Traceability of Vector Network Analyzers

About this Article
Written by Ken Wong of Agilent's Santa Rosa manufacturing division, this article originally appeared in Test & Measurement Europe magazine, February-March 1996. Copyright 1996 Reed Elsevier, Inc. and used with permission.

Introduction

Measurement traceability is becoming an international requirement with the world wide acceptance of ISO quality standards. The issue of traceability, however, is often confusing, especially for microwave network analyzers. There are many definitions of traceability. It is not obvious how the measurement parameters are linked to the basic fundamental constants such as frequency, volt, length etc. The objective of this paper is to clarify these issues.

Traceability must be defined in conjunction with calibration, verification, and the associated uncertainty of a measurement system. The internationally accepted definitions of these terms is documented in ISO Guide 25, General Requirements for the Competence of Calibration and Testing Laboratories. The intent of paragraph 9 of ISO Guide 25, "measurement traceability and calibration", can be summarized as follows: "Traceability to designated standards (national, international, or well characterized reference standards based upon fundamental constants of nature) is an attribute of some measurements. Measurements have traceability to the designated standards if and only if scientifically rigorous evidence is produced on a continuing basis to show that the measurement process is producing measurement results for which the total measurement uncertainty relative to national or other designated standards is quantified."¹

VNA Calibration

Vector network analyzers (VNAs) measure the relationship between incident waves and reflected and transmitted waves. Measured quantities, such as coefficient of reflection, return loss, insertion loss, forward gain, and S-parameters are referenced to the defined impedance of the measurement system.

To minimize measurement uncertainties, VNAs are calibrated using a set of "known " impedance calibration standards such as opens, shorts and loads. The measurement uncertainty of the calibrated VNA system is directly proportional to the uncertainty of the known quantities of these standards. The impact of each standard's uncertainty depends on the calibration method used, stability of the measurement system, and uncorrected performance of the system as illustrated in Figure 1.

Figure 1 -- Propagation of calibration errors

Uncertainty of calibration standards

The key factors that affect the accuracy of RF/microwave calibration standards are connector repeatability, dimensional accuracy, and electrical characteristic accuracy. The most accurate microwave calibration standards are metrology grade precision transmission lines and shorts. (Metrology grade connectors are known also as laboratory precision connectors, LPC.) The impedance and other microwave characteristics of these devices are calculable from their physical dimensions and physical properties of the material used.²  The uncertainty of these devices is a function of the physical metrology systems.

Metrology grade devices are fabricated to the tightest specifications possible using state-of-the-art equipment and methodology. At HP, precision coaxial airlines can be fabricated with a diameter tolerance of <2µm and diameter consistency of <1.5µm. This is equivalent to an impedance accuracy of 0.2% for 7 mm devices and 0.6% for 2.4 mm devices.

For coaxial devices, the female contact must be slotless to minimize performance dependency on mating devices, a prerequisite of metrology, grade coaxial connectors.³  Metrology grade standards reduce calibration uncertainties by a least 50% over instrument grade standards. HP is the only source, worldwide, for microwave metrology grade coaxial connectors. The electrical characteristics of other calibration devices can be derived from these physically calculable standards. These derived electrical characteristics are then fitted to a model using calibration constants.

HP's calibration standards are machined to very tight tolerances, <2.5µm for critical dimensions, and thus are interchangeable. It is not necessary to keep track of which open or short to match to which set of calibration constants or calibration model. Our calibration constants are derived from very accurate computer modeling and experimental data. The model accounts for skin loss, impact of pin depth, and other imperfections.

Traceability of calibration standards

All HP standards are traceable, through precision airlines and shorts and mechanical gages, to the National Institute of Standards and Technology (NIST) in Gaithersburg, MD (see Figure 2). These are the most precise standards available for network analyzers. HP's traceable path is the same path that many national laboratories, such as the National Physical Laboratory (NPL) of UK, are using. Figure 2 also compares HP's traceable path with NIST's six-port traceable path.

Figure 2 -- VNA traceability and uncertainty

The measurement uncertainties of all the applicable mechanical measurement systems are vigorously quantified using well established statistical methods. An accurate electrical model, based on accepted microwave theory and practices and verified data, is created to convert the mechanical/physical properties into microwave S-parameters. The uncertainties also are allowed to propagate through the S-parameter calculations and accounted for in the uncertainties of the S-parameters, as diagrammed in Figure 3.

Figure 3 -- Uncertainty propagation

The modeled S-parameters are then used to calibrate HP's VNA measurement system. This calibration method removes any imperfections that the airlines and shorts may have, and provides the link to the fundamental units of measure length and other physical constants.

VNA verification

To ensure that the VNAs measurements are accurate as specified after calibration, the measurement system must be verified. An internationally accepted method of VNA verification uses a set of verification standards, such as a low loss match airline, a mismatched device and two lossy devices. These devices are first measured by a reference laboratory. The measured data from the VNA in question is compared with the data obtained from the reference laboratory. If the two measurements are statistically similar, then the system is considered verified. This method is also known as "tops down" verification. Important considerations for selecting verification standards are that :

• the devices must be physically different from the calibration devices; and
• the devices cover an adequate range of reflection and transmission characteristics .

The "tops down" verification method works best when the measurement uncertainty of the reference laboratory is much smaller than the measurement uncertainties of the VNA system in question. Unfortunately, the best verification standards and the best calibration standards are at the same accuracy level. Therefore, it is not possible to verify a VNA system's accuracy by measuring a set of verification devices using the traditional tops down method. In addition to performing a tops down verification, statistical data of the verification results should be kept to establish statistical system performance limits. Participation in national and international "measurement comparison programs" will provide added assurance of the VNA's performance and proof of traceability.

A "bottoms up" verification approach has been proposed and used by many calibration laboratories. This approach uses precision airlines, and other physical standards in a different calibration method to extract the "residual errors" of the original calibration. Again, since the original calibration used similarly derived standards, measurement comparisons of interlaboratory standards are still needed. Other error sources of the measurement system are also individually quantified at the most fundamental level. This method provides a more thorough treatment of the VNA system verification, but is much more time consuming and costly.

Conclusion

VNA systems are traceable through calibration and verification standards. The quality and accuracy of these standards determine the achievable accuracy of a VNA . No calibration method can overcome this fundamental limitation. HP provides the highest grade coaxial and waveguide calibration and verification standards today. The manufacturing process, metrology systems and device modeling methodology used to ensure the integrity and accuracy of HP's standards are closely and continuously monitored and improved so that our VNA systems' measurement accuracy are solidly supported.

References

1. Belanger, B.C., "Traceability: An Evolving Concept", ASTM Standardization News , Vol 8, No. 1 1980, pp22-28.
2. Wong, K.H., "Characterization of Calibration Standards by Physical Measurements", ARFTG Digest Spring 1992.
3. IEEE Standard 287, "Precision Coaxial Connectors", (draft version).