Document 364-1

Calibration, Compensation and Correlation

Introduction A typical calibration process involves connecting several When comparing published inductance values, engineers standards to the instrument. The calibration require a common basis for comparison. Ideally, a 1 uH standards should have values that cover the entire mea- inductor from one manufacturer would be equivalent to a surement range of inductance. The instrument is adjusted 1 uH inductor from every other manufacturer. For that to so that it reads the labeled value of the . The be true, all manufacturers would have to use the same instrument is considered calibrated when it displays the standards to design and measure their inductors. correct value of each standard. The National Institute of Standards and Technology (NIST) Calibration Plane provides very limited standards for inductors. The stan- The calibration plane is where measurement accuracy is dards require coaxial connections, styles are limited to defined using the calibration standards. Ideally, it is located air core or non-saturable core inductors, the inductance at the instrument terminals where the inductor is connected value range is 10 nH to 100 mH, and the frequency range for measurement. is limited from 10 kHz to 250 MHz. These standards are not suitable for comparing modern chip inductors and In order to provide a durable and repeatable method of power inductors with their wire lead or surface mount (SM) connecting calibration standards to the instrument, stan- constructions. dard connectors — such as type N, 7 mm, or 3.5 mm threaded connectors — are typically used. Wire lead and As surface mount inductors were developed, new instru- surface mount inductors do not have threaded terminals, ments and fixtures were developed to measure them. As therefore a test fixture is required to connect the part to the instrumentation and fixture technology improved, the ac- instrument. curacy of improved. The relatively rapid change in measurement technology resulted in the manu- Fixture Compensation facture of many different fixture designs. The instrumenta- The Fixture Compensation Plane, shown in Figure 1, is tion and fixtures used are generally based on the best where the inductor is connected to the test fixture. The test available technology at the time for the inductor type, size, fixture extends the measurement plane a physical dis- inductance value, and test frequency. tance from the calibration plane. The test fixture also adds Since inductance standards have not kept pace with in- parasitic resistance, inductance, and capacitance to the ductor technology, each manufacturer is responsible for measurement. Fixture compensation “extends” the cali- the (repeatability) of its inductor bration plane to the test fixture plane and reduces the measurements. The following discussion explains how effects of fixture parasitics. Fixture compensation is similar Coilcraft uses calibration, fixture compensation, and cor- to calibration in that standards of known value are used to relation (where appropriate), to ensure the accuracy and set the instrument to read the correct value of the standard precision of its measurements. at the fixture measurement plane. The Importance of Calibration Fixture compensation plane Accuracy statements assume that a standard of the “true” Instrument value exists to which a measurement can be compared. Calibration plane In the absence of an available standard, a defacto (gen- erally accepted) standard may be used. Defacto calibra- Figure 1 tion standards are typically supplied by the instrument manufacturer. The accuracy of the standards is based on physical principles and measurements traceable to na- Calibration Standards tional standards. Coilcraft performs an instrument calibration before every Calibration sets the instrument to an established mea- measurement session. The selection of calibration stan- surement condition (or state) using physical standards. dards depends on the instrument to be calibrated and the Ideally, after each calibration session, the instrument is set required accuracy. The accuracy of calibration standards to the same standard conditions, resulting in an estab- has increased with improvements in impedance analyzer lished and repeatable measurement condition. technology. The extremes of impedance are: a short circuit

Specifications subject to change without notice. Document 364-1 Revised 09/16/03 Document 364-2

(lowest impedance), and an open circuit (highest imped- ments in instrumentation and fixture technology lead to ance). Most impedance analyzers and network analyzers more accurate measurements, but still can result in differ- that are used to measure inductance have both open and ent measured values of the same component. short calibration standards. Some 50 Ohm (characteristic Test equipment used in the past to design and test our impedance) analyzers also have a 50 Ohm load calibra- inductors required special fixtures to accommodate smaller tion standard. components, particularly surface mount devices. The in- An ideal inductor causes a 90° phase angle between the accuracies of the original test equipment and fixture be- voltage across and the current through the inductor. Accu- came part of the component’s measurement. As a result, rate quality factor (Q) measurements depend on an accu- a part that measured 100 nH on one instrument and fix- rate measurement of the phase angle. A very small error in ture, may measure 95 nH on a different instrument and the phase angle measurement can lead to a large error in fixture. To overcome variations due to different instruments the displayed Q value. Typical open and 50 Ohm load cali- and fixtures Coilcraft uses a correlation process. Correla- bration standards do not have perfect 0° phase character- tion ensures that all manufactured inductor values are istics. Calibrating with an accurate phase calibration consistent with their original design value. standard, such as a low loss air capacitor, can reduce the This process uses standards (correlation pieces) that are phase angle measurement error and improve Q measure- equivalent in size and electrical character to the actual ments. Air capacitor standards are available for use with production parts. These correlation pieces were created, many higher-end impedance analyzers. measured, and labeled (“tagged”) with the measured value Fixture Compensation Standards at the time of design. Measurements of inductors with the Typical fixture compensation includes open and short com- same part number are compared to measurements of the pensation. When measuring series inductance and Q, the correlation standards. The product measurements are open standard is the air gap between the test terminals compensated for the difference between the new mea- with no part connected. The short standard is a very low surement of the correlation piece and that of its tagged impedance (high conductivity) metal bar or wire with di- value. The compensated measurements are then com- mensions approximately the same as the inductor to be pared to specification, providing consistent verification re- measured. Coilcraft uses high quality gold-plated shorting quired by the customer. bars for the short standard to ensure high conductivity and We continue to manufacture our surface mount chip in- good contact with the test terminals. ductors to their original correlation standards, ensuring Some instruments allow a load compensation to be per- that the electrical parameters are consistent over time. By formed. A standard 50 Ohm or 75 Ohm surface mount or manufacturing to a correlation standard, we ensure that wire lead resistor (load) may be used, but real resistors an inductor we produce today or in the future will have the always have some parasitic inductance and capacitance same value as the original design. that add error to the measurement. Combining two resis- tors of double the fixture compensation load value in par- Correlation Is Not Always Required allel cuts the parasitic inductance in half, but may more On low inductance value components with tight tolerances, than double the parasitic capacitance. In most cases, since such as chip inductors, correlation is used for tolerance the parasitic effects of a fixture compensation load stan- control. The large inductance values of power inductors dard are not known, load compensation is not performed. require the use of ferrite or powdered iron core material. The permeability tolerance of ferrites and powered iron Correlation of Chip Inductor Values core materials is typically 20%, although tighter tolerances It is often assumed that after instrument calibration and can be requested at a premium. The difference in induc- fixture compensation, the value displayed by an instru- tance measurements between different instruments is typi- ment is “the” correct or true or absolute value. This re- cally very small compared to the 20% tolerance range of quires that all combinations of instrument, fixture, calibration power inductors. Therefore, correlation is not necessary standards, and fixture compensation standards yield the for power inductors. same value when measuring the same inductor. Improve-

Specifications subject to change without notice. Document 364-2 Revised 10/20/03