Transcript No Slide Title

Technical University of Košice
Faculty of Electrotechnic and Informatics
Department of Electronics and Telecommunications
Education and Training
Data Acquisition Systems
Introduction
part 1
Linus Michaeli
6th Summer School on Data Acquisition Systems
Benevento, Italy, 26.June 2006
Measuring chain with converter
selected parameter
 Structure of measuring chain. Analog part brings systematic error
into measuring chain and interferences with noise sources.
 System configuration of the measuring signal conditioning with analog
multiplex
zi(t)
ki(ntvz)
Output
display
unit
Microcontroler
z1(t)
Conditioning
circuits
Sampling
circuit +ADC
Object
xi(t)
Sensor
Multiplexer
x1(t)
Master
processing
computer
Diagnostic
Autocalibration
Communication bus
 Diagnostic feedback for autocalibration and diagnostic
 Feedback from the controller is dedicated to process control.
Measuring chain with converter
selected parameter
System configuration of the measuring signal
conditioning with digital multiplex
SC+ADC
xi(t)
z1(t)
Conditioning
circuits
zi(t)
k1(ntvz)
SC+ADC
ki(ntvz)
Diagnostic
Autocalibration
Microcontroler
Object
x1(t)
Sensor
Output
display
unit
Master
processing
computer
Communication .bus
Errors in Signal conditioning
blocks
Noise sources affecting measuring signal transmission:
EMC from power line and electronic power supplies.
EMC from telecommunication transmitters (TV, AM and FM radio)
Corona effects in insulators .
Atmospheric ionisation.
Thermal noise,1/F noise, flicker noise
DC drift, instability of parameters.
1/year 1/day 1/hour
50 100 150Hz AM
...
FM/TV
log f ...
GSM
Intelligent sensor
Generalised measuring chain
Is represented by Intelligent sensor
•Embedded ADC
•DAC pulse width modul.
•Microcontroller
•Communication via bus
n. sensor
1.sensor
fx
Internal
sensor
Amplifier,
multiplexer
Excitation
generator
ADC
CPU
Communication bus
EEPROM
Timer,
PWM clock
•Common clock
•Plug&Play
Power
supply
Power supply
Sensors, Intelligent sensors will be presented
by Prof. Helena M.G. Ramos
GND
Real time clock
Serial bus
Instituto Superior Tecnico, Lisboa, Portugal
PWM
Basic operation in mutual
conversion analog and digital signal
Discretization in time and bandwidth constraint
TS  1
2fH
x s (t ) 

 x(t ) (t  nTS )
n 
1
X S 2f  
TS

 
k 
X
2

f


   TS 
k  
Ilustration of the aliasing effect
(t-nTS )
(f-k/TS )
f
Spectrum overlaping
Quantisation
Discretization in level and quantisation noise
 x nTS 
k nTS   round 

 Q
1

2
The quantisation noise
EQ 
k
Ukvantiz.
t
Uanal
UŠ
t
2
1  Q.t 

 dt  Q
12
T  2T 

Measuring chain as a generalised
AD Converter
 The measuring chain could be considered as a generalised ADC
 Input physical quantity x(iT), Output digital sample k(iT)
x(iT)
z(iT)
Conditioning
circuits
Sampling
circuit +ADC
Physical
quantity
Sensor
k(iT)
Generalised ADC
 Analog conditioning block main error sources both systematic & random
 Shape of transfer function with discontinuities and time sampling caused
by ADC
Analog signal preprocessing blocks, trends and limitations: Prof.Ramon Pallas-Areny
Technical University of Catalonia (UPC, Spain)
ADC & DAC Transfer Curve
Number of bits N
Resolution 2N
Full scale range FSR
Signal to Noise ratio
(dB)
S
Effective number of
bits
Code transition level
Code bin width

ENOB  N  log 2 rms
 eq
T (k )
Q
N
 N  6,02  1,76
ENOB
Decreasing
noise power
N
dB
Error sources
Error sources
Stochastic error sources uA:
• Thermal noise- Gaussian distribution of the power 2n
• Noise generated by the EM interference. General
description by the power of induced voltage at the input of
the DAQ 2EMC
• Quantisation noise 2Q
Systematic error sources uB:
• Gain error
• Offset
• Integral nonlinearity
e( x )  G.x  O  INL( x )
Distribution of systematic error parameters is based upon assumption
of the uniform occurrence in the interval G.FS,O,INLMAX with zero mean
 G  G. X
2. 3
; O  O
3
;  INL  INLMAX
3
Total measurement uncertainty
The „total noise“ can be estimated by the limits as a „type A“
uncertainty
2
uB2

G

DC
.X
 O
2
2
 INLMAX 
2
3
Gain, offset and INL errors can be assessed by the „type B“ uncertainty.
u A2 
 n2
Q2

2
 EMC
Q2
  Q2
The combined uncertainty uC and expanded uncertainty U with coverage
factor 2
uC  u A2  uB2
ADC&DAC parameters and
their testing
Problems:
•Unambiguous definition
•Explicit testing approaches
Testing methods:
• Required uncertainty of the instrument ensuring the metrological
property of the testing stand on the accuracy level 10 % corresponds
to the ADC resolution.
1 1
 
2N 10
•Time consumpting  Static and Dynamic Test methods
Definition of the ADC and DAC parameters and Introduction to ADC testing – some nonstandardized methods: Prof.Jan Saliga, Technical University of Kosice
Introduction to ADC testing based on the 1241 standard discussion of basic choices
concerning the 4-parameter algorithm:Prof.Istvan Kollar, University of Technology and
Economics, Budapest
Lecture around the 3 and 4 sine wave parameter fit:Prof.Peter Händel Royal Institute
of Technology (KTH), Stockholm
DAC testing: Eulalia Balestieri, UNI Sannio Benevento
State of the Art in the
Standardisation process
Report about activity of the IEEE TC-10
Standards:
• IEEE 1057
• IEEE 1241
• DYNAD
Thank you for your time and
attention
ADC & DAC Transfer Curve
Problem:
How to test
•Static testing methods
•Dynamic testing methods
Effective number of
bits

ENOB  N  log 2 rms
 eq
Code transition level T ( k )
Code bin width
Q
ENOB
Decreasing
noise power
N