Abstract In this paper, we present an improved version of the unified method (UM) for thermo-electric modules (TEM) characterization with the aim to reduce the overall parameters uncertainty. The parameters that make up the commonly accepted figure of merit Z or z T ¯ are mainly affected by temperature related uncertainty, which also significantly limit the precision of the test conditions and thus the reliability of derived models. To overcome these limitations, we conceived a novel and simplified measurement setup that relies at the same time on thermistor technology rather than thermocouple and on a lower number of required components. The improved setup, exploits a solid calibration procedure to reduce the uncertainty in temperature measurements from 1.4 °C typical of a J-type thermocouple (Class 1 IEC-EN 60584-2) to 0.027 °C with calibrated thermistors, leading to a previously unachievable millidegree-precision temperature control and to an uncertainty drop of two orders of magnitude in all the temperature-derived measurements. Such improvement has a direct feedback on both the α S and Θ , leading to a drop of a least one order of magnitude for a ΔT of 3 °C and more than 44 times for the latter at 30 °C. As result, the figure of merit z T ¯ can now be determined with an uncertainty equal to 0.58%.

中文翻译：

---

用于热电模块表征的高精度测试台

摘要 在本文中，我们提出了一种用于热电模块 (TEM) 表征的统一方法 (UM) 的改进版本，旨在降低整体参数的不确定性。构成普遍接受的品质因数 Z 或 z T ¯ 的参数主要受温度相关不确定性的影响，这也显着限制了测试条件的精度，从而限制了派生模型的可靠性。为了克服这些限制，我们构思了一种新颖且简化的测量设置，该设置同时依赖于热敏电阻技术而非热电偶，并且所需组件数量较少。改进的设置利用可靠的校准程序将温度测量的不确定性从 J 型热电偶（1 类 IEC-EN 60584-2）的典型 1.4 °C 降低到 0。027 °C 使用校准的热敏电阻，导致以前无法实现的毫度精度温度控制，并使所有温度衍生测量的不确定性下降两个数量级。这种改进对 α S 和 Θ 都有直接反馈，导致 3°C 的 ΔT 下降至少一个数量级，而后者在 30°C 时下降超过 44 倍。因此，品质因数 z T 现在可以以等于 0.58% 的不确定性来确定。