Feature: RF and microwave


Figure 4: Typical mmWave test bench


angular miss to 0.5° and the corresponding shortfall to about 0.12dB. Beam state. Te device codebook entry, beam-management behaviour and calibration state at the time of measurement are part of the result. For OTA measurements, the correlation


plan should specify common geometry definitions, correction schemes, alignment and beam-search procedures, and an inter-range comparison using a suitable characterised antenna before DUT results are compared.


Designing for measuring agreement Station disagreements are oſten discussed from a handful of repeated readings. A structured experiment is more useful because it separates the main sources of variation instead of treating every difference as the same problem. If the same units are measured on every


station, the basic unit-by-station design is crossed. Reconnection trials can then be nested within each unit-by-station combination. To estimate short-term repeatability separately from remating variation, take at least two readings without disturbing the connection, then disconnect and remate the unit and repeat. Analyse the result using variance components, ANOVA or a mixed-effects model, following established gauge R&R practice; do not estimate the connection contribution by subtracting one standard deviation from another. For example, ten units across three


stations, with three independent connections and two readings per


connection, produce 180 readings. Tat design can estimate within-connection repeatability, reconnection variation, station effects, unit effects and relevant interactions. Units should span the operating range, and the sequence should be randomised or blocked so that station warm-up and temperature driſt are not confounded with unit order. Te final analysis should relate the


observed measurement-system variation to the product tolerance and the decision rule. If that variation consumes a material fraction of the specification width, the issue is one of measurement capability rather than a simple station offset. Te acceptable fraction depends on the consequences of false pass and false fail and should be set by the test programme rather than taken from a universal rule. Te following controls are worth


building into the station specification: • Define the reference plane in the specification. State where the measurement is referenced and how the correction reaches that plane.


• Calibrate through the path that is used. Include the switch, fixture or extender in the calibration, or characterise it separately with a controlled correction.


• Version the configuration and establish it explicitly. Apply result-critical settings at the start of every sequence instead of assuming residual instrument state.


• Control the physical path mechanically. Fix cable routing, specify connector torque and adapter stacks, and gauge or inspect connectors on an evidence-based schedule.


20 November 2026 www.electronicsworld.co.uk


• Record metadata and trend station differences. Associate every result with station identity, correction-file version, soſtware version, temperature, and calibration status. A stable station difference can be characterised and, where justified, corrected; a changing difference should trigger investigation.


So, which one is right? When two stations disagree, engineers oſten start by asking which one is right. Before that can be answered, it is necessary to establish whether the stations are measuring the same quantity at the same reference plane and under comparable conditions. A valid comparison requires the relevant


parts of the measurement definition to be common, or their differences to be corrected and validated. At lower frequencies, some physical contributions may be smaller, but agreement is never automatic. In the millimetre-wave bands, path loss, connection repeatability, fixture behaviour and OTA geometry make the consequences of an implicit architecture more visible, while 6G research bands, E-band backhaul and integrated phased arrays are bringing these conditions into routine test work. At millimetre-wave frequencies,


reproducibility depends strongly on the station architecture. A defensible comparison also requires a defined measurand, traceable calibration, controlled instrument state and evidence that the comparison process remains stable.


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44