]}
The doubled frequency component will be removed by the low-pass loop
filter. Any phase difference then shows up as the control voltage to the VCO, +VDD/2
a DC or slowly varying AC signal after filtering. KD is the gain of the phase
detector (V/rad).
Vφ(t) = KD (φosc – φin) where KD π/2 = VDD/2  KD = VDD/π
The averaged transfer characteristic of such a phase detector is shown below.
Note that in many implementations, the characteristic may be shifted up in
voltage (single supply/single ended).
–VDD/2
8
 Digital Phase Detector
A simple digital phase detector is an XOR gate with logic low
output (Vφ = 0V) and the logic high output (Vφ = VDD).
An example below shows the PLL is in the locked condition where
Vin and Vosc are two phase-shifted periodic square-wave signals at
𝟏
the same frequency fosc = fin = T , and with 50% duty ratios. The
in
output of the phase detector is a periodic square-wave signal
Vφ(t) at the frequency 2fin , and with the duty ratio Dφ that
depends on the phase difference φ(t) = [φosc(t) - φin(t)] between
φ
Vin and Vosc  Dφ = (for XOR)
𝝅

The dc component Vφ of the phase detector output can be found

T
easily as the average of Vφ(t) over a period 𝟐in
VDD
 Vφ = φ = KD φ KD is called PD gain (for XOR)
𝝅
VDD
where KD = volt/rad for 0 ≤ φ ≤ π
𝝅
V
The average output rise to Vout = πDD ΔΦ = 0 → VDD when ΔΦ
goes from 0 → π. For ΔΦ > π , the average output begins to drop.
9
 Loop Filter
The output Vφ(t) of the phase detector is filtered by the low-pass loop filter. The purpose of the low-pass filter is to
pass the dc and low-frequency portions of Vφ(t) and to attenuate high-frequency ac components at frequencies 2πfin.
The simple RC filter has the transfer function:
𝟏 𝟏
F(s) = =
𝟏+𝒔 𝑹 𝑪 𝟏 + 𝒔/ω𝒑
𝟏 ω
where ωp = and fp = 𝑝 is the cut-off frequency of the filter.
𝑹𝑪 2𝛑
If fp << 2fin  the output of the filter Vout is approximately equal to the
dc component V𝝓 of the phase detector output.
In practice, the high-frequency components are not completely eliminated
and can be observed as high-frequency ac ripple around the dc or slowly-varying Vo.

In general, the filter output Vout as a function of the phase

difference. Note that Vout = 0 if Vin and Vosc are in phase (𝝓 = 0),
and that it reaches the maximum value Vout = VDD when the two
signals are exactly out of phase (𝝓 = π).
For 0 ≤ 𝝓 ≤ π, Vo increases, and for 𝝓 > π, Vo decreases. The
characteristic of periodic in 𝝓 with period 2π.
The range 0 ≤ 𝝓 ≤ π is the range where the PLL can operate in
the locked condition.
10
 Voltage Controlled Oscillator (VCO)
In PLL applications, the VCO is treated as a linear, time-invariant system. To obtain an arbitrary output frequency (within
the VCO tuning range), a finite Vout is required. Let’s define φosc– φin = φo.
The XOR function produces an output pulse whenever there is a phase misalignment. Suppose that an output frequency
ω1 is needed. From the upper right figure, we see that a control voltage V1 will be necessary to produce this output
frequency. The phase detector can produce this V1 only by maintaining a phase offset φ at its input. In order to minimize
𝑽 ω − ω𝟎
the required phase offset or error, the PLL loop gain, KD KO, should be maximized, since φ = 𝟏 = 𝟏
𝑲𝑫 𝑲𝑫 𝑲𝑶

Thus, a high loop gain KDKO is beneficial for

reducing phase errors.
Note:
From Phase detector:
𝑽
V1 = KD φ  φ = 𝟏
𝑲𝑫
From VCO:
ω − ω𝟎 𝑽 ω − ω𝟎
V1 = 𝟏  φ = 𝑲𝟏 = 𝟏
𝑲𝑶 𝑫 𝑲 𝑫𝑲 𝑶

11
 PLL 4046 Design Example
The filter output Vo controls the VCO, i.e., determines the frequency fosc of the VCO output Vosc . From PLL 4046 circuit
below, the voltage Vo controls the charging and discharging currents through capacitor C1. As a result the frequency fosc
of the VCO is determined by the Vo. The VCO output Vosc is a square wave with 50% duty ratio and frequency fosc.

The VCO characteristics are adjustable by three components:

R1, R2 and C1.
When Vo = 0, the VCO operates at the minimum frequency
fmin given approximately by:
𝟏
fmin =
𝑹𝟐(𝑪𝟏+𝟑𝟐 𝒑𝑭)
When Vo = VDD, the VCO operates at the maximum frequency
fmax given approximately by:
𝟏
fmax = fmin+ 𝑹𝟏(𝑪𝟏+𝟑𝟐 𝒑𝑭)
For fmin ≤ fosc ≤ fmax, the VCO output frequency fosc is ideally a
linear function of the control voltage Vo.
Δf
The slope Ko = osc of the fosc(Vo) characteristic is called the
Δ𝑽𝒐
For proper operation of the VCO, components C1, R1 and R2 gain or the frequency sensitivity of the VCO, in Hz/V.
should satisfy: 100pF ≤ C1 ≤ 100nF and 10kΩ ≤ R1, R2 ≤ 1MΩ 12
 PLL Design Example
Given PLL 4046 circuit on previous page.
1) Select C1, R1 and R2 so that the VCO operates from fmin = 8 kHz to fmax = 12 kHz.
2) Find the frequency sensitivity (gain) KO if the VCO characteristic fosc (Vo) is linear for 0 ≤ Vo ≤ VDD where VDD = 15V.
3) Select Cf and Rf so that the cut off frequency of the low-pass filter fp = 1KHz
4) Assume Vin(t) is square-wave signal at frequency fin. Determine voltage Vo for:
a) fin = 9 kHz
b) fin = 10 kHz
c) fin = 11 kHz

Solution:
1) Select C1 = 1 nF
 R2 = 1 / (1, 032 nF x 8000) ≈ 121 kΩ
 R1 = 1 / (1, 032 nF x 4000) ≈ 242 kΩ
2) KO = (12 – 8)kHz / (15 – 0)V = 267 Hz/V or 1676 rad/V
3) Select Cf = 10 nF  Rf = 1 / (1kHz x 2π x 10nF) ≈ 16 kΩ
4) For: 𝟏
fmin = 𝑹𝟐(𝑪𝟏+𝟑𝟐 𝒑𝑭)
a) fin = 9 kHz  ΔVo = Δfosc / KO = (12 – 9) / 267 = 11.2 V  VO ≈ 3.8 V
b) fin = 10 kHz  ΔVo = Δfosc / KO = (12 – 10) / 267 = 7.5 V  VO = 7.5 V 𝟏
fmax = fmin+ 𝑹𝟏(𝑪𝟏+𝟑𝟐 𝒑𝑭)
c) fin = 9 kHz  ΔVo = Δfosc / KO = (12 – 11) / 267 3.7 V  VO ≈ 11.3 V
13
 Examples
Problem 1.
Determine the change in frequency for a voltage controlled oscillator (VCO) with a transfer function of KO = 2.5KHz/V
and a DC input voltage change of ΔVO = 0.8V.
Solution: Δf = ΔVO KO  Δf = (0.8 V)(2.5 kHz/V) = 2 kHz
Problem 2.
Calculate the voltage at the output of a phase comparator with a transfer function of KD = 0.5V/rad and a phase error
of Vϴ = 0.75 rads.
Solution: VD = KD Vϴ = (0.5 V/rad)(0.75 rad) = 0.375 V
Problem 3.
Determine the hold in range, (i.e. the maximum change in frequency) for a phase lock loop with an open loop gain of
KV = 20kHz/rad.
Solution: Δfmax = KV π/2 = (20 krad) π/2 rad = 31.4 kHz
Problem 4.
Find the phase error necessary to produce a VCO frequency shift of Δf = 10KHz for an open loop gain of
KV = 40KHz/rad.
Solution: Vϴ = Δf / KV = 10 kHz / 40 kHz/rad) = 0.25 rad
Problem 5:
Given fosc = 1.2 MHz at VCOin = 4.5 V and fosc = 380 kHz at VCOin = 1.6 V. Find Ko
Solution: Ko = 2π x (1.2 MHz – 380KHz) / (4.5V – 1.6V) rad/V = 1,777 krad/s/v

14
 PLL Overall Transfer function
Consider a PLL with feedback = 1.
𝐾𝐷 𝐾𝑂 𝐹 𝑠 𝐾𝑉 𝐹 𝑠 1 𝟏
Forward path gain (open loop gain): G(s) = = where KV = KDKO, F(s) = =
𝑠 𝑠 1+𝑠 𝑅𝐶 𝟏 + 𝒔/ω𝒑
Transfer function (close loop gain):

Φ𝑜𝑢𝑡 𝐺 𝑠 𝐾𝑉 𝐹 𝑠 /𝑠 𝐾𝑉 𝐹 𝑠
H(s) = = = =
Φ𝑖𝑛 1+𝐺(𝑠) 1+𝐾𝑉 𝐹 𝑠 /𝑠 𝑠+𝐾𝑉 𝐹 𝑠
𝟏
𝐾𝑉 𝐾𝑉 𝐾𝑉ω𝒑
𝟏 + 𝒔/ω𝒑
= 𝟏 = 𝒔
𝟐 =
𝑠+𝐾𝑉 𝟏 + 𝒔/ω + 𝑠 +𝐾𝑉 𝑠2+ 𝑠ω𝒑+𝐾𝑉ω𝒑
ω𝒑
𝒑
Standard form:
𝐾𝑉ω𝒑 ω𝒏𝟐
H(s) = =
𝑠2+ 𝑠ω𝒑+𝐾𝑉ω𝒑 𝑠2+𝟐ζω𝒏𝑠+ω𝒏𝟐
Therefore:
Natural frequency: ω𝒏 = 𝐾𝑉ω𝒑 = 𝐾𝐷𝐾𝑂ω𝒑

ω𝒑 ω𝒑 1 ω𝒑 1 ω𝒑
Damping factor: ζ = = = =
𝟐ω𝒏 𝟐 𝐾𝑉ω𝒑 2 𝐾𝑉 2 𝐾 𝐷𝐾 𝑂
15
 PLL Examples
A phase-locked loop has a center frequency of ω0 = 105 rad⁄s, KO = 103 rad/s per V, and KD = 1 V/rad. There is no
other gain in the loop. Determine the overall transfer function H(s) for:
a) The loop filter F(s) = 1 (all pass filter).
b) The loop filter F(s) is shown below.
c) Loop filter F(s) as in part b) , XOR for the phase detector and VDD = 5V.
d) Natural frequency ω𝒏 and damping factor ζ for part c)
Solution:
a) The loop gain KV = KD KO = (103 rad/s-V)(1 V/rad) = 103 s-1.
𝐾𝑉 𝐹 𝑠 103
The transfer function for F(s) = 1 : H(s) = =
𝑠+𝐾𝑉 𝐹 𝑠 𝑠+103
2525
b) The transfer function for F(s) = 𝑠+2525: ωp = 1/RC
3 2525 = 1/(120kΩ x3.3nF) = 2525rad/s
𝐾𝑉 𝐹 𝑠 10 2525x10 3
𝟏 1𝑠 2525
𝑠+2525 
2525 = 2
H(s) = = F(s) = =
𝟏 + 𝒔/ω𝒑 = 1+𝑠/2525 𝑠+2525
𝑠+𝐾𝑉 𝐹 𝑠 𝑠+103 𝑠 +2525𝑠+2525x103
𝑠+2525
c) KD = VDD/π = 1.6 V/rad  KV = KD KO = (103 rad/s-V)(1.6 V/rad) =1600 s-1
2525
𝐾𝑉 𝐹 𝑠 1600 𝑠+2525 4.044x106
H(s) = = 2525 =
𝑠+𝐾𝑉 𝐹 𝑠 𝑠+1600 𝑠2+2525𝑠+4.04x106
𝑠+2525
ω𝒑 𝟐𝟓𝟐𝟓
d) ω𝒏 = 𝐾𝑉ω𝒑 = 4.04x106 = 2010 Hz and ζ= = = 0.63
𝟐ω𝒏 𝟐(𝟐𝟎𝟏𝟎)
16
 Synthesize PLL
We will now add the divider 1/N to the feedback path. This architecture is called an “integer-N” synthesizer.
𝐾 𝐾𝑂 𝐹 𝑠 𝐾 𝐹 𝑠
Forward path gain (Loop gain is reduced by a factor of N): G(s) = 𝐷 = 𝑉
𝑁𝑠 𝑁𝑠
(Note: In most applications, N is not constant, so KV = KDKO is not a constant – varies with frequency according to
the choice of N).
𝟏
Φ𝑜𝑢𝑡 𝐺 𝑠 𝐾𝑉 𝐹 𝑠 /𝑁𝑠 𝐾𝑉 𝐹 𝑠 𝐾𝑉 𝐾𝑉ω𝒑
𝟏 + 𝒔/ω𝒑
Transfer Function: H(s) = = = = = 𝟏 =
Φ𝑖𝑛 1+𝐺(𝑠) 1+𝐾𝑉 𝐹 𝑠 /𝑁𝑠 𝑁𝑠+𝐾𝑉 𝐹 𝑠 𝑁𝑠+𝐾𝑉 𝑁𝑠2+𝑁𝑠ω𝒑+𝐾𝑉ω𝒑
𝟏 + 𝒔/ω𝒑
𝐾𝑉ω𝒑/𝑵 ω𝒏𝟐
Standard form: H(s) = =
𝑠2+ 𝑠ω𝒑+𝐾𝑉ω𝒑/𝑵 𝑠2+𝟐ζω𝒏𝑠+ω𝒏𝟐
Therefore:
𝐾𝑉ω𝒑 𝐾𝑉 ω𝒑 ω𝒑 1 𝑵ω𝒑 1 𝑵
Natural frequency: ω𝒏 = = Damping factor: ζ = = = =
𝑁 𝑅𝐶𝑁 𝟐ω𝒏 𝐾 𝑉 ω𝒑 2 𝐾𝑉 2 𝐾𝑉𝑅𝐶
𝟐
𝑁

17
Phase Lock Loop
Applications
EE174 – SJSU
Tan Nguyen
 Common PLL Applications
• Clock multiplier/Clock Generator
• Input: Fixed frequency clock
• Output: Multiple of input clock frequency/Multiple of clock outputs

• Frequency synthesizer (Fractional-N, Integer-N)

• Input: Fixed frequency clock
• Output: Clock signal with arbitrary frequency

• Clock and data recovery

• Input: Data signal (from a serial link)
• Output: Digital data as well as clock signal with phase detector is different than other
applications

• FM demodulation
• Input: Radio signal
• Output: Demodulated signal
Basic PLL
Basic Clock Multiplier
 Integer-N Synthesizer
• The resolution of the output frequency is determined by the reference frequency applied to the phase detector. Step
size or frequency resolution - the smallest frequency increment possible.
• To obtain a stable low frequency source is not easy, because a quartz crystal oscillating in kHz region is quite bulky
and not practical. A sensible approach is to take a good stable crystal-based high frequency source and an integer-N
synthesizer to divide it down.

Example: Given an oscillator EXTAL = 10 MHz, a step size/frequency resolution of 200 KHz is required. Determine counter
values of R and N to produce outputs fOUT = 900.2, 900.4, … MHz.

Solution: Step size/frequency resolution: fREF = 200 KHz = 10 MHz / R  R = 10 MHz / 200 kHz = 50
FOUT = 900.2 MHz = 200 kHz x N  N = 900.2 MHz / 200 kHz = 4501  N = 4501, 4502, …
 Fractional-N Synthesizer
Fractional-N allows the resolution at the PLL output to be reduced to small fractions of the PFD frequency as shown
below, where the PFD input frequency is 1 MHz. It is possible to generate output frequencies with resolutions
of 100s of Hz, while maintaining a high PFD frequency. As a result the N-value is significantly less than for integer-N.

𝑨(𝑵)+𝑩(𝑵+𝟏)
NEFF =
𝑨+𝑩

A: Cycle for N counter

B: Cycle for N+1 counter
A + B = 10
Toggling between the two integer
division ratios, a fractional division
ratio can be achieved by time-
averaging the divider output.

𝟖(𝟗𝟎𝟎)+𝟐(𝟗𝟎𝟏)
NEFF = = 900.2
𝟏𝟎
 Programmable Phase-Locked Loop Clock Generator
The FS7140/FS7145 is a monolithic CMOS clock generator/regenerator IC. Via the I2C−bus interface, the FS7140/45 can
be adapted to many clock generation requirements. The length of the reference and feedback dividers, their fine
granularity and the flexibility of the post divider make the FS7140/45 the most flexible stand alone PLL clock generator
available.
 Integer-N Frequency Synthesizers with Prescalers
Four-modulus prescalers
To extend the upper frequency range of a frequency synthesizer but still allows the synthesis of lower
frequencies. The solution is the four-modulus prescaler. The four-modulus prescaler is a logical extension
of the dual-modulus prescaler. It offers four different scaling factors, and two control signals are required to
select one of the four available scaling factors.

fosc = 10MHz
 Integer-N Frequency Synthesizers with Prescalers cont.
The four-modulus prescaler can divide by factors of N2: N3: Divide by
100, 101, 110, and 111. The internal logic of the four- Add 10 Add 1
modulus prescaler is designed so that the scaling 0 0 100
factor is determined by control signals “N2: Add 10” 0 1 101 N1 N2 N3 Div
and “N3: Add 1” as shown in next table.
1 0 110 10 2 3 111
1 1 111 9 1 2 111
There are three programmable /N counters in the system: /N1, /N2, and /N3 dividers. 8 0 1 101
The overall division ratio is given by: NEFF = 100N1 + 10N2 + N3 7 0 0 100
where N3 represents the units, N2 the tens, and N1 the hundreds of the division ratio Ntot. 6 100
N2 and N3 range: 0 – 9, and N1 ≥ N2 and N3 because when the content of N1 becomes 0, all
/N1, /N2 and /N3 counters are reloaded to their preset values, and the cycle is repeated. 5 100
Output /N2 and /N3 counters are HIGH when counter value ≠ 0 and go LOW when counter 4 100
value = 0. Note: For a reference frequency f1 = 10 kHz, the lowest frequency to be 3 100
synthesized is therefore: 100 x f1 = 1 MHz.
Example: We wish to generate a frequency that is 1023 times the reference frequency. The 2 100
division ratio Ntot is thus 1023; hence N1 = 10, N2 = 2, and N3 = 3 are chosen. 1 100
Both outputs of the /N2 and /N3 counters are now HIGH, a condition that causes the four- 0 Reload values for
modulus prescaler to divide initially by 111. N1, N2, N3
NEFF = 2(111) + 1(101) + 7(100) = 1023
10 2 3 111
 Frequency Synthesis and Frequency Dividers
A frequency synthesis technique and frequency dividers are used to generate multiple frequencies
from an accurate reference frequency, usually a crystal oscillator.
In this manner, the non integer frequencies can be developed.

Example: A system requires CPU clock 1.6 GHz, memory clock = 200MHz and I/O bus clock = 400 MHz.
A crystal oscillator of 30 MHz is used for fREF. Determine counters /R, /N, /P and /Q.
Solution: Choose R = 3  fPFD = 30 MHz / 3 = 10 MHz, N = 160  fOSC = 160 x 10 MHz (CPU),
P = 4  fOSC = 1.6 GHz / 4 = 400 MHz (I/O Clock), Q = 8  fOSC = 1.6 GHz / 8 = 200 MHz (Memory).

10 MHz /4
fOSC

30 MHz /3

/8
/160
 Clock Data Recovery
Different Techniques of Data Communication
2. Parallel Data Communication:
1. Serial Data Communication: Data bits are driven on multiple wires simultaneously.
Data bits are transmitted sequentially
one by one a) Skew
Travelling path length
for every bit is going to
be different. Due to this
some bits can arrive
early or before than
others which may
corrupt the information.

b) Inter symbol interference (ISI) and Cross talk

Due to several parallel links ISI and Cross talk is introduced in the system
which gets more severe as length of link is increased. So this limits the
length of a connection.
c) Limitation of I/O pin count
Parallel data communication requires a lot more I/O pins than what is
required by serial data communication.
 Clock Data Recovery
A Clock and Data Recovery (CDR) circuit is an essential block in many high-speed wire-linked data transmission
applications such as optical communications systems, backplane data-link routing and chip-to-chip interconnection.
Sometimes the data is sent over the high speed serial interfaces without an accompanying clock, the receiver needs to
recover the clock in order to sample the data on serial lines.
The important role of a Clock and Data Recovery (CDR) is to:
• Detect the transitions in the received data and generates a periodic recovery clock – Clock recovery.
• Retime the received data which samples noisy data and then regenerates it with less jitter and skew driven by the
recovered clock – Data recovery.
Note: A primary difference between CDR and PLL is that the incoming data is not periodic like the incoming reference
clock of a PLL
 Clock Data Recovery
• Phase Detector: Generates an output signal in relation to the phase difference of both inputs
• Linear – PLL can analyzed in a similar manner as frequency synthesizers
• Nonlinear – PLL operates as a bang-bang control system (hard to rigorously analyze in many cases)

• Charge-Pump: Output pulses of PD are converted to current

• Loop Filter: Integrates the output of the charge pump and produces the control voltage

• Voltage Controlled Oscillator: Generates a periodic output whose frequency depends on the control voltage

30
 Phase Detectors in Clock and Data Recovery Circuits

Hogge Phase Detector (Linear PD):

Error output, e(t), consists of two pulses with opposite polarity

• Positive polarity pulse has an area that is proportional to the phase error between the data and clock
• Negative polarity pulse has a fixed area corresponding to half of the clk period
• Overall area is zero when data edge is aligned to falling clock edge

31
 Phase Detectors in Clock and Data Recovery Circuits
Alexander Phase Detector (Bang-Bang PD):
Data is sampled at 3 equidistant points A, B and C
• XOR gates combine nodes A, B and C: X = A xor B and Y = B xor C
• Performs an early-late detection
Clock is early: Y = Low and X = High
Clock is late: Y = High and X = Low

32
 Phase Detectors in Clock and Data Recovery Circuits
Alexander Phase Detector (Bang-Bang PD):

33
 Clock Data Recovery
• Clock Jitter:
Jitter is a shift in the edges of a periodic signal. This
breaks the periodicity of the signal.
– AC - jitter: The uncertainty of the output phase
– DC - phase offset: Undesired difference of the
average output phase relative to the input phase.
Clock w/o jitter Clock w/ jitter
A data bit should be sampled at the
centre. It is the optimum position
where maximum shift in the edges
on either side (from left to right or
right to left) can be encountered.
However if the shift in an edge
becomes greater than half of the bit
period then there will be a bit error.
• Data Jitter: Jitter tolerance is defined as the amount of jitter that
the CDR circuit must tolerate on the input data without increasing
the bit error rate (BER).
If the jitter on the input data varies slowly, the recovered clock will
track the transition in the data and always sample the data in the
middle of the bit period as shown below. This will guarantee a low
BER.
If the jitter on the input data varies fast, the recovered clock will not
be able to track the transition in the data and will fail to sample the
data in the middle of the bit period as shown in Figure 2.14. This will
result in a greater BER.
Clock Data Recovery
 Differentiation CDR
The steps taken by the algorithm to obtain the recovered data. The first plot is the input data, the
second is the differentiated input data. We can see that the peaks occur at the zero crossings of the
input data. The third plot is the fullwave rectified differentiated data. This data is used to create a
clock, which is then used to create the fourth plot, the regenerated data
References:
http://www.onsemi.com/pub/Collateral/FS7140-D.PDF
http://www.scribd.com/doc/237983665/PLL

http://www.delroy.com/PLL_dir/tutorial/PLL_tutorial_slides.pdf

Phase Locked Loops 6/e, 6th Edition by Roland Best

https://www.google.com/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-
8#q=an535

http://eprints.lancs.ac.uk/52334/1/PLLbook_chapter_final_2.pdf

http://www.ti.com/lit/ds/symlink/lm565.pdf

PLL-74HC4046_Application_Note%20(1).pdf

http://users.ece.gatech.edu/pallen/Academic/ECE_6440/Summer_2003/L170-
FreqSyn-I(2UP).pdf

http://iris.lib.neu.edu/cgi/viewcontent.cgi?article=1007&context=elec_comp_theses
References:
ecee.colorado.edu/~ecen4002/manuals/dsp56300family/ch6-pll-clk.pdf
http://www.arrowdevices.com/blog/beginners-guide-to-clock-data-recovery/
http://www.twyman.org.uk/clock_recovery/#pn-seq
http://web.stanford.edu/class/archive/ee/ee371/ee371.1066/lectures/Old/Older/lect_17_CDR_2up.pdf

http://www.seas.ucla.edu/brweb/teaching/215C_W2013/PLLs.pdf
http://www.ece.ucsb.edu/~long/ece594a/PLL_intro_594a_s05.pdf
http://www.ti.com/lit/an/snoa351/snoa351.pdf

http://memo.cgu.edu.tw/jtkuo/files/eelab%202014%28III%29/1230_Lab12_Expxx_PhaseLockedLoop.pdf

http://siihr64.iihr.uiowa.edu/MyWeb/Teaching/ece_55141_2013/Homework/HomeworkAssignment08Solution.pd
f
http://www.freeclassnotesonline.com/VCO-and-PLL-Calculations-HW.php
http://ecee.colorado.edu/~ecen4618/lab4.pdf

http://www.analog.com/media/en/training-seminars/tutorials/MT-086.pdf
www.ti.com/lit/an/swra029/swra029.pdf
References:
http://www.scribd.com/doc/237983665/PLL

http://www.delroy.com/PLL_dir/tutorial/PLL_tutorial_slides.pdf

Phase Locked Loops 6/e, 6th Edition by Roland Best

https://www.google.com/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-
8#q=an535

http://eprints.lancs.ac.uk/52334/1/PLLbook_chapter_final_2.pdf

http://www.ti.com/lit/ds/symlink/lm565.pdf

PLL-74HC4046_Application_Note%20(1).pdf

http://users.ece.gatech.edu/pallen/Academic/ECE_6440/Summer_2003/L170-
FreqSyn-I(2UP).pdf

http://iris.lib.neu.edu/cgi/viewcontent.cgi?article=1007&context=elec_comp_theses

39
References:
http://scholarworks.sjsu.edu/cgi/viewcontent.cgi?article=8032&context=etd_theses
http://www.scribd.com/doc/237983665/PLL
http://www.seas.ucla.edu/brweb/teaching/215C_W2013/PLLs.pdf
http://www.ece.ucsb.edu/~long/ece594a/PLL_intro_594a_s05.pdf
http://www.ti.com/lit/an/snoa351/snoa351.pdf
http://memo.cgu.edu.tw/jtkuo/files/eelab%202014%28III%29/1230_Lab12_Expxx_PhaseL
ockedLoop.pdf
http://siihr64.iihr.uiowa.edu/MyWeb/Teaching/ece_55141_2013/Homework/HomeworkAs
signment08Solution.pdf
http://www.freeclassnotesonline.com/VCO-and-PLL-Calculations-HW.php
http://ecee.colorado.edu/~ecen4618/lab4.pdf
http://www.eas.uccs.edu/~mwickert/ece5675/lecture_notes/N5675_1.pdf
http://hft.uni-duisburg-essen.de/arbeiten/Vortrag_Forcan_Milan.pdf
https://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-976-high-
speed-communication-circuits-and-systems-spring-2003/lecture-notes/lec21.pdf
https://pdfs.semanticscholar.org/de17/8a088507a387e98a1dc06ccf628e3015fcd3.pdf

40