Non-Contact Reflection-Coefficient Measurement of Automotive Transmission Fluid Degradation with a Hexagonal Patch...

This study formulates and experimentally characterizes a non-contact reflection-coefficient (S₁₁) measurement method for assessing the degradation state of automotive transmission fluid (ATF). The measurand is the reflection coefficient, at the SMA (SubMiniature version A) reference plane, of a 4 × 4 hexagonal patch array antenna whose near field is loaded by a steel-backed measurement cell containing a fluid layer of approximately 9-10 mm; fluid condition is read from the frequency and level of the fluid-loaded resonance notch. The antenna was characterized by full-wave simulation (resonance at 4.852 GHz) and by SOLT (short-open-load-thru)-calibrated bench measurements (4.725 GHz, -24.32 dB), with the far-field pattern measured in an anechoic chamber. Fluid spectra were acquired using a SOLT-calibrated LibreVNA (2-6 GHz, 4 MHz step) within a single session under one calibration in a bracketed sequence in which the fresh fluid was measured both before and after the degraded fluid. Repeated acquisitions (n = 6 fresh; n = 4 degraded, ~200,000 km of service) yield a fresh-fluid notch of -28.6 ± 0.6 dB at 4.755 GHz and a degraded-fluid notch of -43.0 ± 4.3 dB at 4.775 GHz. Decomposition of the measured complex reflection coefficient shows that the resistive part of the loaded impedance at the notch is essentially unchanged (49.9 → 50.5 Ω) whereas the reactive part collapses (+3.7 → +0.4 Ω); the upward frequency shift and the deepening are therefore two consequences of a single reactive mechanism rather than two independent signatures. The two specimens were separated at p = 0.0095 (level) and p = 0.0080 (frequency) by the Mann-Whitney U test against a bracketed session drift of -1.3 MHz and +0.31 dB. Type A repeatability results and Type B considerations are reported following the GUM/VIM. These results establish a metrologically characterized, low-cost, non-invasive proof of concept for ATF condition monitoring, validated here on a single fresh/degraded sample pair.