Transcript Introduction

Analog electronics – general introduction







Analog – continuous in time
Digital – discrete in time
Design of amplifiers and filters
ADCs
Logic gates
Receivers, transmitters
Storage cells
L
001
010
100
Sensor
Amplifier
Filter
ADC
DSP
E
S
1
Advanced analog circuit design
1
0
1
0
1
Analog electronics – general introduction
 Digital design: compromise between power consumption and processing speed
 Analog design: compromise between speed, power consumption, resolution, supply voltage,
linearity…
 Analog circuit are crosstalk and noise sensitive
 Analog design can‘t be automatized
 Different levels of abstraction
G
D
S
B
PMOS
A
Transistor
Verstärker
Advanced analog circuit design
System
2
Analog electronics in scientific applications
 Particle detectors with high spatial resolution
- Semiconductor detectors with spatial resolution are today widely used in consumer digital cameras,
professional HDTV cameras, medical imaging and in science-grade instruments for particle physics,
astronomy, material and biology studies (x-ray diffraction imaging, electron-microscopy) and many
other fields.
- Spatial resolution of semiconductor detectors is achieved by segmenting the sensor surface into
many small picture elements ("pixels"). Every segment has its own signal collecting region that can
be readout individually.
- These detectors are distinguishable from the sensors for consumer electronics either by its low noise
and single-particle detection capability or by other properties such as 100% fill-factor, high time
resolution, high dynamic range, radiation tolerance, etc.
 Multi-channel systems
 Pixel electronics
 Signal amplification, signal transmission, sampling, comparison, A/D conversion, timemeasurements, amplitude measurement
 Amplifiers, filters, switched-voltage/current circuits, comparators, A/D convertors,
oscillators…
 AC analysis, feedback
 Transistor models
 Noise, threshold dispersion
 Semiconductors – solid state physics
Advanced analog circuit design
3
Pixel electronics
Amplifier
P-”guard-ring”
N-well
Filter
Comparator
SRAM
Hit memory
DAC
55 μm
Advanced analog circuit design
4
Pixel sensors for particle physics
 Pixel sensors are used to detect high-energy charged particles, and to determine particle
trajectories.
 Since particles tracking requires many layers of planar detectors, tracking sensors should be
as transparent for particles as possible. They should be very thin, otherwise the particles will
be deflected from their initial trajectories.
 Silicon is the best material for such detectors since silicon-based technologies offer the
possibility to implement any possible semiconductor device (from PN junction to the
completed signal processing electronics) on the sensor.
Advanced analog circuit design
5
Pixel sensors for particle physics
Advanced analog circuit design
6
Pixel sensors for particle physics
Advanced analog circuit design
7
Pixel-sensors for medical imaging
 In the case of high energy photon (x-ray or gamma) detection for medical imaging, the
requirements are opposite. Photon sensors should be thick enough to absorb the largest
part of the radiation. Due to its low absorption coefficient, silicon is not the best material for
high-energy photon detection.
 The most of practical pixel sensors for such radiation are based on indirect detection. Such
sensors consist of a layer of scintillator material that converts the high-energy photons into
visible light. The light detection is then performed by a silicon pixel sensor layer.
Advanced analog circuit design
8
Pixel-sensors for medical imaging
g
Scintillators
SIPMs
PCB1
SIPM signals
Readout chip
Bias voltages
PCB2
Digital output signals – time & energy
Control
PCB3
USB Cable
FPGA
Digital output signals
PCB4
USB Chip
Supply voltages
Advanced analog circuit design
9
Classification of pixel-sensors
 Hybrid- and monolithic detectors
- Monolithic pixel detectors: An n x m pixel matrix is placed on one chip and usually
connected by means of signal multiplexing to n (or less) readout channels placed on the
same or different chip. Pixels of a monolithic detector must be equipped with a certain
readout electronics that at least perform the simplest tasks such as signal clearing,
multiplexing and in most cases the amplification. (Some of monolithic detectors employ
even more complex in-pixel signal- processing and data reduction. In this case we are
talking about "intelligent" pixels that can e.g. detect particle hits, perform A/D
conversion, transmit pixel addresses, perform time measurements, etc.) There are n or
less connections between the pixel matrix and the block of readout channels.
- Hybrid pixel detectors: Each pixel on the sensor chip has its own channel on the readout
chip. There are n x m connection between two chips.
 Technology – custom or specific
- The development of such detectors is relatively low-cost since they use modern
commercially available and well characterized CMOS technologies.
- Pixel detectors in the technologies that are specially developed or adjusted for particle
(or visible light) detection, like the technologies on high resistance substrate, thick epilayer, etc.
Advanced analog circuit design
10
Hybrid detectors with fully-depleted sensors
n-type collecting region
(n-diffusion)
Pixel i
Pixel i
Substrate
Signal collection
P-type Si - depleted
P-type Si - undepleted
Advanced analog circuit design
Potential enegry (e-)
P-type Si - depleted
P-type Si - undepleted
11
Hybrid detectors with fully-depleted sensors
 Standard (bump-bonded) hybrid
pixel detectors
Pixel
Readout chip
Min. pitch ~50 μm
Advanced analog circuit design
Fully-depleted sensor
Signal charge
Bumps
- The bump-bonded hybrid pixel
detectors are used in highenergy physics for particle
tracking, and in medicine and
synchrotron experiments as
direct detectors for x-rays. They
are based on a relatively simple
pixel sensor (ohmic or with pn
junctions) without any pixel
electronics and bumpconnections between the pixel
sensor and the readout pixel chip
- The connection between the
sensor and the readout chip is
mechanically complex and
expensive, especially in the case
of small pixel sizes.
12
Hybrid-detector for cell imaging
Power/signal supply for RO-chip
Bonding matrix for one RO-chip
Pixel matrix
Advanced analog circuit design
RO-chip (in a “gel”-pack)
13
Capacitive coupled hybrid detector
Pixel
Readout chip
Glue
Advanced analog circuit design
Smart diode- or fully-depleted sensor
Signal charge
14
Capacitive coupled hybrid detector
Power supply
and cont. signals
for the readout chip
1.5 mm
Power supply
and cont. signals
for the sensor
Readout chip (CAPPIX)
Sensor chip (CAPSENSE)
Advanced analog circuit design
15
3D hybrid-detector
Pixel
Readout chip2 Readout chip1
 3D-integration is a technology
that allows for both vertical and
horizontal connection between
electronic components placed on
different chips (thinned dies)
stacked vertically.
Wafer bond
TSV
Wafer bond
Advanced analog circuit design
Fully-depleted sensor
Signal charge
16
Standard monolithic detector - MAPS
NMOS transistor in p-well
N-well (collecting region)
Pixel i
P-type epi-layer
P-type substrate
Energy (e-)
Charge collection (diffusion)
MAPS
 In the case of a standard monolithic CMOS sensor ("Monolithic Active Pixel Sensor“) - the
sensitive area is undepleted epitaxially-grown silicon layer and the charge is spread and
separated by diffusion. Some part of the charge is finally attracted by the next
well/diffusion.
Advanced analog circuit design
17
Standard monolithic detector - MAPS
Select(i)
Select(i+1)
Signal out
P-type epi-layer
P-type substrate
 Pixel rows are consecutively "selected" by connecting their outputs (usually single-transistor
amplifier outputs) to column lines. The pixel signals are in this way transported to the
readout channels. Such a multiplexing requires at least one electronic switch per pixel
implemented with a transistor.
Advanced analog circuit design
18
Standard monolithic detector - MAPS
 MAPS are slower and not as radiation tolerant as the hybrid detectors.
 standard MAPS do not allow implementation of complete set of CMOS electronics inside
pixels (only n-channel FETs - NMOS transistors - can be used)
N-well (collecting region)
Pixel i
NMOS transistor in p-well
PMOS transistor in n-well
P-type epi-layer
P-type substrate
Signal loss
Advanced analog circuit design
Signal collection
Energy (e-)
MAPS with a PMOS transistor in pixel
19
Enhanced MAPS
Pixel
PMOS in a shallow p-well
NMOS shielded by a deep p-well
N-well (collecting region)
P-doped epi layer
INMAPS
Advanced analog circuit design
20
T-well detector and smart diode array
P-well
Deep n-well
Pixel
2. n-well
NMOS
PMOS
Diffusion
Epi-layer
T-well MAPS
Pixel
Potential energy (e-)
“Smart” diode
Deep n-well
Drift
Potential energy (e-)
Depleted E-field region
P-substrate
“Smart diode” array
Advanced analog circuit design
21
SOI monolithic detector
CMOS pixel electronics
Connection
Electronics layer Buried oxide
 An SOI detector is based on a
modified SOI process. SOI detectors
use the electronics layer for the
readout circuits and the highresistivity support layer as a fullydepleted (drift-based) sensor. The
sensor is typically 300um thick and
has the conventional form of a
matrix of pn junctions. A connection
through the buried oxide is made to
connect the readout electronics with
the sensor.
Energy (e-)
Support layer
Advanced analog circuit design
22
DEPFET monolithic detector
Pixel
PMOS
Ext. gate
Clear
Elect. Interact.
Int. gate
Int. gate
Signal clearing
Potential en. (e-)
Signal collection
N-substrate (depleted)
P-type backside contact
Advanced analog circuit design
23
SDD monolithic detector
Drift “rings”
N-doped collecting region
Energy (e-)
Depleted n-type substrate
Undepleted p-type backside contact
Advanced analog circuit design
24
Monolithic detector - SDA
ADC channel
Pixel matrix
2.7 mm
Advanced analog circuit design
25
Amplification
 In its simplest form, pixel signal amplification is performed using a single-transistor amplifier.
In the case of Field Effect Transistors (FETs), a single-transistor amplifier is sensitive to the
voltage change on its input (gate). The charge signal generated by ionization is first
collected by the collecting region. The amplifier is coupled with the collecting region by
means of DC-coupling (wire) or by use of AC-coupling (capacitance). The conversion factor
between the charge signal and the voltage change is the capacitance of the collecting
region, referred to as detector capacitance. Clearly the voltage signal will be higher if the
collection region has smaller capacitance.
 More efficient amplification is achieved by multi-transistor amplifiers. Such amplifiers are
typical for hybrid detectors and advanced CMOS monolithic detectors. They are often
equipped with feedback circuit which makes the amplification more linear. An example of an
amplifier with feedback is the charge sensitive amplifier - CSA. CSA is sensitive only to the
charge injected into its input, the capacitance of the input node does not influence the
output signal amplitude.
Advanced analog circuit design
26
Amplification
Bias V
Bias V
Charge sensitive amplifier
Bias R
Bias R
Out
Out
Isig
Cdet
Isig
Simple voltage amplifier
(source follower)
Detector (equivalent circuit)
Advanced analog circuit design
Cdet
Detector
27
Noise
 An amplifier not only performs the amplification of the input signal; unfortunately it also
introduces electronic noise. Let us explain this: Every amplifier needs to be biased in order
to achieve the desired amplification, which means that the amplifier transistor(s) must
conduct a certain bias- (DC) current. The signal on transistor's gate will then modulate the
current. Thermal motion of the charge carriers inside the transistor active region (channel),
leads to bias current fluctuations. These fluctuations are small compared to the bias current
itself, but since the bias current is almost always much larger than the signal, its noise can
in many cases exceed the signal. A way to decrease the noise is to extend the measurement
time (or add a low-pass filter/shaper). Noise signals are random signals with expected value
zero and if the measurement takes long time, the average of the noise during measurement
interval will in fact approach zero. Most signals, however, have nonzero DC value and they
are unaffected by the measurement time.
 We could conclude that the detector capacitance does not play any role if we use CSA. This
is, however, not true. The noise of a charge sensitive amplifier depends linearly on the
detector capacitance. The reason for this is that the negative feedback which cancels the
output noise becomes less efficient if the input amplifier node is loaded with a large
capacitance.
Advanced analog circuit design
28
Noise
25.0m
Noiseless signal
Signal with noise
20.0m
Signal [V]
15.0m
10.0m
5.0m
0.0
-5.0m
-10.0m
0.0
500.0n
1.0µ
1.5µ
2.0µ
2.5µ
3.0µ
Time [s]
Advanced analog circuit design
29
„Time walk“
0.14
Response to 600 e
Response to 6000 e
0.12
0.10
Signal [V]
0.08
0.06
0.04
Threshold
0.02
0.00
-0.02
0.0
500.0n
1.0µ
1.5µ
2.0µ
Time [s]
Time walk ~ 70 ns
Advanced analog circuit design
30
KTC Noise
 Almost every electronic circuit that employs transistors will be affected by their noise. This
holds also for the transistor-based pulsed-reset circuit. During the pulsed reset, i.e. when
the reset switch is closed, the potential of the collecting region will fluctuate around the
desired reset value due to the thermal noise in the reset transistor. When the reset transistor
is turned off, the instantaneous value of the reset voltage will be frozen. The instantaneous
value is the sum of the desired reset-voltage and the reset error. The reset error superposes
to the signal and leads to a measurement uncertainty. It is interesting to note that the reset
noise only depends on the detector capacitance (not on the reset transistor resistance):
 σ2v = kT/Cdet,
 with σ2v variance of the voltage reset error, k Boltzmann's constant, T temperature and Cdet
detector capacitance.
Advanced analog circuit design
31
KTC Noise
Reset switch closed
Reset switch opened
Reset voltage
1.806
Reset
Reset voltage [V]
1.804
1.802
Reset error
1.8 V
Reset switch
1.800
Reset voltage
Desired reset voltage = 1.8 V
1.798
Detector c. = 1 fF
1.796
1.794
960.0n
980.0n
1.0µ
1.0µ
1.0µ
Time [S]
Advanced analog circuit design
32
Properties of pixel sensors











Properties
Pixel size
Detector capacitance
Noise
- readout amplifier
- reset- and bias-resistor noise
- The leakage-current noise
- σ2v = kT/(gm t).
- The magnitude of the noise determines the smallest detectable signal.
Signal to noise ratio (SNR)
- SNR is the ratio between a chosen reference signal and the noise.
- SNR ~ (gm t)0.5/Cdet
Dynamic range
- Dynamic range is the ratio between the greatest undistorted signal (the greatest signal for which the
readout does not saturate) and the smallest detectable signal (determined by the noise).
Time resolution
Power consumption
- FOM = P t / SNR2
Radiation tolerance
Fixed pattern noise
- FPN refers to a non-temporal spatial noise and is due to device mismatch in the pixels and/or
readout channels.
Radiation length
Advanced analog circuit design
33