transparencies - Indico
Download
Report
Transcript transparencies - Indico
AD/DA Conversion Techniques
An Overview
J. G. Pett
Introductory tutorial lecture for :‘Analogue and digital techniques in
closed-loop regulation applications’
17/09/2002
for terminology see Analog Devices Inc.
AD/DA
Introduction to the subject
Understanding conversion methods
Methods
Parameters
The past, the present and the future
Introduction
What are AD/DA Converters
What are they used for
Why do you need to know how they work
Digital coding methods
Waveform digitising
CERN examples
What are AD/DA
Converters (1)
An Analog to Digital converter [AD or ADC]
is an electronic circuit which accepts an
analog input signal (usually a voltage) and
produces a corresponding digital number at
the output
An Digital to Analog converter [DA or DAC]
is an electronic circuit which accepts a
digital number at its input and produces a
corresponding analog signal (usually a
voltage) at the output
They exist as modules, ICs, or fully
integrated inside other parts, e.g. µCs
Photos
What are AD/DA Converters (2)
Analog
Digital
discrete time world
continuous time world
+/-10v
ADC 1
DAC 1
16
12
COMPUTER
or µP/µC
+/-5v
ADC 2
16
Typical AD & DA Application
+/-10v
The Real World
The Real World
continuous time world
Analog
What are they used for
Any time a real world analog signal is
connected to a digital system
CD players, GSMs, DVMs, Digital Camcorders
etc, etc
CERN control systems & instruments
HOWEVER, each application has particular
needs
Resolution - number of bits
Speed and Accuracy
Level of input/output waveforms
Cost etc
Why do you need to know
how they work
Because the theoretical course you will
shortly undertake assumes perfect
converter products - BUT
Practical converters have :
Many conversion methods - why
Trade-offs between resolution and speeds +
delays
Different methods of “sampling” the
waveforms
A large number of basic and method-dependent
error sources
Manufacturers specifications which ‘differ’ AND
Almost all converters need some analog ‘signal
conditioning’ which is application dependent
Digital coding methods (1)
8,10,12,14,16,18, 20-24bits?
+10v
Most/Least significant bit
MSB/LSB
Uni-polar, bipolar, straight
binary, 2’s complement invert MSB
Parallel I/O or serial [delay]
0v
Bytes or words
Double buffering
Digital ‘breakthrough’
Digital correction methods
Time skewing & jitter
-10v
AD/DA Transfer Characteristic
0000
8000
8000
FFFF
0000
7FFF
FFFF
Digital coding methods (2)
Resolution = 2n-1
n
8bits
10bits
12bits
14bits
16bits
18bits
20bits
22bits
24bits
[n = number of bits]
2n
256
1024
4096
16384
65536
262144
1,048576
4,194304
16,777216
1bit ppm [1x10-6]
3906
976
244
61
15
3.8
0.95
0.24
0.06
Digital value
Waveform digitising (1)
time
A waveform is ‘digitised’ (sampled) at a constant
rate D t
Each such sample represents the instantaneous
amplitude at the instant of sampling
Between samples the value remains constant [zero
order hold]
What errors can occur in this process ?
Waveform digitising (2)
C
A
D
B
A & B show aliasing in the time domain
C & D show a different case in the frequency
domain
- it is important to understand these effects
Waveform digitising errors
For a DAC
output waveform is a ‘distorted’ version of original
higher frequencies not reproduced - aliasing ?
‘average shape’ displaced in time
‘sharp’ edges need filtering
For an ADC
converter sampling errors
with a ‘sample & hold’ circuit ahead of the converter?
integrating action during part, or all of the sample-time
?
conversion time
data ‘available’ delay
aliasing - [ is multiplication of input spectrum and
fs]
…[must ‘remove’ all spectrum > fs/2 before
sampling]
Sampling rate
Nyquist rate = 2x highest frequency of
interest
Practically, - always sample at least 5x, or
higher
Ensure ADCs have input filtering [anti-alias]
where necessary [large hf signals]
Filter DAC outputs to remove higher
frequencies and switching ‘glitches’
‘Over-sampling’ converters sample x4 to
x500 - this may reduce above problems
and/or extend resolution
CERN examples
Many PLCs with analog values, such as
temperature, to measure : 10 - 12bit <10kHz
PS, SPS, LHC control instrumentation, such
as power converter control, regulation and
monitoring : 16 - 22bit <1kHz
Beam instrumentation, experiments : high
speed: 10 - 12bit 25ns
ETC ETC
Photos
1969
ISR Beam-Transfer DAC
[5 decimal decades]
1973
ISR Main Bends DAC
[16bit binary
Relay switching
Kelvin-Varley divider
All electronic switching
Photos
ADC Sigma-Delta 1998
1989
LEP 16bit Hybrid DAC
Understanding Conversion
Methods
AD/DA Methods
Some very simple ideas
DAC circuits
Basic ADC circuits
Successive approximation, flash - S&H
Integrating - single/dual/multi slope
Charge balance, D
Some very simple ideas
‘Digitally set’
potentiometer
Comparator
DAC
ADC
Vref
dial
Vdac
equal
=
Unknown
voltage
ADC =
precise reference voltage
comparison of divider value with unknown [analog input]
“digitally adjustable” divider or potentiometer [output
value]
DAC =
precise reference voltage ……. {multiplying dac}
“digitally adjustable” divider or potentiometer [input
value]
optional output amplifier of pot. value [analog output]
DAC circuits (1)
Simplified binary weighted resistor DAC
R - 2R ladder DAC
8.75V
9.375
max.
Summation of binary weighted currents
Modern DACs use the ‘R-2R ladder’
DAC circuits (2)
Important circuit concepts
Resistor tracking - temp. & time > ratios
Switch is part of R [on & off resistance]
Limits for tracking and adjustment
Switch transition times - glitches
Switched current sources are faster
Other DAC methods
DC performance not needed for all uses
Different ladders, Caps. as well as Resistors
PWM, F>V
Sigma-Delta
Performance cannot be better than the Reference
- {multiplying DAC concept}
Basic ADC circuits (1)
Simple ramp and comparator ADC
Unknown
analog
input
start
Binary output
Digitising begins with a ‘start’ pulse
DAC is ramped up from zero
counter stopped by comparator when Vin = DAC out
ADC output is counter value
Tracking ADC
Basic ADC circuits (2)
This ADC circuit is limited and rarely used
WHY
slow
variable time to give result
input signal can vary during digitising
Successive Approximation ADC solves these
problems - using
complex logic to test and retain each DAC bit
a sample and hold circuit ahead of the
comparator
Successive Approximation
ADC
Fast process - 1 100µsecs
Result always n clocks
after start
Used extensively for
12-16bit DAQ systems
Vref
Flash ADC
Half-Flash
Vref
analog
input
The fastest process <50nsecs
Limited resolution typically 8 10bits
Half-flash technique is cheaper
analog
input
Flash
Sample & Hold Circuit (1)
LF398
Essential for defining the ‘exact’ moment of
sampling
Circuit introduces other error sources [ see (2) ]
Sample & Hold Circuit (2)
Storage Capacitor Waveform