Confused by calibration concepts - Seeking help with the basics

[nick1980]

I really don't understand the concept of calibration. Who can help me please? Who can suggest me some good websites or FREE materials to look at? I have read the QS9000 MSA book but I don't understand what it said.

My major confuses:

1. A master equipment and a under-test equipment calibration concept.

i) master equipment has its error from the manufacturer, e.g. +/-1V

ii) under-test equipment has its error from other manufacturer,e.g. +/- 2V

iii) From my calibration form for the master equipment, it has specification limit, e.g. +/- 0.1V.

iv) From my calibration form for the under-test equipment, it has the specification limit, e.g. +/- 0.1V.

v) Can I use a less accurate equipment as a master to calibration a more accurate equipment? (Someone have answered me before it can't. But I think the error of the under-test equipment can be downgraded by master equipment. e.g., master error=+/-2V, under-test eq.=+/-1V, result: under-test eq. error = +/-2V after calibration by less accurate master eq.)

vi) how to decide the specification limit of master equipment and under-test equipment? Who decides them? by what reasons? (From last audit, auditor said the master eq.'s tolerance must be 10% the under-test eq.'s tolerance) Is it right? Why?

a) tolerance related to equipment error??

b) Tolerance must include master eq. error and under-test eq. error?

c) e.g., master voltmeter error=+/-1V, master spec. limit=+/-0.1V, under-test voltmeter error=+/-2V, under-test spec. limit=+/-0.5V. I checked a voltage: under-test showed 5V, master showed 4.6V, 5-4.5=0.4V (within limit)

ci) I guess that: 0.4V+/-2V (under-test voltmeter error)+/-1V (master error)---> it must out of limit!!! 1V error means when I measure a reading 0.4V +/-1V = 1.4V or -0.6V.---which out of 0.5V under-test spec. limit.

cii) am i right?

 

[Jim Howe]

For whatever reason I had this saved in my library.

Softer ratio requirements

We can reduce, but not eliminate the problem by softening the requirements for a ratio. ISO 10012-1, for example, supporting the ISO 9000 series, recommends at least 3:1 and prefers 10:1 (but does not require any specific value). ANSI/NCSL Z-540-1, another important measurement standard, calls for 4:1.


PHILIP STEIN is a metrology and quality consultant in private practice in Pennington, NJ. He holds a master's degree in measurement science from George Washington University, in Washington, and he is a Fellow of ASQ.
Hope it helps!

 

[Hershal]

Nick,

You have quite a range of questions. Let' see if I can get through them. Others will also contribute.

First, a working definition. Calibration in its simplest form is a comparison of a device or instrument to a more "accurate" device or instrument, under controlled conditions, to establish how accurate and repeatable the device or instrument is, so you know how accurate and repeatable your measurement is. The lineage of the measurements must be to what is known as SI Units, and have what is known as measurement uncertainty calculated at each step going from your instument back to SI Units.

SI Units are the International System of units of measurement. They come in two varieties. Base SI, of which there are 7, and derived SI, which take the 7, mix and match, and do some math to obtain an infinite number of units.

Measurement uncertainty is basically the mathematical description of errors associated with any given measurement.

In the case of the Master and instrument under test, what is meant here is that th master is a "higher level" instrument and can be used to calibrate "lowere level" instruments. Good calibration practices, ANSI/NCSL Z540-1-1994, and ISO/IEC 17025 all require the master to be used for calibration only, not calibration plus something else.

The calibration ratio, also known as a test accuracy ratio (TAR), is a ratio of how accurate the master is compared to the instrument being calibrated. Ratios will vary depending on standard (documentation standard) used. ISO 10012 suggests a minimum 3:1, meaning the master is at least three times more accurate than what is being calibrated. ANSI/NCSL Z540-1 requires a minimum 4:1 in absence of measurement uncertainties. ISO/IEC 17025 does not recognize the ratio, and requires measurement uncertainties. A 10:1 ratio is considered optimal. In theory, a 1:1 ratio can be used, and in some cases that is the best available, but there are a lot of additional steps to take to assure the validity of the measurements.

The best advice I can give is to take some basic measurement courses at a local college, and outsource the calibration to a calibration laboratory accredited to ISO/IEC 17025. Finding a laboratory should be fairly easy by going to an accrediting body's website.

In the U.S., accreditation is administered by a number of accrediting bodies. The main ones include:

IAS http://www.iasonline.org
A2LA http://www.a2la.org
NVLAP http://ts.nist.gov/ts/htdocs/210/214/214.htm
L-A-B http://www.l-a-b.com

There are others also.

Countries other than the U.S. have their own accrediting bodies.

I hope this helps.

Hershal

 

[Hershal]

Nick,

One final note. Given your location, the accrediting body close to you is HKAS. I don't have their address handy, but you can find it by going to http://www.aplac.org

Hershal

 

[nick1980]

I can't send them out for calibration

THANKS FOR ALL YOUR CONCERN!!!

I hv totally hundreds of equipments such as voltmeters, ammeters, pH meters, etc. needed to be calibrated so I must do it internally.

That's why I choose the "master and under-test equipment" calibration method to save the money.

I 'm very confused about accuracy, error, tolerance between master and under-test equipments. I don't know which step I should consider first.

e.g.
calibration decision process
step 1) Master eq. accuracy >= under-test eq. accuracy
step 2) Determine master tolerance
...... I don't know the decision steps.

 

[Al Rosen]

For whatever reason I had this saved in my library.

Softer ratio requirements

We can reduce, but not eliminate the problem by softening the requirements for a ratio. ISO 10012-1, for example, supporting the ISO 9000 series, recommends at least 3:1 and prefers 10:1 (but does not require any specific value). ANSI/NCSL Z-540-1, another important measurement standard, calls for 4:1.


PHILIP STEIN is a metrology and quality consultant in private practice in Pennington, NJ. He holds a master's degree in measurement science from George Washington University, in Washington, and he is a Fellow of ASQ.
Hope it helps!

I believe Philip Stein passed away in June.

 

[Jerry Eldred]

There were many very good responses above. Nevertheless, I'll give my "two cents" worth of input...

First, do I understand correctly that you are in Hong Kong? If so, I have contacts there who may be able to give some input. I am not free to mention them in this forum. But I can contact them for input and advice on what resources may be available in your area.

There is a company in the U.S. which sells calibration training on CD ROM. Send me an email and I can give you that information if you wish. Since it is only one company, I wouldn't want to commercialize this forum by naming them. If anyone knows of multiple companies that provide calibration training on CD ROM, please feel free to reply.


Calibration can be simply thought of as comparing an unknown less accurate instrument against a known more accurate instrument. "Unknown" means it hasn't been tested or verified, and "Known" means that it has been certified.

It all comes down to making sure that what you measure is the same as what someone else or some other instrument measures. Most countries maintain a set of EXTREMELY accurate instruments or standards in a highly controlled laboratory. Those standards are compared against those of other countries to make sure there is international agreement on a measured value. Companies such as yours or mine, compare our less accurate unknown instruments against higher accuracy known instruments, which are compared eventually against the standards in our national standards labs. This is known as TRACEABILITY (very important concept).

So when you check your unit under test with a master (or standard) the master (or standard) must be of known accuracy, and more accurate than the unit you are calibrating. There are detailed reasons why it is important the the master unit to be more accurate (which I won't cover for now). A rule which has long been used is 4 to 1 accuracy (uncertainty) ratio. This provides you a statistical safeguard to assure your under-test instrument really meets specs when you calibrate it.

The spefications for your master unit must be at least four times more accurate than the unit under test. The actual values that the manufacturer measured when calibrating the master must be within those specifications.

When you calibrate your under-test unit using the master unit, the error on your under test unit must not be outside it's specified tolerances.

There is a term used by some in the calibration world called "Uncertainty Growth". This means that from the moment you calibrate an instrument, it will statistically become less and less accurate over time. Therefore it is also important to determine how much time there should be between each calibration of an instrument, so that you maintain an acceptable confidence that the measurements made are still within the instruments specified tolerances. This is also a complicated topic, which I can not adequately cover here. I advise you to learn about that as well.

If you calibrate a more accurate under-test unit with a less accurate master unit, there is the danger that you will actually make the under test less accurate after calibration. This depends on how your under-test units are used. You need to compare the specifications on the under-test units against those of your product.

Since you mentioned MSA, I presume that possibly you have to meet QS9000 or TS16949 requirements. If this is correct, and if your under test units are used to measure important product parameters, specifications for your under-test units will be less than those of the master unit. Therefore, there is possibility that this will cause inadequate stability in MSA.

Specifications for a master unit and under-test unit are defined by the manufacturer. If the manufacturer did not define specifications, you must statistically determine what specifications each will meet. This is also a complex topic, and difficult for a user to do.

The specifications for a master unit are defined by a manufacturer by thorough testing under very controlled conditions. This amounts to verification over time of short term and long term variability for the master unit, and under what conditions. This requires use of a higher accuracy standard than the master unit, known use conditions and defined amount of time. This is done by engineering and/or design persons where it is manufactured.

Other specifications can be derived, but it is a long, detailed process. So I also can not cover that here.

The limit (specification) is the maximum amount of error allowed for the instrument to still be considered to meet its specifications.

the error is the actual measured value during calibration, before any adjustments are made. Your error must not exceed specifications.

I'm not sure if my additional comments have been any more help. Please let me know for any further questions.

 

[qualitygoddess - 2010]

RE: I can't send out for calibration...............there are many accredited labs that will come to your site and do the calibration for you. They will typically ask for an owner's manual of the device, as this helps the calibration company do the best job for you. The manual contains information about the device and its capabilities. I have used outside services for many years and have always found them to do an outstanding job. Be sure you use a calibration lab that is certified.

I do agree that you should take a calibration course, or read up on the basics of metrology. I did that, and with the outside lab, was able to do everything needed. I have done this method at 3 different companies now, and have never failed an audit for calibration.

 

[nick1980]

Any good suggetions about calibration course?

I worked in a leadframe which supplies for semionductor firms.
I found this one:
Is it OK?
Who can suggest more suitable for me?
What should I study? ISO 17025? ISO 11012?


A)
http://www.hktrainingonline.com/eng/training/detail.cfm?itemcode=04015934&navid=3&sort=10&idnav=99

B)
ISO Gum and ISO 10012: Workshop in Laboratory Equipment Calibration and Measurement:
This course will explain the ISO 9000 and ISO/IEC 17025 calibration requirements for clinical and testing laboratories and their underlying principles. It will provide participants with hands-on experience of the calibration of common laboratory instruments based on the standard criteria specified in the "ISO Guide to the Expression of Uncertainty in Measurement" (GUM) and Guidance from ISO 10012

 

[Charmed]

NIST Calibration Services

Dear Nick1980:

Please try the following link to NIST Calibration Services.

http://ts.nist.gov/ts/htdocs/230/233/calibrations/

Calibration Services, NIST, 100 Bureau Drive, Stop 2330, Gaithersburg, MD 20899-2330
Telephone: 301-975-2092, Fax: 301-869-3548, E-Mail: [email protected]


When on their website, click on Electromagnetic Measurements. Then navigate your way and click on the measurement of interest to you. Just see how comprehensive the list is for Electromagnetic Measurements. Of course, many other types of measurements and calibrations are covered. So, I decided to post this info here.

Your problem is somewhat different, as I understand. You are trying to understand the meaning of calibration. Others have provided inputs in that regard. Hope this helps too. Good luck.

Charmed


Electromagnetic Measurements
Fees for services are located directly below the technical contacts for each service category link.
Links referencing the service id numbers are for detailed descriptions of specific services.

Resistance Measurements
DC Resistance Standards and Measurements
Special Resistance Measurement Services, by Prearrangement (51100S)
Measurement Assurance Program for Resistance (51110M)
Standard Resistors, 1 and 10 k (51130C and 51131C)
Standard Resistors 10-4 to 106 (51132C-51142C)
High-Value Standard Resistors: 10 7 to 10 12 (51143C-51154C)
High-Current Standard Resistors—Shunts (51160C-51163C)

High-Voltage Standard Resistors
High-Voltage Standard Resistors (51210C)
High-Frequency Standard Resistors
High-Frequency Standard Resistors; Two-Terminal (51310S)

Impedance Measurements (Except Resistors)
Low-Frequency Capacitance and Inductance Measurements and Standards
Special Four Terminal–Pair (4TP) Capacitance and Dissipation Factor Characterization (52100S)
Low-Frequency Capacitance Measurements and Inductance Measurements and Standards (52110S-52181C)

Special LF Impedance Measurements, by Prearrangement (52110S)
Special Measurement Assurance Program for Standard Capacitors (52120S)

Fused-Silica Dielectric Standard Capacitors (52130C-52131C)
Standard Capacitors (52140C-52176C)
Standard Inductors, Self or Mutual (52180C-52181C)

High-Frequency Standard Capacitors and Inductors
High-Frequency Standard Capacitors and Inductors (52210S-52310S)
Two-Terminal Low-Loss Standard Capacitors (52210S-52211S)
Three-Terminal Low-Loss Standard Capacitors (52221C)
Two-Terminal, High-Q Standard Inductors (52310S)

Power-Frequency Capacitors
Power-Frequency Capacitors (52400C)
Q-Standards
Q-Standards (52710C-52711C)

Voltage Measurements
DC Voltage Measurements and Standards
General Information— DC Voltage Measurement Standards
Special DC Voltage Measurements, by Prearrangement (53110S)
Saturated Standard Cells (53130C-53140C)
Unsaturated Standard Cells (53150C)
Solid-State Voltage Reference Standards (53160C-53161C)

AC Voltage Measurements
Digital Multimeters (DMMs) and Multifunction Calibrators (53200S)
Low-Voltage AC-DC Transfer Standards (53201S)
Special 25-Point Test of Digital Multimeters (DMMs), by Prearrangement (53202S-53203S)

AC-DC Thermal Voltage and Current Converters (to 1 MHz)
General Information— Thermal Voltage and Current Converters
Special AC-DC Measurement Services, by Prearrangement (53310S)
AC-DC Difference Calibration of a Standard or Standards Set (Voltage or Current) (53350C-53352C)
RF-DC Thermal Voltage and Current Converters (100 Hz-1 GHz)
General Information—RF-DC Thermal Voltage and Current Converters, 100 Hz to 1 GHz (53405S-53445S)
RF Voltage Comparators (53405S)
Thermal Voltage Converters (TVCs) (53410C-53421C)
Peak-to-Peak Detectors (53430S-53431S)
RF Micropotentiometers (53440S-53445S)

Precision Ratio Measurements
Inductive Dividers
Special Ratio Measurements and Tests of Inductive Voltage Dividers, by Prearrangement (54110S)
Inductive Voltage Dividers (54120C-54131C)

Resistive Dividers
Resistor and Resistive Dividers, DC Measurements (54210C-54211S)
Resistor and Resistive Dividers, 60 Hz Measurements (54212C-54213S)
Resistor and Resistive Dividers, Pulsed High-Voltage Conditions (54214S)

Capacitive Dividers
Capacitive Dividers, 60-Hz Measurements (54310S)
Capacitive Dividers, Pulsed High-Voltage Conditions (54311S)

Mixed Dividers
Mixed Dividers (54410S)

Voltage and Current Transformers
Voltage Transformers (54510C-54513C)
Current Transformers (54520C-54522C)
Special Tests of Dividers and Transformers (54600S)

Phase Meters and Standards and VOR Measurements
Special Tests of Phase Standards and Related Instruments, by Prearrangement (55110S)
Phase Meters (55120C-55141C)

Power and Energy Measurements, Low-Frequency
Special Tests of AC-DC Wattmeters (56110S)
Power and Energy Measurements, Low-Frequency (56200C-56202C)
Measurement Assurance Program for Watthour Meters (56210M)
Fast Turn-Around Energy Measurements, Low-Frequency (56220S)

RF, Microwave and Millimeter-Wave Measurements
Thermistor Detectors
General Information
Commercial Coaxial Thermistor Detectors (61110S-61136S )
NIST Model CN Reference Standard (61137C-61138C)
Waveguide Thermistor Detectors (61144S-61155S)
High-Power Wattmeter (61160S)

Scattering Parameters of Passive Multi-Port Devices
General Information
Coaxial Fixed and Variable Attenuators (61210S-61221S)
Rectangular Waveguide Fixed and Variable Attenuators (61230S-61249S)
Time Delay, Coaxial and Waveguide (61250S)
Coaxial One-Port Devices (61260S-61271S)
Waveguide One-Port Devices (61280S-61294S)
Phase Shifters, RF and Microwave (61295S-61297S)

High-Accuracy Attenuation Measurements
Coaxial Fixed and Variable Attenuators (61310C)
Waveguide-Below-Cutoff (Piston) Attenuator Measurements at 30 MHz (61320S)
Attenuation Measurements at 1.25 MHz (61330S)
Phase Shifters (61350C))

Thermal Noise Measurements
Noise Temperature Measurements (61410S-61465S)
Special Noise Temperature Measurements (61495S)

Dimensional Verification of Coaxial Air Line Standards
Coaxial Air Lines (61510S)
Microwave Dielectric and Magnetic Material Measurements
Special Tests for Dielectric and Magnetic Materials (61620S)
Special Consulting and Advisory Services for Dielectric and Magnetic Materials, by Prearrangement 61640S)

Electromagnetic Field Strength and Antenna Measurements
Microwave Antenna Parameter Measurements
General Information—Antenna Parameter Measurements (Microwave)
Gain and Polarization Calibrations of Standard Antennas Using Extrapolation Range (63100S)
Measurement of Pattern, Gain, and Polarization of Arbitrary Antennas Using Near-Field Scanning Techniques (63200S)
Special Test Service for Calibrating Probes Used with Near-Field Scanning Facilities (63300S)
Special Consulting, Advisory, and Other Services (63400S)
Field Strength Parameter Measurements
Antennas/Field Strength Measurements, Utilizing the Transverse Electromagnetic (TEM) Cell Method (64100S)
Antennas/Field Strength Measurements, Utilizing the Open Area Test Site and Standard Antenna Method (64200S)
Antennas/Field Strength Measurements, Utilizing the Anechoic Chamber and Standard Field Method (64300S)

Pulse Waveform Measurement
General Information—Pulse Waveform Measurements
Impluse Spectrum Amplitude (65100S)
Fast Repetitive Pulse Transition Parameters, 50 (65200S)
Repetitive Pulse Waveform Measurements, Including Settling Parameters (65250S)
Network Impulse Response (65300S)
Pulse Time Delay Interval (65400S)

Calibration Services, NIST, 100 Bureau Drive, Stop 2330, Gaithersburg, MD 20899-2330
Telephone: 301-975-2092, Fax: 301-869-3548, E-Mail: [email protected]

Date created: 06/30/1999
Last updated: 06/16/2003