Upload
eliasationgkiao
View
212
Download
0
Embed Size (px)
DESCRIPTION
.
Citation preview
All Digital CMOS Self-sample Vernier Delay Line
Circuit for Clock Jitter Measurement
(95 7) ( 1)/()
( 9 ) () ( ) () ( ) () ( )
: 93521006
:
`
95 7 17
1. 15 3 http://thesis.lib.ncu.edu.tw/paper.htm
2.
3.
4.
i
(PLL)(jitter)
one-period-delay
circuit jitter-measurement circuit
One-period-delay circuit Jitter
Measurement circuit
TSMC 0.35um 2P4Mone-period-delay circuit
300ps jitter-measurement circuit 15ps
100MHz~400MHz
ii
Abstract As the improvement of semiconductor technology, VLSI circuits has developed
into System-On-a-ChipSoC. When many systems integrated into a chip, the
sequence of clock of every circuit must be accurate. In the system, clock skew
will affect the performance of the system. The Phase-Locked Loop (PLL) is
recognized as one of the important components for clock recovery.
In PLL circuits, the value of output clock jitter effect the performance of PLL. In
the past, the jitter is measured by the external equipment. But, with the
increased operating frequency, it will have a high cost on jitter measuring by
external equipments. Sometimes probes of external equipments will be
induced noise. The measurement result will be different. Because this reason,
the built-in clock jitter measurement circuits are proposed. Using built-in clock
jitter measurement circuits to measure the clock jitter that can reduce testing
cost, decrease the effect of noise, and speeding up the jitter measurement.
In the past, the vernier delay line circuit has high circuit resolution but high chip
area. In this all digital CMOS self-sample vernier delay line circuit has two
stages. One is one-period-delay circuit, the other is jitter measurement circuit.
In this circuit, the one-period-delay circuit is used to delay the clock to
measurement point quickly and the jitter measurement circuit created high
circuit resolution to measure the jitter. In this way, the circuit can measure the
clock jitter quickly and accurately and reduce the chip area.
The proposed circuit is designed in TSMC 0.35um 2P4M CMOS process. The
resolution of one-period-delay circuit is 300ps and that of jitter measurement
circuit is 15ps. The measured frequency of the proposed circuit is 100MHz to
400MHz.
iii
Contents Abstract.................................................................. ii List of Table ............................................................. v List of Figure........................................................... vi Chapter 1 Introduction.............................................1
1.1 Motivation.....................................................................................................1 1.2 Organization of Thesis ..................................................................................2
Chapter 2 Preview Work .........................................4 2.1 Jitter Definition ...........................................................................................4
2.1.1 Cycle-to-Cycle jitter...........................................................................4 2.1.2 Period Jitter ........................................................................................6 2.1.3 Long-Term Jitter I ..............................................................................7 2.1.4 Long-Term Jitter II.............................................................................8
2.2 Jitter Histogram...........................................................................................8 Chapter 3 Jitter Measurement Technology ........... 11
3.1 Off-chip Jitter Measurement ..................................................................... 11 3.2 On Chip Jitter Measurement .....................................................................12 3.3 Operation of BIJM Circuits ......................................................................14 3.4 Traditional Methodology ..........................................................................15 3.5 Delay Chain Circuit ..................................................................................16
3.5.1 Operation of Delay Chain Circuit ..................................................16 3.5.2 Jitter Measurement of Delay Chain Circuit ...................................17 3.5.3 Characteristic of Delay Chain Circuit............................................18
3.6 Delay Loop Circuit ...................................................................................18 3.6.1 Operation of Delay Loop Circuit .....................................................18 3.6.2 Jitter Measurement of Delay Chain Circuit ...................................19 3.6.3 Characteristic of Delay Chain Circuit............................................19
3.7 Vernier Delay Line Circuit........................................................................203.7.1 Operation of Vernier Delay Line Circuit..........................................20 3.7.2 Jitter Measurement of Vernier Delay Line Circuit...........................21 3.7.3 Characteristic of Vernier Delay Line Circuit ...................................22
Chapter 4 Proposed BIJM Circuit..........................23 4.1 Structure of the Proposed BIJM Circuit....................................................23 4.2 Proposed Methodology .............................................................................24 4.3 One-period-delay Circuit ..........................................................................25
4.3.1 Structure of One-period-delay Circuit .............................................25 4.3.2 Operation of One-period-delay Circuit ............................................26 4.3.3 Design of One-period-delay Circuit.................................................27
4.4 Jitter-measurement Circuit..........................................................................28 4.4.1 Structure of Jitter-measurement Circuit ...........................................28 4.4.2 Operation of Jitter-measurement Circuit..........................................29 4.4.3 Design of Jitter-measurement Circuit ..............................................29 4.4.4 Control Logic Circuit .......................................................................30
4.5 Operating Principle ...................................................................................31 Chapter 5 Chip implementation.............................33
iv
5.1 Simulation of the proposed circuit ............................................................34 5.1.1 Consideration of Simulation ............................................................35 5.1.2 Simulation of One-period-delay Circuit ..........................................35 5.1.3 Simulation of Jitter-measurement Circuit ........................................36 5.1.4 Simulation of Control Logic Circuit ................................................37
5.2 Layout of the Proposed Circuit .................................................................38 5.3 Measurement Result..................................................................................39
5.3.1 Concept of Measurement .................................................................39 5.3.2 Chip Measurement ...........................................................................40 5.3.3 Measurement Result.........................................................................41 5.3.4 Specification of Circuit ....................................................................44
Chapter 6 Conclusion and Future Work................45 6.1 Conclusion ................................................................................................45 6.2 Future Work ..............................................................................................45
Reference material ................................................46
v
List of Table Table 5.1 Specification of BIJM circuit44
Table 5.2 Comparison of BIJM circuit.44
vi
List of Figure
Figure 2.1 Cycle-to-cycle jitter..........................................................................5 Figure 2.2 Application of cycle-to-cycle jitter ....................................................5 Figure 2.3 Period jitter......................................................................................6 Figure 2.4 Application of period jitter measurement .........................................6 Figure 2.5 Long-term jitter I..............................................................................7 Figure 2.6 Long-term (tracking) jitter II.............................................................8 Figure 2.7 Jitter histogram ...............................................................................9 Figure 2.8 Jitter histogram II ............................................................................9 Figure 2.9 Jitter histogram III ...........................................................................9 Figure 2.9 Jitter histogram IV.........................................................................10 Figure 3.1 Operation of BIJM circuits.............................................................14 Figure 3.2 Traditional methodology................................................................15 Figure 3.3 Delay chain circuit.........................................................................16 Figure 3.5 Jitter measurement of delay chain circuit......................................17 Figure 3.6 Operation of delay chain circuit.....................................................19 Figure 3.7 Operation of vernier delay line circuit............................................20 Figure 3.8 Jitter measurement of vernier delay line circuit .............................21 Figure 4.2 Propose methodology ...................................................................24 Figure 4.3 Structure of one-period-delay circuit .............................................25 Figure 4.4 Operation of one-period-delay circuit ............................................26 Figure 4.6 Structure of jitter-measurement circuit ..........................................28 Figure 4.7 Operation of jitter-measurement circuit .........................................29 Figure 4.8 MOS capacitance..........................................................................30 Figure 4.9 Control logic circuit........................................................................30 Figure 4.10 Operating principle......................................................................32 Figure 5.1 Design procedure..........................................................................34 Figure 5.2 Simulation consideration...............................................................35 Figure 5.3 Simulation of circuit resolution in one-period-circuit ......................36 Figure 5.4 Simulation of circuit resolution in jitter-measurement circuit .........37Figure 5.5 Simulation of control logic circuit...................................................38 Figure 5.6 Photograph of whole chip .............................................................39 Figure 5.7 Layout block of whole chip............................................................39 Figure 5.8 concept of measurement ..............................................................40 Figure 5.9 Measurement setup of the chip.....................................................41 Figure 5.10 100MHz signal mixed with random signal...................................42 Figure 5.11 Measurement result of proposed circuit I ....................................42 Figure 5.10 100MHz signal mixed with square noise.....................................43 Figure 5.11 Measurement result of proposed circuit II ...................................43
1
Chapter 1 Introduction
1.1 Motivation
As the improvement of semiconductor technology, VLSI circuits has developed
into System-On-a-ChipSoC. When many systems integrated into a chip, the
sequence of clock of every circuit must be accurate. In the system, clock skew
will affect the performance of the system. In many practical applications, such
as, wireless phone, optical fiber links or macro-computers. The Phase-Locked
Loop (PLL) is recognized as one of the important components for clock
recovery.
In high-speed system, the PLL circuit is an important component for the clock
recovery. In theory, the PLL circuit is a good clock recovery circuit [1], however,
in the circuit implementation; there is non-ideal effect, clock jitter, in the PLL
circuit. This effect affects the timing accuracy and the signal to noise ratio in
circuits base on a PLL. Sometimes, the clock jitter may cause unacceptable
errors in high-speed systems.
In PLL circuits, the jitter affects the performance of PLL. So, the clock
measurement and testing is the most effective method to estimate the PLL
performance. With increasing frequency resulting from new submicron
technologies, the impact of clock jitter on all function has become more and
more critical. The clock measurement has become a very important job for the
design PLLs [2]. There is a strong pressure to find a low-cost test solution for
PLLs[3].Although measuring timing margins and synchronization requirements
push the limits of traditional test equipment [3], the urgent need for high-speed
clock jitter testing is still not satisfied. It makes designers find a new technology
2
that can substitute for the traditional equipments. However, traditional ways are
still not enough in measuring the clock jitter.
In traditional way, the external equipment would distort the tested clock signal
and change the clock measurement result. In order to achieve a more
convenient clock jitter measurement a Time to Digital ConverterTDC
technology is used to output an all-digital data in the propose method. A built-in
jitter test method is used to realize to the propose method. However, even if
some usable built-in clock jitter measurement BIJMmethods also have been
proposed, the requirements of the test time and circuit area still limited the
circuit application. In order to release these requirements, a new built-in clock
jitter measurement method is proposed. A continuous clock jitter measurement
method is adopted to make a real-time measurement. The proposed circuit
area and test time cab be reduce. And, with the improve circuit structure; no
more external jitter-free clock is needed to sample the clock jitter.
1.2 Organization of Thesis
In PLL circuit design, the half-digital PLL is the most popular used circuit. It has
advantages of less area requirement, simple circuit structure. However, unlike
all-digital PLL, clock jitter of the half-digital PLL circuit cannot be extracted
directly by the circuit internal signal. It needs a specific jitter test circuit to
measure the clock jitter of the half-digital PLL circuit.
In the chapter 2, the different clock jitter is defined. Normally, the common
tested clock jitter can be divided into four parts, period jitter, cycle-to-cycle jitter,
and two different long term jitter. Different jitter definitions play different in the
circuit.
In the chapter 3, the on-chip and off-chip jitter measurement techniques will be
3
introduced. All techniques have their-self benefits. Off-chip jitter measurement
techniques base on external equipments. On-chip jitter measurement
techniques will introduce delay chain circuit, delay loop circuit, vernier delay
line circuit. With these existed techniques, readers would have a brief concept
in clock jitter measurement techniques.
In the chapter 4, the proposed circuit is introduced. The propose circuit used
new technique to measure the clock jitter. This new technique can reduce
chip area. In the proposed circuit, it only needs only one input, without extra
jitter-free clock.
In the chapter 5, the simulation and layout of chip will be introduced. The
approach of measuring chip and the result of the chip measurement will be
shown.
In the final chapter, some general conclusions from this work are presented.
4
Chapter 2 Preview Work
PLLs can be used in timing applications such as clock synthesis, clock
recovery, and clock skew compensation. In these applications, the most
important set of specifications for a PLL is its jitter characteristic [4]. In fact,
jitter specifications are critical in high speed interfaces because of a limit
available timing budget. Although jitter testing is expensive test because it
requires costly equipment and long test time [5]. Jitter testing is still a very
important in analyzing the circuit performance of PLL.
2.1 Jitter Definition
Jitter can be defined as the deviations in an output transition of clock from their
ideal positions. The deviations can be leading or lagging the ideal position.
Jitter is usually specified in positive or negative pico-seconds. Jitter
measurement can be classified into three categories: cycle-to-cycle jitter,
period jitter, and long-term jitter. All jitter measurements are at a specified
voltage, usually Vcc/2 [6].
2.1.1 Cycle-to-Cycle jitter
Cycle-to-cycle jitter is the change in a clocks output transition from its
corresponding position in the previous cycle. This form of jitter is the most
difficult to measure and require a Timing Interval Analyzer (TIA). Figure 2.1
depicts the graphical representation of cycle-to-cycle jitter. J1 and J2 are the
jitter values measured for single ends signals [6].
Figure 2.1 Cycle-to-cycle jitter
The maximum of such values measured over multiple cycles in the maximum
cycle-to-cycle jitter. Large cycle-to-cycle jitter can cause system to fail.
Consider the example shown in Figure2.2 where the output of PLL1 is the
reference of PLL2. If PLL2 cannot lock to the reference frequency, the
cycle-to-cycle jitter of the output of PLL1 may have exceeded the maximum
jitter allowable for PLL2 to lock. The cycle-to-cycle jitter of PLL1 must be low
enough for PLL2 to lock [6].
Figure 2.2 Application of cycle-to-cycle jitter
5
2.1.2 Period Jitter
Period jitter is the maximum change in a clocks output transition from its ideal
position. Figure 2.3 is a graphical representation of period jitter [7]. Period jitter
measurements are used to calculate timing margins in systems.
Figure 2.3 Period jitter
For example, a macro-processor-based system in which the processor
requires 2ns of a data set-up time [6]. Assume that the clock driving the
macro-processor-based system has a maximum of 2.5ns period jitter. In this
example, the rising edge of the clock can occur before the data is valid on the
data bus. The processor will be given incorrect data and the system will not
operate correctly. The example is shown in Figure 2.4. The system designer
must consider period jitter when calculating the timing margin of the system [6].
Figure 2.4 Application of period jitter measurement
6
2.1.3 Long-Term Jitter I
Long-term jitter measures the maximum change in a clocks output transition
from its ideal over a large number of cycles. Figure 2.5 is a graphical
representation of long-term jitter. The actual number of cycles depends on the
application and the clock frequency. For PC motherboards and graphics
applications, this is usually 10~20 microseconds. For other application, this
number will be different [6].
One example of a system affect by long-term jitter is a graphics card driving a
CRT. Assume that a pixel of data is specified for the pixel at coordinates (10,
24) on the CRT. Because of long-term jitter, this data may drive at the wrong
location (11, 28). Since the effect of a jittery clock is consistent over all pixels,
the overall effect of a jittery clock is to cause an image to shift from its ideal
display position on the screen. This effect is also known as running of the
screen. The long-term jitter is the time difference of the first rising edge and the
time delayed edge [6].
Figure 2.5 Long-term jitter I
7
2.1.4 Long-Term Jitter II
Long term, or tracking, jitter, which is measure of the phase variation between
PLLs input and output clocks, as shown in Figure 2.6. Tracking jitter
represents how closely the PLL output tracks the reference clock. This jitter is
important only in inter-chip communication, where synchronization must be
maintained [7].
Output clock of PLL
Lag
Ideal
Lead
Reference Clock
Jitter
Figure 2.6 Long-term (tracking) jitter II
2.2 Jitter Histogram
The variation of the clock jitter can be census to jitter histogram. The jitter is
random in the clock. The jitter histogram can shown the distribution of jitter. In
general, the jitter histogram will be shown in the Gaussian histogram without
other noise. See the Figure 2.7 [8].
When some noise brings to the clock jitter, the jitter histogram will be change. It
will be no longer in the Gaussian histogram. The noise may be the square,
triangle, and sinusoid waves. The different jitter histogram will be shown in
Figure 2.8, Figure 2.9, and Figure 2.10.
8
Figure 2.7 Jitter histogram
Figure 2.8 Jitter histogram II
Figure 2.9 Jitter histogram III
9
Figure 2.9 Jitter histogram IV
10
11
Chapter 3 Jitter Measurement Technology
Jitter measurement technology divided into two parts. One is off-chip jitter
measurement technology; it is depended on the test equipment, it is also called
equipment based techniques [9]. The other is on-chip jitter measurement, it
included delay chain techniques, delay loop techniques and vernier delay line
techniques. All these techniques have their-self benefits.
3.1 Off-chip Jitter Measurement
In high-speed circuit design, clock jitter measurements have become a very
important job in analyzing the circuit performance. From a users perspective,
there are four traditional ways of making jitter measurements.
Spectrum analyzer: From this measurement, the jitter is modeled as phase
noise of the signal. The power of the phase noise can be measured with high
frequency range
ATE: The signal is repeatedly acquired at slightly different time settings for the
capture strobe. With the distribution of signal timing, jitter can be easily
measured.
Real time sampling oscilloscope: The clock signal is over-sample at a fixed
voltage threshold. Special software can extract the jitter from the sample data.
12
Dedicated jitter instrumentation: The signal jitter is directly measured from
the dedicated jitter instrumentation. The measurement result is report as a
digital number. There are two main approaches in this way: one is based on
Time-Interval-analyzer (TIA) ; the other is based on counter-timer.
In these traditional ways, all jitter measurements rely on the external
equipment. This will cause some serious problem. The most obvious problem
is the loading problem. In making jitter measurements, the output signals of the
circuits have to connect many probes. These probes loading makes the
measured signal decrease or distort seriously. Therefore, the reported result
will be not the original data, but the fake data. Moreover, these traditional ways
also have some other problem, resolution, accuracy, test time, testable clock
speed, and so on. In order to reduce these problems effect, the built-in clock
jitter measurement method has been report [10]-[14]. In the clock jitter
measurement, the built-in clock jitter measurement method uses the Time to
Digital Converter (TDC) technique to deal with the analog clock signal. The
TDC technique is like the ADC technique. It can translate the analog signal into
all-digital clock signal. The translated digital signal makes the clock jitter
measurement become easier. Users can use general equipment, logic
analyzer, to measure the clock jitter directly. In the built-in clock jitter
measurement method, there are fewer measurement limitations than the
traditional ways.
3.2 On-chip Jitter Measurement
On chip jitter measurement can also called built-in clock jitter measurement
(BIJM). Using BIJM method, it can avoid many problems from outside testing
equipment. The BIJM method also has many properties as following.
13
1. Low cost on testing equipment: In recent year, the output frequency of
PLL is increasing, because of the improvement of semiconductor
technology. Therefore, it is difficult to measure the clock jitter of PLL by the
traditional testing equipment. Users must buy more high level testing
equipment and software that support testing equipment to measure jitter.
When using BIJM methods, users can measure clock jitter by general
testing equipment.
2. Low noise independent of test equipment: When the output frequency
of PLL is increasing, if the clock jitter measurement is still relying on the
external equipment, it will cause many problems. The most important
problem is loading problem, the output signal of circuit will connect many
probes. This probes may induce noise to signal, these may cause output
signal distort. The measurement result of output signal will be not correct.
Using BIJM circuits, the jitter measurement is measured in chip, the output
signal will not be effect by the external equipment. The BIJM circuits can
reduce the effect of noise to minimum, so the measurement result will be
more accuracy.
3. Full speed measurement: The measurement accuracy is according to the
sampling frequency of the testing equipment. When sampling frequency is
high, the measurement result is more accuracy. But the sapling frequency
has limitation. When measuring the clock jitter in the chip, users can
measure the jitter at full speed.
4. Suitable for SoC system: In high speed communication system, the
system will compose of many circuits, such as transceiver, receiver, and
PLL. When users only want to measure the clock jitter of PLL, it is difficult
to measure. Using BIJM circuits can solve this problem.
3.3 Operation of BIJM Circuits
BIJ
M
Ou t
p ut
Figure 3.1 Operation of BIJM circuits
Seeing Figure 3.1 that shown the operation of BIJM circuits.
First step: The output signal of PLL or DLL or VCO, which has jitter, feed into
the BIJM circuit. The BIJM circuit will produce a serious digital data.
Second step: The serious digital data feed into the logic analyzer, then the
logic analyzer will measure this digital data and transfer this data to computer
format file.
Third step: Using computer to deal with the digital data that is computer
format file. The digital data can be translated into jitter histogram by mathlab.
14
3.4 Traditional Methodology
In BIJM method, Time to Digital Converter (TDC) technique used to deal with
the clock signal. Figure 3.2 will introduce the traditional TDC methodology.
Figure 3.2 Traditional methodology
Seeing Figure 3.2, input clock pass through delay cell, then input clock
become delayed clock. When passing more and more delay cells, the first
rising edge of delayed clock can sample the second rising edge of input clock.
In this way, the delay time of delay cells had been known, the numbers of delay
cells that input clock passed had also been known, the period of clock can be
quantify as P=Td*N. Td is delay time of delay cells, N is the numbers of delay
cells that input clock pass through. In this example, the input clock passed
through six delay cells, then, the first rising edge of delayed clock can sample
the second rising edge of input clock. Therefore, P= Td*6. In TDC circuits, the
circuit resolution is the delay time of delay cells. In order to quantify the period
of clock more accuracy, the circuit resolution of TDC circuits must be improve.
15
In following sections, some of BIJM circuits will be introduce.
3.5 Delay Chain Circuit
In delay chain circuit, the main components are constant delay chain,
adjustable delay chain, D-type flip/flop, and error counter. See the Figure 3.3
Figure 3.3 Delay chain circuit
3.5.1 Operation of Delay Chain Circuit
At first, the output of PLL passed through the constant delay, the input of PLL
passed through adjustable delay chain. So the jittery clock will have a constant
delay time, them the delay time of input clock can control by the adjustable
delay chain. Figure 3.4 is shown the adjustable delay chain. In the adjustable
delay chain, the digital buffers are the delay cells, it can control the delay time.
When the binary counter is increasing is increasing, the delay time will also
increase. When the delay time of adjustable time delay is increasing, the input
clock of PLL can sample the output clock of PLL. As the rising edge of input
clock sample the rising edge of output clock, the numbers of delay cells of
delay chain will be known [10]. According to the number of delay cells, the
16
clock can be quantified.
Figure 3.4 Adjustable delay chain
3.5.2 Jitter Measurement of Delay Chain Circuit
In this circuit, if the output clock of PLL is no jitter, when the input clock pass
through the adjustable to sample the output clock, the counter signal will rise
up in the same delay cell stage. If the output clock of PLL has jitter, the counter
signal will rise up in different delay cells, because that the rising edge of output
will shift left or right.
Figure 3.5 Jitter measurement of delay chain circuit
17
18
3.5.3 Characteristic of Delay Chain Circuit
Low circuit resolution: Because that the circuit resolution is delay cells of
adjustable delay chain. The delay cells are digital buffers.
Measurement error: Because that the input clock may be have jitter, when
input clock used to sample output clock, the measurement result may be have
error.
Limitation of measurement: In this circuit, it only can measure the internal
and external input of PLL.
Real time testing: This circuit can process the clock continuously. The clock
can be measured one by one.
3.6 Delay Loop Circuit
The delay loop circuit composed of NOR gate , latches, counters, encoder, and
D-type flip/flops [14].
3.6.1 Operation of Delay Loop Circuit
See the Figure 3.6. In this circuit, the clock that wants to measure will treat as
the enable signal. The enable signal can trigger the latches. In the beginning,
all outputs of latches set on 0, the rst signal sets on 1, so that output of NOR
gate sets on 0. When the clock that wants to measure feed into the circuit, the
rst signal sets on 0, then output of NOR gate will change to 1, the 1 signal will
pass through all latches, when the last stage signal 1 return to the NOR gate,
the rst signal is still on 0, so the output change to 0, then the 0 signal will also
pass through all latches, then the 0 signal will return to the NOR gate, the
output signal of NOR gates will change to 1. In the way, the signal will loop in
this circuit until the enable signal disabled. The counter will count the numbers
of 1 and 0. The enable signal can be quantified by the signal that combine
counter signal and encoder signal.
Enable 11111100000011111100000011111100000
Counter
D Q
Latch
D Q
Latch
D Q
Latch
D Q
Latch
Encoder
Counter
DFFs
RST
Enable
Output
1
0
0 0 0 00
1
1 1 1 1 1
0
Figure 3.6 Operation of delay chain circuit
3.6.2 Jitter Measurement of Delay Chain Circuit
In this circuit, if the clock has no jitter, the output of the circuit will be the same.
If the clock has jitter, the output of the circuit will be different. Comparing the
different signal, the difference between the clock will be known. In this way, the
clock jitter can be measured.
3.6.3 Characteristic of Delay Chain Circuit
Low circuit resolution: In this circuit, the circuit resolution is determined by
the latches. The delay time of latches is long.
Low chip area: Because the circuit is used in the loop type, so that the
numbers of latches can be reduced, and the chip area alas can be reduced.
19
Limitation of measurement: Before measuring the jittery clock, the clock
must divide by 2. That, the clock signal can start the circuit.
Discrete time testing: Because that the circuit is loop circuit, so that the
circuit can only process one data once time. When the process of one data is
finished, then the next data will be process.
3.7 Vernier Delay Line Circuit
Vernier delay line circuit is composed of two different time of delay cells, D type
flip/flop, and counters. The basis concept of vernier delay line is to produce
high circuit resolution [13].
3.7.1 Operation of Vernier Delay Line Circuit
Figure 3.7 Operation of vernier delay line circuit
Seeing figure 3.7, it is shown that operation of vernier delay line circuit.
Previously, the circuit resolution of the BIJM circuit is only decided by one type
delay cells. Sometimes, inverter may be used to be a delay cell, although the
20
propagate delay of inverter is very small. If users want to quantify the period of
clock more accuracy, the delay time of inverter is still too big. In order to
achieve high circuit resolution, vernier delay line circuit was proposed. In
instead of one type delay cell, vernier delay circuit used two types of delay
cells that had different delay time. The high circuit resolution was produce by
the difference of two types of delay cells that have different time. Using high
circuit resolution, the period of clock can be quantified more accuracy. In this
circuit, the rising edge of signal B is used to sample the rising edge of signal A.
When the rising edge of signal B can sample the rising edge of signal A, signal
C and D will be rise up.
3.7.2 Jitter Measurement of Vernier Delay Line Circuit
In this circuit, when measuring the clock jitter, an extra jitter free clock is need;
it is feed into the clock side, the clock that users want to measure is feed into
the data side. See the Figure 3.8.
Jitter-less Data
Ideal Clock
N-1 Counter
N Counter
N+1 Counter
Jitter Data
Jitter
N-1 Counter
N Counter
N+1 Counter
1 2 3 1
2
3
Jitter are detected on Signal 2 and Signal 3
1 2 3
2
1 2
D QData
Clock
Counter
1 N M
1
Stage 1 Stage 2 Stage N Stage M
D1
C1
A1
B1
TA
TB
D Q
Counter
TA
TB
2
D Q
Counter
TA
TB
N
2
D Q
Counter
TA
TB
M
N M
Figure 3.8 Jitter measurement of vernier delay line circuit
21
22
If the data that users want to measure is jitter-less, when it feed into to the
circuit, the signal of counter will rise up in the same position. Because that the
second rising edge of data that users want to measure is always in the same
position. If the data that has jitter feed into the circuit. Because that the second
rising edge of the jittery data will shift right of left, when the extra no jitter clock
sample the jittery data, the counter signal will rise up in different positions.
Jitter is detected on signal 2 and signal 3.
3.7.3 Characteristic of Vernier Delay Line Circuit
High circuit resolution: The circuit resolution is created by the difference of
two different delay cells. The difference can be very small, so that the period of
clock can be quantify more accuracy.
Large chip area: In order to create high circuit resolution, the circuit needs
double delay cells. Therefore, the chip area is also increase. More higher
circuit resolution, more layout area is need.
Jitter-free clock is need: When measuring clock jitter, the jitter-free clock is
feed into the clock side.
Real time testing: This circuit can process the data continuously. The data
can be measured one by one.
Chapter 4 Proposed BIJM Circuit
In proposed BIJM circuit, the circuit is improving on vernier delay line circuit.
The characteristic of vernier delay line circuit is high circuit resolution but large
chip layout area, it also need a jitter free clock. So that in the proposed circuit,
it should be have high circuit resolution, low chip layout area, and the jitter free
clock is not need. The structure and operation of the circuit will be introduce in
this chapter.
4.1 Structure of the Proposed BIJM Circuit
One Period Delay
Jitter Measurement
Figure 4.1 Structure of the proposed BIJM circuit
The structure of the proposed BIJM circuit is shown in Figure 4.1. The
proposed circuit has three components. There are one-period-delay circuit,
jitter-measurement circuit, and control logic circuit. When the circuit measured
the clock jitter, first, the circuit needs only one input, the output signal of PLL or
DLL or other circuits feed in the BIJM circuit, then the one-period-delay circuit
will produce two outputs. This two outputs will feed in the jitter-measurement
23
circuit, the control logic circuit will control the jitter-measurement circuit to
measure clock jitter. The measurement result will be show in the BIJMoutput. In
the propose BIJM circuit, the one-period-delay circuit is used to reduce the
whole chip area , the jitter-measurement circuit has high circuit resolution, so
that the jitter-measurement circuit can measure the clock jitter more accuracy.
4.2 Proposed Methodology
The proposed methodology can achieve reduce chip layout area. Proposed
methodology use a new way to quantify the period of clock.
TOTOTO TP TPScale TP
CLKin
TP = TN - TNP
Ti
0 1 2 3 Q
M0 5 10
One Period Delay Jitter Measurement
CLKin TNP TNPTN TN
1th 2th
TNPTN
Qth
MthTOTOTO
1st 2nd Nth
TO TO TO TP...
Figure 4.2 Propose methodology
The new methodology is that using two circuit resolutions to quantify the
period of the clock. Before the second edge of clock, low circuit resolution to
quantify the clock. When closing to the second rising edge, high circuit
resolution to quantify the clock. In this way, users can quantify the clock more
quickly and accurately. Because that traditional way always using the same
circuit resolution, when using low circuit resolution, clock can be quantify
24
quickly but not accuracy, when using high circuit resolution, clock can be
quantify accuracy but slowly. The new methodology used two circuit
resolutions that can achieve accuracy and quick. In one-period-delay circuit,
the circuit resolution is low and the circuit can delay the clock almost one
clock. In jitter-measurement circuit, the circuit resolution is high and the clock
can measure the clock jitter. In following sections, these two circuits will be
introduced.
4.3 One-period-delay Circuit
In this section, the structure, detail components, and operation of the circuit will
be introduced.
4.3.1 Structure of One-period-delay Circuit
Figure 4.3 Structure of One-period-delay circuit
This circuit composed of adjustable delay cells, AND gates, OR gates and
D- type flip/flops. The adjustable delay cells can control by the Ctrl signal, so
that the delay cells have two mode delay time. When Ctrl=0, delay time of
25
delay cells is TON+TO. When Ctrl=1, delay time of cells is TON. In the circuit,
Ctrl signals of the upward delay cells are always set on 1, so that the delay
time of the delay cells is TON. The Ctrl signals of the downward delay cells
are changeable. The D-type flip/flops are used to determine that whether the
CLKD+OPD outstrip one period than CLKD+IN. If the CLKD+OPD outstrip one
period than CLKD+IN, the Ctrl signals of the delay cells will be set on 1.
4.3.2 Operation of One-period-delay Circuit
In the circuit, at beginning, all Ctrl signals of downward of delay cells are set
on 0 and all Ctrl signals of upward of delay cells are set on 1. So that the
stage dealt time is TON+TO-TON=TO. In this way, the CLKD+OPD can delay TO
more than CLKD+IN in every stage. When CLKD+OPD delay almost one period
of the clock than CLKD+IN, the d-type flip/flop will detect, then, the Ctrl signals
of downward delay cells will be change on 1. In this way, the stage delay
Figure 4.4 Operation of One-period-delay circuit
time will be change to TO-TO=0. Therefore the difference of delay time
26
between the CLKD+OPD and CLKD+IN is the same until the signal pass through
all delay cells. Seeing Figure 4.4, it is shown the whole process.
4.3.3 Design of One-period-delay Circuit
In one-period-delay circuit, the circuit compose of D-type flip/flop, AND
gates, OR gates and adjustable delay cells. The D-type flip/flop was design
in TSPC circuit. The OR and AND gates was design in static circuit. Seeing
Figure 4.5, it was shown the design of adjustable delay cells.
Figure 4.5 Design of adjustable delay cell
In this circuit, M1, M2, M3, M4 can be treated as current source. M1 and M2
are always turning on, M3 and M4 is control by the Ctrl signal. The size of
M3 and M4 is bigger than that of M1 and M2. When Ctrl signal is set on 1,
M3 and M4 are turn on, the current of M3 and M4 is bigger than that of M1
and M2, therefore, the delay time of the delay cell will be decrease. When
there are only M1 and M2 turn on, the delay time of delay cell will longer
than that M3 and M4 turn on.
27
4.4 Jitter-measurement Circuit
In this section, the structure, detail components, and operation of the circuit will
be introduced.
4.4.1 Structure of Jitter-measurement Circuit
CLKD+OPD
CLKD+INVSS
buffer
buffer
D Q
BISToutputCLKtest
Control Logic
10
0
MOS1
MOS1
10
0
MOS2
MOS2
10
0
MOSN
MOSN
buffer
buffer
10
0
MOSM
MOSM
1 2 N M
Figure 4.6 Structure of jitter-measurement circuit
Seeing Figure 4.6, it is shown that the structure of jitter-measurement circuit.
This circuit composed of MOS capacitances and D-type flip/flop. The output
signal s of one-period-delay circuit will feed into the jitter-measurement circuit.
In the circuit, MOS capacitances are used to be the delay cells. Using MOS
capacitances can create high circuit resolution than using logic gates. In this
way, the circuit can quantify the clock more accuracy. The high circuit
resolution is produced by the difference of MOS capacitances turn on and off.
In the circuit, the upward MOS capacitances are always turning off; the
downward capacitances are controlled by the control logic circuit. The CLKtest
can control the control logic circuit that the signals of control logic rise up one
by one.
28
4.4.2 Operation of Jitter-measurement Circuit
In this circuit, the way of jitter-measurement works as the vernier delay line
circuit. See the Figure 4.7. At the beginning, the rising edge of CLKD+IN and the
rising of CLKD+OPD is very closing. With increasing the control signal, the
difference between the rising edge of CLKD+IN and CLKD+OPD will decrease.
When the rising edge of CLKD+OPD had sampled the rising edge of CLKD+IN, the
output signal of D-type flip/flop will rise up. According to the numbers of control
signals, the clock can be quantified. When measuring the jittery clock, because
that the rising edge of jittery clock will shift right or left, so that, the numbers of
control signal will be different.
Ti Ti-TF
1st Control signal ON
Ti-2*TF
1st 2nd Control signal ON
Ti-N*TF
1st 2nd Nth Control signal
ON
CLKD+IN
CLKD+OPD
Control Signal
BISTOutput
Figure 4.7 Operation of jitter-measurement circuit
4.4.3 Design of Jitter-measurement Circuit
The circuit is composed of MOS capacitances, control login circuit, D-type
flip/flop. The control logic circuit will be introduced in next section. The D-type
flip/flop is design in TSPC circuit. The MOS capacitances are the most
important design in the circuit. According to the property of MOS capacitances,
the very small delay time is created. See the Figure 4.8. When control the
voltage of the gate side of the MOS capacitance, the control voltage is high,
the MOS capacitance is lager, the control voltage is low, and MOS capacitance
is small. The MOS capacitance is very stable. When the MOS capacitance is
29
large, it will take long time to charge. So that, the delay time will increase. In
this way, the difference of the MOS capacitances turn on and off will be created
easily.
Figure 4.8 MOS capacitance
4.4.4 Control Logic Circuit
The control logic circuit is composed of D-type flip/flops, seeing the Figure 4.9.
Figure 4.9 Control logic circuit
Before the last D-type flip/flop, the output Q of D-type flip/flop is always
connect to the next input D of the next D-type flip/flop. In the last D-type
30
31
flip/flop, the output of the Qbar will connect to the input D of the first D-type
flip/flop. In the beginning, all the D-type flip/flops should be reset. Then the
output Qbar of the last D-type will be set to high. When control clock start the
circuit, the control signals will rise up one by one. So that, the control logic
circuit can control the jitter-measurement circuit.
4.5 Operating Principle
Data flow of self-sampled vernier delay line (VDL) structure is divided into
two steps. First step is one-period-delay and second step is jitter-measurement.
TOPD is the timing difference between CLKD+IN and CLKD+OPD of
one-period-delay (OPD) block. TJ is the timing delay used to control the sample
window of jitter-measurement (JM) block. M is the number of processed input
clock period, and X is the upper limit number for M. N is the controlled delay
stage number for TJ, and C is counted number for the sample window of TJ.
Furthermore, TO and TF are timing units in OPD block and JM block. And, TP is
the period of each clock cycle.
When input clock is fed into the self-sampled VDL structure in first step, TOPD
is set to zero initially. Then, TOPD starts to increase with the fixed timing delay
TO continuously to produce internal signals, CLKD+IN and CLKD+OPD. When TOPD
is larger than TP, TOPD is forced to decrease with one fixed timing delay
conversely. At this time, TOPD is smaller than TP again, but TOPD has
approached to TP. Therefore, the first rising edge of CLKD+OPD signal will leads
before the second rising edge of the CLKD+IN signal. And, the timing difference,
TOPD, between CLKD+IN and CLKD+OPD is approached to one-period-delay time
in first step.
In second step, sample window, TJ, is defined by delay stage number, N, in
the beginning. And, the timing difference between CLKD+IN and CLKD+OPD is
extended to TOPD plus TJ. Then, the jittery clock period TP is compared with the
system defined timing difference TOPD plus TJ .And the comparison result is
stored in the counter value, C, in JM block. At this moment, the number of
processed input clock period, M, is increased. And, its compared with the
system defined upper limit number of processed input clock period, X
continuously until M is larger than X. In the end, the measurement result is
translated into the hit number, C divided by X, in JM block to generate the
cumulative distribution function (CDF) of input clock jitter.
Figure 4.10 Operating principle
32
33
Chapter 5 Chip implementation
The simulation result, layout of the chip, chip measurement is shown in this
section. Figure 5.1 is the design procedure.
Reference Material
Circuit Structure Specification
Draw Schematic Decide Size of MOS
Netlist CDL out
Pre-Simulation
Timing & Cost fit specification
Layout
DRCDesign Rules
Checking
LVSLayout versus
Schematic
Post Simulation
Timing, Cost, Area fit specification
Put PAD in Layout
DRC & LVS
Output GDS File IC fabricate
IC Measurement
Figure 5.1 Design procedure
5.1 Simulation of the proposed circuit
The circuit is design for measuring clock frequency from 100MHz to 400MHz.
The circuit resolution of jitter-measurement circuit is 15ps. The circuit is design
in TSMC 0.35um 2P4M CMOS process.
34
5.1.1 Consideration of Simulation
When the chip was fabricated, it has inductances on its bounding pads. When
chip working, the inductances will cause the ground bounce effect. The ground
bounce effect can induce noise into the power line and ground line, so that the
chip function may fail or the performance will decrease. Therefore, in the
simulation, the ground bounce is also simulated. See Figure 5.2. In the power
land ground line, there are 3nH inductances.
Figure 5.2 Simulation consideration
5.1.2 Simulation of One-period-delay Circuit
In this circuit, the circuit resolution is design in 320ps. In other words, TO is
320ps in a stage. Because that the input frequency is 100MHz from 400MHz,
in order to measure 100 MHz frequency, the circuit must almost delay 10000ns.
35
In this way, the circuit needs 10000/32031 stages at least. Considering of
process variation, see Figure 5.3. When delay cells work on FF mode, the
circuit resolution TO is 250ps, so that the circuit designed in 40 stages.
Figure 5.3 Simulation of circuit resolution in one period circuit
5.1.3 Simulation of Jitter-measurement Circuit
In this circuit, the circuit resolution designed in 15ps in a stage. The total delay
time of all stages is double of the circuit resolution of one-period-delay circuit.
So, the circuit resolution of one-period-delay circuit is 300ps, and the numbers
36
of stage in jitter-measurement circuit is 40(300*2/15=40). Figure 5.4 is shown
the process variation of circuit resolution in jitter-measurement.
Figure 5.4 Simulation of circuit resolution in jitter-measurement circuit
5.1.4 Simulation of Control Logic Circuit
The control clock can control the control signals rise up. It can control how long
the control signals rise up. In this design, the clock pass through 16 times clock,
then the next control signal will rise up. See Figure 5.5
37
Figure 5.5 Simulation of control logic circuit
5.2 Layout of the Proposed Circuit
In the Figure 5.6, it is the photograph of whole chip layout. The layout block of
whole chip is shown in Figure 5.7. In this chip, it included the proposed BIJM
circuit and the PLL circuit. The chip area of the proposed BIJM circuit is
38
500um*750um.
Figure 5.6 Photograph of whole chip
Figure 5.7 Layout block of whole chip
5.3 Measurement Result
5.3.1 Concept of Measurement
In order to measure the chip, the measure signal is created from the pulse
generator, because that the signal is not really ideal, so it will have jitter. So
39
that the signal feed into the BIJM circuit, the circuit can measure the signal.
Then the output of BIJM will be measured by the logic analyze. The output
signal of BIJM circuit will be process by computer, then, the jitter histogram will
be created by computer.
Figure 5.8 concept of measurement
5.3.2 Chip Measurement
The setup of measurement in the chip is shown in Figure 5.9. The power is
provided by the Agilent E3646A. The input signal is created from Agilent
81334A, and HP 33120A pulse generator provided different noise into Agilent
81334A, so that the signal will have different jitter histogram. In this way, the
chip can measure differ histogram.
40
Agilent E3646APower Supply
Agilent 81134APulse Generator
HP 33120APulse Generator
Hp 16702ALogic Analysis system
Figure 5.9 Measurement setup of the chip
5.3.3 Measurement Result
In order to measure the chip, the input signal is induced noise. In this way,
the chip can measure the signal. In Figure 5.10, the input signal is 100MHz
that mixed with the random noise, so that the signal has jitter that is 214ps
(peak to peak). In the Figure 5.11, it is the result that measured from the BIJM
circuit. In the measurement result, the jitter (peak to peak) is 17*15ps=255ps.
In Figure 5.12, the input signal is 100MHz that is mixed with the square signal
that is 100KHz , 300mVolt. In the Figure 5.13, it is the result that measured
from the BIJM circuit. In the measurement result, the jitter (peak to peak) is
18*15ps=270ps.
41
Figure 5.10 100MHz signal mixed with random signal
150
350
550
750
950
1150
1350
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28Stage number
Hit number
Figure 5.11 Measurement result of proposed circuit I
42
Figure 5.10 100MHz signal mixed with square noise
0
200
400
600
800
1000
1200
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32Stage number
Hit number
Figure 5.11 Measurement result of proposed circuit II
43
44
5.3.4 Specification of Circuit
Technology TSMC352P4M
Power Supply 3.3V
Input Frequency 100MHz~400MHz
Jitter Resolution 15ps
Jitter Range 15ps~330ps
Power Dissipation 29mw(at 100MHz)
Measurement VS. BIJM 210ps(peak-peak) VS.
255ps(peak-peak)
Table 5.1 Specification of BIJM circuit
Reference Logic Vision [12] G. W.
Robers[13] G. W. Robers[15]
Propose Cirucit
Technology 0.6um 0.35um 0.18um 0.35um Resolution(ps) 40ps 118ps 18,9ps 15ps Area (gate) 3000~6000(gates) 5.5mm2 0.12mm2 0.375mm2
Technique Delay Chain Vernier Delay
Vernier Delay
Vernier Delay
Table 5.2 Comparison of BIJM circuit
45
Chapter 6 Conclusion and Future Work
6.1 Conclusion
In this work the circuit for clock jitter-measurement is proposed. The proposed
circuit is divided into two structures. One is one-period-delay circuit; the other
is jitter-measurement circuit. In one-period-delay circuit, the circuit can reduce
the chip area. In jitter-measurement circuit, the circuit has high circuit
resolution, in order to measure the clock jitter accurately. The testing frequency
of the proposed circuit is 100HMz to 400MHz. The circuit resolution is 15ps.
The circuit is fabricated in TSMC 0.35um 1P4M CMOS process. The chip area
of the proposed circuit is 500umx750um. The power consumption is 29mw.
6.2 Future Work
The circuit resolution must be improve, because the frequency of PLL will
increase. The most important factor of circuit resolution is the noise. Because
that the circuit resolution will smaller than 10ps, so that, when there is noise in
the circuit, the circuit resolution will be interfere with the noise. The circuit must
resist the noise. The D-type flip/flop is the circuit is used to determine the jitter.
But the D-type flip/flop has setup time and hold time, those factors will affect
that determine the jitter. So the design of D-type flip/flops can be improve.
46
Reference material
[1] C.C. Tsai, On-Chip jitter measurement for phase-locked loop. MSc. Thesis,
National Chiao Tung University, Taiwan, 2002.
[2] M. M. Gourary, S. G. Rusakov, S. L. Ulyanov, M. M. Zharov, K. K. Gullapalli,
and B. J.Mulvaney, A new approach for computation of timing jitter in phase locked
loop, Proc.of Design, Automation and Test in Europe and Exhibition 2000, pp.
345-349, 2000.
[3] F. Azais, M. Renovell, Y. Bertrand, A. Ivanova, and S. Tabatabaei, A unified
digital test technique for PLLs: catastrophic faults covered, Proc. of Int. Mixed
Signal Testing Workshop, pp. 269-292, June 1999.
[4] Bell Research Laboratories, SONET transport systems: Common criteria network
element architecture features, GR-253-core, Issue 1, pp. 5-81, Dec. 1994.
[5] W. Dalal and D. Rosenthal, Measuring jitter of high speed data channels using
under-sampling technique, Proc. of Int. Test Conf., pp. 814-818, 1998.
[6] Nelson Soo, Jitter measurement techniques, Pericom Application Brief AB36,
Nov.2000.
[7] K. A. Jenkins and J. P. Eckhardt, Measuring jitter phase error in microprocessor
phase-locked loop, IEEE Design & Test of computers, vol.17, pp. 86-93, Apr-Jun
2000.
[9] Bozena Kaminska, BIST means more measurement options for designers, EDN
Magazine, Dec. 2000.
[10] M. Frisch Arnold and H. Thomas Rinderknecht, Jitter Measurement System and
Method. US Patent # 6,295,315 assigned to Frisch, et al. Sep. 2001.
47
[11] S. Tabatabaei and A. Ivanov, Embedded timing analysis: A SoC infrastructure
IEEEDesign & Test of Computers, vol. 19, pp. 22-34, May-June 2002.
[12] S. Stephen and R. Audin, BIST for Phase-Locked Loops in Digital
Applications Proc.
of Int. Test Conf., pp. 532-540, Sep 1999.
[13] A. H. Chan and G.W. Roberts, A synthesizable, fast and high-resolution timing
measurement device using a component-invariant vernier delay line Proc. of Int. Test
Conf., pp. 858-867, Nov 2001.
[14] T. Lin, K. L. Luo, Y. J. Chang and W.C. Wu, A testable design of on-chip jitter
measurement, VLSI Design/CAD Symposium, Taidong, Taiwan, Aug 2002, pages
182-185
[15] Antonio H. Chan and Gordon W. Roberts, A Jitter Characterization System Using a
Component-Invariant Vernier Delay Line Very Large Scale Integration (VLSI)
Systems, IEEE Transactions on Volume 12, Issue 1, Jan. 2004 Page(s):79 - 95
.pdf93ncu_au(digital).pdf.pdfTable 5.2 Comparison of BIJM circuit.44List of FigureChapter 1 Introduction1.1 Motivation1.2 Organization of Thesis
Chapter 2 Preview Work2.1 Jitter Definition2.1.1 Cycle-to-Cycle jitter2.1.2 Period Jitter2.1.3 Long-Term Jitter I2.1.4 Long-Term Jitter II
2.2 Jitter Histogram
Chapter 3 Jitter Measurement Technology3.1 Off-chip Jitter Measurement3.2 On-chip Jitter Measurement3.3 Operation of BIJM Circuits3.4 Traditional Methodology3.5 Delay Chain Circuit3.5.1 Operation of Delay Chain Circuit3.5.2 Jitter Measurement of Delay Chain Circuit3.5.3 Characteristic of Delay Chain Circuit
3.6 Delay Loop Circuit3.6.1 Operation of Delay Loop Circuit3.6.2 Jitter Measurement of Delay Chain Circuit3.6.3 Characteristic of Delay Chain Circuit3.7 Vernier Delay Line Circuit3.7.1 Operation of Vernier Delay Line Circuit3.7.2 Jitter Measurement of Vernier Delay Line Circuit3.7.3 Characteristic of Vernier Delay Line Circuit
Chapter 4 Proposed BIJM Circuit4.1 Structure of the Proposed BIJM Circuit4.2 Proposed Methodology4.3 One-period-delay Circuit4.3.1 Structure of One-period-delay Circuit4.3.2 Operation of One-period-delay Circuit4.3.3 Design of One-period-delay Circuit
4.4 Jitter-measurement Circuit4.4.1 Structure of Jitter-measurement Circuit4.4.2 Operation of Jitter-measurement Circuit4.4.3 Design of Jitter-measurement Circuit4.4.4 Control Logic Circuit
4.5 Operating Principle
Chapter 5 Chip implementation5.1 Simulation of the proposed circuit5.1.1 Consideration of Simulation5.1.2 Simulation of One-period-delay Circuit5.1.3 Simulation of Jitter-measurement Circuit5.1.4 Simulation of Control Logic Circuit
5.2 Layout of the Proposed Circuit5.3 Measurement Result5.3.1 Concept of Measurement5.3.2 Chip Measurement5.3.3 Measurement Result5.3.4 Specification of Circuit
Chapter 6 Conclusion and Future Work6.1 Conclusion6.2 Future Work
Reference material