Session: Session 3 | Calibration Techniques |
Chair: | Dave Blackham, Agilent Technologies, Inc |
Abstract: | In this session we present a range of calibration techniques. These include a method directed towards calibration of large-signal nonlinear measurement systems; a new technique for calibration of the Vector Signal Analyzer, an instrument frequently used for modulated signal measurement; a novel method of making calibrated mixer measurements; a submillimeter-wave power measurement standard; and a paper discussing uncertainties in VNA measurements. |
| | Session 3-1 | A multi-step phase calibration procedure for closely spaced multi-tone signals |
1:20 PM-1:40 PM | M. Mirra1, M. Marchetti1, F. Tessitore1, M. Spirito1, L. C. de Vreede1, L. Betts2, 1TU Delft, Delft, Netherlands, 2Agilent technology, Santa Rosa, United States |
(45) | A phase calibration procedure suitable for closely spaced multi-tone signals is presented. The proposed multi-step calibration approach can provide high accuracy for complex modulated signals (e.g. WCDMA) and is compatible with a PNA-X platform, which uses hardware extensions for the measurement of the baseband frequencies. |
| |
Session 3-2 | Traceable calibration of Vector Signal Analyzers |
1:40 PM-2:00 PM | D. A. Humphreys, M. R. Harper, M. Salter, National Physical Laboratory, Teddington, United Kingdom |
(49) | We have developed a traceable calibration technique for Vector Signal Analysers (VSA). A multi-sine signal generator provides detailed magnitude and phase stimulus over the instrument bandwidth. A Digital Sampling Oscilloscope (DSO) measures the multi-sine waveform. Traceability is provided via the DSO, which can be traceably calibrated to the primary standard (Electro-Optic Sampler). The VSA can be characterised over its full-bandwidth (magnitude and phase) and this technique can be implemented using commercial equipment. |
| |
Session 3-3 | A Novel Method for Measuring Phase and Group Delay of Mixers Without a Reference Mixer |
2:00 PM-2:20 PM | J. P. Dunsmore, J. Ericsson, Agilent Technologies, Santa Rosa, United States |
(24) | Making phase and group delay response measurements on mixers and frequency converters has traditionally been a difficult and complicated process, requiring either modulated-signals (and demodulation), two-tone signals (a special form of modulation) or a system with a reference or re-converting mixer. This paper presents a method to measure the phase response of a mixer on a VNA using a novel phase-measurement method which can measure the input and output phase response independently, computing the transfer response directly. This method relies on a VNA system which utilizes swept-frequency phase-coherent sources and receivers to maintain a common phase reference. A software phase-locking method allows it to operate on frequency converters which have an in-accessible embedded LO, along with a well established method of calibration. |
| |
Session 3-4 | A new 75-110 GHz primary power standard with reduced thermal mass |
2:20 PM-2:40 PM | D. Adamson, J. Miall, J. Howes, M. Harper, R. Thompson, National Physical Laboratory, Teddington, United Kingdom |
(14) | Primary power standards are required to provide traceability for measurements. NPL has recently built a new primary power standard for use in the 75 GHz to 110 GHz band. Compared to NPL’s older standard (which has been used for several decades) the new system has a reduced thermal mass and improved thermal isolation. This gives slightly reduced overall uncertainties and an improvement in usability. Improvements in manufacturing technique offer the promise of using a similar design in the 110 GHz to 170 GHz band and work on a calorimeter for this band has begun. |
| |
Session 3-5 | Some effects of error term interpolation on network analyzer uncertainties |
2:40 PM-3:00 PM | J. Martens, Anritsu Company, Morgan Hill, United States |
(30) | Practically, many network analyzer measurements (both large signal and small signal) are performed at frequencies other than those where the calibration was performed in order to save time and effort. Typically, the error terms are simply interpolated in some way to calibrate at the new frequencies. Under what circumstances and how this interpolation is done can significantly affect the uncertainties of the measurement. This paper explores the characteristics of this interpolative error and how its impact can sometimes be assessed. Of particular note is that the hardware dependence of this effect is considerably larger than for many other components of uncertainty. |
| |
