| 摘要: | 在當前無線與有線通訊快速演進的趨勢下，高頻寬、高功率、低功耗與高速傳輸日益受到重視，而「電磁相容性（ EMC）」與「高速訊號／電源完整性（ SI/PI）」遂成產品設計與驗證的核心。混響室因具備統計均勻場特性，已成為常用的電磁兼容量測環境，但其量測時間長，如何兼顧效率與精確度仍是重要議題。
為提升量測效率，本研究比較「模式調諧」與「模式攪拌」兩種攪拌方式對混響室場均勻性的影響。實驗採用 0.53 m × 0.60 m × 0.91 m 之小型混響室，搭配兩具不同結構與轉速的攪拌器，並以修正 Weyl 公式估算最低可用頻率約 900 MHz。結果顯示，兩種方法皆符合 IEC 61000‑4‑21 場均勻性要求；相較之下，模式攪拌在每頻點取 100 筆樣本僅耗時 4 小時，即能達到與模式調諧（ 30 筆樣本、 15 小時）等同的準確度，效率約提升四倍。
在混響室量測中，電磁場必須滿足隨機統計模型，也就是功率呈指數分布、電場強度呈瑞利分布，這樣才能確保場均勻性與吸收截面等指標的物理意義。若量測數據偏離理論分布，則後續評估與校正皆可能失準。 為驗證本研究模式攪拌量測資料之統計有效性，並探討分布符合程度與場均勻性的關聯，本文針對每一頻率點（樣本數 N = 100，α = 0.05）執行科摩哥洛夫-史密諾夫檢定(K-S)檢定，並以分位數– 分位數（ Q-Q）圖作視覺比對；接著計算各頻點的 K-S 檢定其未通過率並與 IEC 61000-4-21 定義的場均勻性標準差進行皮爾森相關係數分析。結果顯示：在 4GHz 以上頻段， K-S 檢測通過率高於 80%；4 GHz 以下僅約 0– 60%，且隨頻率逼近最低可用頻率而下降。進一步相關性分析得到 r ≈ 0.8，呈顯著正相關，顯示量測場分布越接近理論模型，其場均勻性表現亦越佳。
為了量化吸波材料在隨機電磁場中的能量耗散能力，本文採用吸收截面做為評估指標。吸收截面定義為吸波體在均勻隨機場中所等效吸收的等效面積(單位： m2)；數值越大，表示材料可吸收的平均功率越高。混響室具有統計均勻場特性，可透過比較空室平均功率與放入被測物後平均功率來推算其吸收截面。由於吸收截面不僅反映材料本身的損耗，也間接代表其對系統雜訊與散射能力的抑制潛力。 故本研究以此指標評估三種頻率選擇表面吸波體，並驗證其對筆記型電腦平台雜訊的抑制效果。 在 2–18 GHz 頻段，高電阻吸波體吸收截面約 0.002–0.004 m² (幾何截面積為 144cm2)，低電阻型約 < 0.003 m² (幾何截面積為 144cm2)，商用吸波體約 0.001–0.006 m² (幾何截面積為 361cm2)；實驗結果顯示， 平台雜訊可降低 5–10 dBm，其中以商用吸波體表現最佳；綜合而言， 吸收截面可作為衡量吸波效能與雜訊抑制能力的重要指標。
最後， 本研究利用頻域與時域吸收截面分析等效性的推導結果，提出一套免校正天線以及不需要參考天線的天線效率即可估算天線效率的方法。將所得韋瓦第與微帶貼片天線效率與無反射室量測結果相比， 2.3–9.7 GHz 之誤差不超過 20%， 10.3–17.7 GHz 約為20%，彰顯混響室量測兼具精度與時間效益。;In the rapidly evolving landscape of wireless and wired communications, wider bandwidths,higher power levels, lower power consumption, and faster data rates have elevated electromagnetic compatibility (EMC) and signal/power integrity (SI/PI) to core requirements in product design and compliance. Reverberation chambers (RCs) provide statistically uniform fields for EMC testing but suffer from long test times, making the balance between efficiency and accuracy a critical concern.
This study compares mode tuning and mode stirring for improving field uniformity in a compact RC measuring 0.53 m × 0.60 m × 0.91 m. Two stirrers with different geometries and rotational speeds are employed, and a modified Weyl formula places the chamber’s lowest usable frequency (LUF) at roughly 900 MHz. Both stirring techniques satisfy the IEC 61000‑4‑21 field‑uniformity criterion; however, mode stirring acquires 100 samples per frequency point in only 4 h, whereas mode tuning needs 15 h for 30 samples, so mode stirring achieves comparable accuracy while offering a four-fold gain in efficiency.
In reverberation-chamber (RC) measurements, the electromagnetic field must obey a random statistical model—namely, power follows an exponential distribution and electric-field strength follows a Rayleigh distribution—so that metrics such as field uniformity and absorption cross section (ACS) retain physical meaning. Deviations from these theoretical distributions can invalidate subsequent evaluations and calibrations. To verify the statistical validity of the modestirring data obtained in this study and to explore how distribution conformity relates to field
uniformity, a Kolmogorov–Smirnov (K-S) test was applied to every frequency point (sample size N = 100, α = 0.05), and quantile–quantile (Q-Q) plots were used for visual confirmation. The K-S pass rate at each frequency was then correlated with the field-uniformity standard deviation prescribed in IEC 61000-4-21 via Pearson’s correlation coefficient. Results show that above 4 GHz the K-S pass rate exceeds 80 %, whereas below 4 GHz it is only 0–60 % and declines further as the frequency approaches the lowest usable frequency (LUF). The subsequent analysis yields r ≈ 0.8, indicating a significant positive correlation: the closer the measured field distribution is to the theoretical model, the better the field-uniformity performance.
To quantify the energy-dissipation capability of absorbing materials in a random electromagnetic environment, this work adopts the absorption cross section (ACS) as the evaluation metric. ACS is defined as the equivalent area (m²) that an absorber effectively removes from a uniform random field; a larger value indicates higher average power absorption. Owing to the statistically uniform field in an RC, ACS can be estimated by comparing the chamber-averaged power before and after inserting the device under test. Because ACS reflects not only material losses but also the potential for suppressing system noise and scattering, three frequency selective-surface (FSS) absorbers were evaluated, and their ability to mitigate noise on a notebook-computer platform was verified. Over 2–18 GHz, the high-resistance FSS shows
an ACS of 0.002–0.004 m²(geometric cross-sectional area = 144 cm²), the low-resistance FSS less than 0.003 m² (geometric cross-sectional area = 144 cm²), and the commercial absorber 0.001–0.006 m²(geometric cross-sectional area = 361 cm²). The platform noise was reduced by 5–10 dBm, with the commercial absorber performing best. These results confirm that ACS is a meaningful indicator of both absorption performance and noise-suppression capability.
Finally, by exploiting the equivalence between the frequency- and time-domain formulations of ACS, we propose an antenna-efficiency estimation method that requires neither antenna calibration nor a reference antenna. Applied to a Vivaldi antenna and a microstrip patch antenna, the estimated efficiencies agree with anechoic-chamber measurements to within 20 % over 2.3–9.7 GHz and about 20 % over 10.3–17.7 GHz, demonstrating that RC measurements can deliver both high accuracy and substantial time savings. |
