Text preview for : 1260355E_Artifact_Calibration_AN_w.pdf part of Fluke 1260355E_Artifact_Calibration_AN_w Fluke 5700a 1260355E_Artifact_Calibration_AN_w.pdf
Back to : 1260355E_Artifact_Calibra | Home
Artifact calibration
theory and application
Application Note
Introduction
Electronic instruments gen-
erally contain both a large
number and a wide variety of
components. The circuit con-
figuration and the values of
the components determine the
characteristics of the instru-
ment. Unfortunately, because
nothing is absolutely stable, the
value of any component varies
with time, and because of this,
instruments require periodic
calibration to assure continued
compliance with specifications.
Until the advent of the micro-
processor, periodic calibration
generally required the physi-
cal adjustment of components
within the instrument. This was
done to bring the instrument
into compliance with external
standards. Complex instruments
might contain dozens of internal
physical adjustment points such
as potentiometers and vari-
able capacitors. The adjustment
process could take many hours Figure 1. A Calibration Lab in an Instrument.
to complete.
This approach to calibration
requires traceable stimulus and
measurement at each of these In the mid-1970s, instru- standards) has been utilized
points. The systems used have mentation broke new ground extensively in the calibration
been both manually operated by using the microprocessor, of instruments. Today, inter-
and complex. Such systems not only to enhance capabili- nal software corrections have
may include various refer- ties and operation, but also to eliminated the need to remove
ence components or stimulus simplify the calibration process. instrument covers to make
values, as well as bridges and For example, the Fluke 8500A physical adjustments in almost
other instruments. The support (a high-accuracy multimeter) all types of instrumentation.
of these complex and lengthy was designed to store and use However, for instruments that
calibrations required a large software correction factors do not support Artifact Calibra-
and costly array of equipment, to compensate for gain and tion (defined below), it is still
processes and manpower. zero errors on each range of necessary to provide a large
the instrument. This process array of external stimulus or
of storing constants (based measurement capability for
on comparison to external purposes of calibration.
F ro m t h e F l u k e C a l i b r a t i o n D i g i t a l L i b r a r y @ w w w. f l u k e c a l . c o m / l i b r a r y
Artifact standards and The expansion included the
null detector and decade
artifact calibration additional functions of alter-
divider built in. And it would
nating voltage, resistance and
use those built-in devices to
An artifact standard is a stan- direct and alternating current.
transfer the accuracy of the
dard that maintains a small, The 5730A is a higher perfor-
artifact to the many ranges of
concise set of derived values. mance variant of the 5700A.
the instruments. In essence, an
An example of this is a 10 V Inside the 5730A there is a
instrument capable of Arti-
zener reference such as the null detector for making com-
fact Calibration takes over the
Fluke 732B DC Transfer Stan- parison measurements and
manual metrology functions of
dard. Typically, the artifact divider for scaling between
establishing ratios and making
standard is in the category (and ranges. The inclusion of the
comparisons. This is done by
of the technology) commonly measurement system in the
placing circuitry, microproces-
considered to be a transfer instrument being calibrated
sor control and software inside
standard. This is in contrast to eliminates the need for the
the instrument so that it can
an intrinsic reference such as operator to read the difference
perform these same functions.
the Josephson Voltage Refer- between the externally applied
The driving force behind this
ence which generates values voltages and internally gener-
change has been the need to
based on physical constants. ated voltages and allows the
reduce the time and equipment
Artifact Calibration is the instrument's software to control
costs associated with conven-
process of transferring the the nulling process. The null
tional manual or semi-manual
assigned value(s) of an artifact detector zero is calibrated and
calibration and to provide more
to a large array of multidimen- uniform quality. made traceable by periodic
sional parameters. Typically the adjustment against an internal
Technological advances in
term Artifact Calibration is used short.
components and software are
to describe the process when Internally, the 5730A is
now allowing manufacturers
it is implemented internally in configured to emulate activities
to emulate what is humorously
an instrument. For example, in a conventional metrology
illustrated in Figure 1. Null
consider the calibration of a dc lab. A microprocessor controls
detectors can be built on a chip.
source that has several ranges all functions and monitors
Ratio systems can be reduced to
extending from millivolts to performance, routing signals
a single circuit board. Thin-film
one kilovolt. To calibrate such between modules by way of a
resistor networks can replace
an instrument, whether it uses switch matrix. Like all modern
bulky wire-wound resistors.
internally stored constants or instruments, no physical cali-
requires manual adjustment, Artifact calibration bration adjustments are made.
you ordinarily need an exter- and the Fluke 5730A Instead, correction constants
nal reference voltage such as
Calibrator are stored in non-volatile
a zener reference or standard memory. Numerous internal
cell; a null detector to make The Fluke 5440A Precision
checks and diagnostic routines
comparisons; a Kelvin-Varley DC Voltage Calibrator1 was
ensure that the instrument is
ratio divider (which is usually introduced in 1982 and was
always operating at optimum
self-calibrating), and a decade the first instrument to employ
performance. A proprietary ultra
divider. For calibration, this Artifact Calibration. This limited
linear pulse width modulated
array of equipment is connected embodiment of Artifact Cali- digital-to-analog converter
in various configurations to pro- bration uses an external 10V (DAC) functions as a divider
vide the traceable source and reference and decade divider
within each calibrator. This
measurement parameters. as traceable standards. Com-
divider, like any ratio divider
Now consider the calibra- parisons are made using an
such as a Kelvin-Varley divider,
tion of the dc source with the external null detector, and
functions on the basis of
capability of Artifact Calibra- through this process internal
dimensionless ratio. That is,
tion. Then all that is necessary references and dividers are
there are no absolute quantities
is to apply the artifact, in this calibrated. The Fluke 5700A,
involved. The repeatable linear-
case a dc reference. The dc introduced in 1988, expanded
ity of a pulse width modulated
source being calibrated would on the capability of the 5440A's
DAC depends only on a highly
have to have the equivalent Artifact Calibration techniques. 2
reliable digital pulse train. To
of the Kelvin-Varley divider, maintain high confidence, this
2 Fluke Calibration Artifact calibration theory and application
DAC Setting
10 Volt Range
N4
N3
20 Volt Range
N2 Artifact Calibration
N1
Artifact Ratio Internal Ratio Ranges and
Standard Transfer References Transfer Functions
Output
V1 V2 Voltage
Figure 2. Digital-to-analog converter verification. Figure 3. The traceability chain for the 5730A.
linearity is checked and verified 5730A.4 One thermal converter then establishes ratios of
during Artifact Calibration. This makes real time ac/dc compari- various other values within the
is done by comparing two fixed son measurements to maintain calibrator, and stores them in
voltages on different ranges of the output voltage. A second is non-volatile memory.
the DAC. Figure 2 illustrates used only during Artifact Cali-
this comparison. It should be bration to compare the external The traceability path
noted that the precise values of dc artifact to the internally gen- Traceability is often defined
the two voltages V1 and V2 are erated ac voltages. To maintain as "the ability to relate indi-
unimportant; it is only required confidence, a software routine vidual measurement results to
that they be stable during the directs intercomparison of the national standards or nation-
measurement process. two converters to ensure that ally accepted measurement
If the DAC is perfectly linear, their characteristics track each systems through an unbroken
then: N4/N3 = N2/N1 other. The traceability of this chain of comparisons ...." This
An analog-to-digital con- internal ac/dc reference, used requirement must be met with
verter (ADC) provides null only during Artifact Calibration, Artifact Calibration as rigorously
detection capability. Using the is verified by periodic compari- as it is with all other calibra-
ADC together with the DAC, son to an external ac/dc transfer tion methods. This means that
comparisons are made and reference like the Fluke Calibra- no adjustments can be made
values assigned for the cor- tion 792A or 5790A. without comparison to traceable
rection constants stored in Such external ac/dc transfer standards, and that all transfer
memory. verifications are recommended of values must be done using
Two reference amplifiers3 to be done once every two reliable ratiometric techniques.
similar to those used in the years, as is common for such The Artifact Calibration block
Fluke Calibration 732B DC thermal transfer devices. diagram shown in Figure 3
Transfer Standard maintain the The transfer of resistance illustrates the unbroken trace-
5730A's accuracy and stability. references to the 5730A is ability chain. The values of the
These references are calibrated similar in concept to the trans- external artifacts are trans-
by comparing them to the fer of direct voltage described ferred to the internal references
external 10 V artifact standard. earlier. Two resistors, having by a built-in self-calibrating
This comparison takes place values of 1 and 10 k , form ratio device (like the self-cal-
internally, using the DAC and the working internal refer- ibrating Kelvin-Varley divider
null detector to assign correc- ences for the calibrator. Their in the lab). The ratio device
tion values. values are assigned using then transfers values from
Two Fluke Calibration solid- the DAC and null detector by the references to the output
state thermal rms converters comparing them to external parameters. The integrity of the
form the alternating voltage resistance artifact standards system is enhanced by built-in
measurement reference for the like the Fluke Calibration Model self-check routines and through
742A-1 and 742A-10K. The periodic verification by external
DAC and null detector system comparison.
3 Fluke Calibration Artifact calibration theory and application
With Artifact Calibration, Consequently it is only neces- Data collection
there are several important sary for the user to reverify
The driving force behind Arti-
factors to consider during the the function of the circuitry on
fact Calibration has been the
design and manufacturing a very infrequent basis. The
need to reduce the operator time
processes. They assure that performance of this group of
required to calibrate precision
manufacturing reliably pro- instruments demonstrates that
instruments, along with a reduc-
duces instruments that truly the Artifact Calibration pro-
tion in the amount of support
and fully meet their calibration cess has properly adjusted the
equipment required. A second-
criteria. instruments. The measured
ary benefit--one with potentially
These vital factors are: results indicated in Figure 4 are
more impact on the metrology
1. The design must be correctly similar to those that would have
function--is the opportunity for
analyzed to identify sources been obtained using traditional
data generation, collection and
of error. The possibility of a manual calibration methods.
analysis.
design oversight cannot be The results are shown to fall in
In order to implement Artifact
ignored. Rigorous testing and a Gaussian (normal) distribution
Calibration in an instrument,
analysis must be performed with the predominant value
that instrument must include
during the instrument's (mean) centered at nominal.
sophisticated analog hardware
development. The 5730A production
as well as a microprocessor and
2. Manufacturing processes criteria is set so that each
software. With internal refer-
must ensure that compo- verification point must show
ences and internal comparison
nents and construction a 3-sigma normal distribution
capability, the capacity is there
meet design criteria. These limited to 80 % of the instru-
to collect data at the time of
processes must be monitored ment's 24 hour specifications.
Artifact Calibration. Perhaps
to ensure consistency of This is equivalent to 3.75-sigma
more significantly, the capability
production. relative to 100 % of the speci-
is there to execute these rou-
3. Instrument operation must be fication, or approximately 1 out
tines between calibrations. This
fully verified in production toof 10,000 will fall beyond the
allows the measurement of drift
eliminate the possibility of limits. Results have shown that
and of performance changes rel-
unusual faults.5 This veri- the Artifact Calibration process
ative to the internal references.
fication must itself be fully exceeds these criteria for most
A traditional instrument that
traceable. The production verification points. In addition
is reviewed only during calibra-
process of the Fluke Calibra- to the 3-sigma criteria, other
tion (say once every 6 months to
tion 5730A includes 237 statistical tests are performed
one year) may go out of calibra-
verification points. Data from on the data to ensure that the
tion without the knowledge of
each of these verifications production process is in control.
the user. Where critical tests rely
is collected and analyzed to 4. The instrument must have
on the instrument's accuracy,
ensure that the production diagnostic routines capable
this lack of awareness may have
process is in control. A repre- of verifying that its inter-
extremely costly and potentially
sentative chart is illustrated nal calibration system is
dangerous consequences.
in Figure 4. A more compre- functioning correctly. These
hensive display of the data is routines should establish the
shown in Figure 7. same confidence in calibra-
tion as is expected with an
The chart shows data collected operator performing manual
on 100 instruments produced calibration, using conven-
over a 60-day period. It is tional techniques. This may
typical of data collected on well be achieved (as is the case
over many thousands of instru- with the 5730A) by main-
ments. This external verification taining a set of internal,
confirms the integrity of the environmentally controlled
circuitry used to assign values references. These are used
based on Artifact Calibration. to make periodic internal
comparisons.
Figure 4. Measured results on Artifact Calibrated Instrument.
4 Fluke Calibration Artifact calibration theory and application
The ability to run internal 3. All the individual data points Using the data collected from
Calibration Checks between measured in the cal check an instrument it is possible
external calibrations allows process. to detect a potential problem
the operator to monitor the 4. Analysis of the measured with that instrument. This will
performance between calibra- shifts in this data since the eliminate the implications and
tions and helps to avoid these time of the previous calibra- cost of an instrument being out
situations. If the instrument's tion shown in measured of specification during use. Also,
internal references are well absolute shift values and if enough history has been col-
controlled and impervious to shifts relative to the applica- lected, it is possible to extend
environmental changes, then ble instrument specification. the period between Artifact
these Calibration Checks can be Calibrations and reduce future
performed with the instrument maintenance and calibration
in its working environment. costs.
This instills confidence without
the need to return the instru- FLUKE Model
Serial Number
5730A
5730005
ment to the calibration lab. Note Report String 5730A
that these Calibration Checks Report:
Printed:
CALIBRATION CHECK
6/27/2014 7:14:13
do not adjust the instrument's SW version
Installed
v1.05.00
A8: Switching Matrix
output, but merely evaluate the Installed A11: DC Volt Module
instrument's output against Installed
Installed
A12: AC Volt Module
A16: 220V Module
internal references. Comparison Installed
Installed
A14: 1100V/2A Module
A7: Current Module
of the internal reference values Installed A9/A10: Ohms Module
to external traceable standards Installed
Installed
A13: Hires Osc Module
A21: Rear Panel
is necessary to make traceable Reference Cal Date Temp Prev Date Temp CalCheck Temp
internal adjustments. MAIN OUTPUT: 5/13/2014 23 12/04/2013 23 5/13/2014 23
Using statistical techniques6 ZERO: 6/26/2014
it is possible to analyze the DC VOLTAGE
DC
Range
220 mV
Zero Shift Magnitude
1.16E-07
Abs Shift
2.20E-01 -3.60E-07 V
Rel Shift
-1.637
Spec
9.2727 ppm
Pct of Spec
-17.654
data collected during Arti- DC 220 mV 1.16E-07 -2.20E-01 5.93E-07 V 2.6932 9.2727 ppm 29.0446
fact Calibration and internal DC
DC
2.2V
2.2V
4.07E-09
5.81E-12
2.20E+00 -3.01E-06 V
-2.20E+00 3.01E-06 V
-1.3684
1.3702
4.3636 ppm
4.3636 ppm
-31.3585
31.4009
Calibration Checks. With a DC
DC
11V
11V
2.03E-08
0.00E+00
1.10E+01 -7.85E-06 V
-1.10E+01 7.87E-06 V
-0.714
0.7158
3.2727 ppm
3.2727 ppm
-21.8166
21.873
computer, data can be imported, DC 22V -2.85E-07 2.20E+01 -9.79E-06 V -0.4452 3.2273 ppm -13.7948
processed, displayed and used DC
DC
22V
220V
5.29E-07
-2.66E-05
-2.20E+01
2.20E+02
1.00E-05 V
7.24E-06 V
0.4563
0.0329
3.2273 ppm
4.2273 ppm
14.1385
0.7781
to assess the performance of DC
DC
220V
1100V
-2.70E-05
-4.81E-11
-2.20E+02 -6.09E-05 V
1.10E+03 7.78E-04 V
-0.2766
0.7071
4.2273 ppm
6.4545 ppm
-6.5433
10.9546
the instrument. DC 1100V -2.03E-06 -1.10E+03 -7.80E-04 V -0.7089 6.4545 ppm -10.9833
RESISTANCE Range Magnitude Abs Shift Rel Shift Spec Pct of Spec
Constant performance Resistance
Resistance
0.00E+00
1.00E+00
0.00E+00
0.00E+00
Ohm
Ohm
0.00E+00
0.00E+00 95
0 ppm
ppm
0
0
Figure 5 shows data collected Resistance
Resistance
1.90E+00
1.00E+01
0.00E+00
-2.13E-05
Ohm
Ohm
0.00E+00
-2.13E+00
95
25
ppm
ppm
0
-8.517
using the Calibration Check Resistance
Resistance
1.90E+01
1.00E+02
-6.51E-05
-7.67E-05
Ohm
Ohm
-3.43E+00
-7.67E-01
25
11
ppm
ppm
-13.7021
-6.9728
function of a particular 5730A, Resistance 1.90E+02 -3.78E-04 Ohm -1.99E+00 11 ppm -18.0869
where the output parameters Resistance
Resistance
1.00E+03
1.90E+03
2.10E-04
2.21E-04
Ohm
Ohm
2.10E-01
1.16E-01
7.2
7.2
ppm
ppm
2.9157
1.6149
are measured relative to its Resistance
Resistance
1.00E+04
1.90E+04
5.50E-03
8.94E-03
Ohm
Ohm
5.50E-01
4.70E-01
7
7
ppm
ppm
7.8582
6.7201
internal references. This Cal Resistance 1.00E+05 5.46E-02 Ohm 5.46E-01 8 ppm 6.8274
Check data is available in a Resistance
Resistance
1.90E+05
1.00E+06
1.64E-01
1.06E+00
Ohm
Ohm
8.64E-01
1.06E+00
10
14
ppm
ppm
8.637
7.585
comma separated variable data Resistance
Resistance
1.90E+06
1.00E+07
1.90E+00
1.80E+01
Ohm
Ohm
9.99E-01
1.80E+00
17
37
ppm
ppm
5.8754
4.8713
format (CSV) through the USB Resistance 1.90E+07 2.15E+01 Ohm 1.13E+00 47 ppm 2.4084
port on the calibrator or through Resistance 1.00E+08 -3.02E+02 Ohm -3.02E+00 110 ppm -2.7412
inquiry from a computer via a AC VOLTAGE
AC
Range
2.2 mV
Magnitude Frequency
2.20E-03 All Freqs.
Abs Shift
-5.39E-09 V
Rel Shift
-0.0002
Spec
0.2363 %
Pct of Spec
-0.1036
remote interface. AC 2.2 mV 2.20E-03 2.00E+04 5.59E-09 V 0.0003 0.2363 % 0.1075
The data file shows: AC
AC
2.2 mV
2.2 mV
2.20E-03
2.20E-03
5.00E+04 -1.20E-08
1.00E+05 -1.10E-08
V
V
-0.0005
-0.0005
0.2503 %
0.3267 %
-0.2174
-0.1531
AC 2.2 mV 2.20E-03 1.20E+05 -6.27E-09 V -0.0003 0.6655 % -0.0428
1. Identification and con- AC 2.2 mV 2.20E-03 1.20E+05 -1.07E-08 V -0.0005 0.6655 % -0.0728
figuration information of AC
AC
2.2 mV
2.2 mV
2.20E-03
2.20E-03
2.00E+05 1.33E-08
3.00E+05 -2.41E-08
V
V
0.0006
-0.0011
0.6655 %
0.6655 %
0.0906
-0.165
the specific 5730A being AC 2.2 mV 2.20E-03 4.00E+05 1.48E-08 V 0.0007 1.2864 % 0.0521
AC 2.2 mV 2.20E-03 5.00E+05 3.59E-08 V 0.0016 1.2864 % 0.1268
checked. AC 2.2 mV 2.20E-03 6.00E+05 -1.73E-08 V -0.0008 1.4464 % -0.0543
2. Dates for the most recent AC
AC
2.2 mV
2.2 mV
2.20E-03
2.20E-03
7.00E+05 -1.65E-08
8.00E+05 -2.70E-08
V
V
-0.0007
-0.0012
1.4464 %
1.4464 %
-0.0518
-0.0847
calibrations and zero- AC 2.2 mV 2.20E-03 9.00E+05 -1.30E-08 V -0.0006 1.4464 % -0.0408
ing events on the 5730A
calibrator. Figure 5. Data collected using the Calibration Check function of a 5730A, where the output parameters are
measured relative to its internal references.
5 Fluke Calibration Artifact calibration theory and application
Verification system and you will notice through this and you can avoid the use of
analysis program procedure that the equipment potentially out of calibration
is only off line when it is found equipment. By this means you
Each and every 5730A is Arti- to be necessary rather than on may predict problems before
fact Calibrated, monitored for a rigorous scheduled basis. In they occur.
stability using the Calibration addition, if something shows Figure 6 shows Process
Check function and then finally signs of going wrong before a Metrology.
verified using the custom-built regularly scheduled calibra-
automatic test equipment. The tion that will also be obvious,
parametric capabilities of such
equipment are described in
detail in An Automatic Test Statistically
Analyzed
System for a Multifunction
Calibrator, Measurement Sci-
ence Conference, 1989.5 Each
ca