Semi-device independent (Semi-DI) quantum random number generators (QRNGs) gained attention for security applications, offering an excellent trade-off between security and generation rate. This paper presents a proof-of-principle time-bin encoding semi-DI QRNG experiments based on a prepare-and-measure scheme. The protocol requires two simple assumptions and a measurable condition: an upper-bound on the prepared pulses’ energy. We lower-bound the conditional min-entropy from the energy-bound and the input–output correlation, determining the amount of genuine randomness that can be certified. Moreover, we present a generalized optimization problem for bounding the min-entropy in the case of multiple input and outcomes, in the form of a semidefinite program. The protocol is tested with a simple experimental setup, capable of realizing two configurations for the ternary time-bin encoding scheme. The experimental setup is easy-to-implement and comprises commercially available off-the-shelf components at the telecom wavelength, granting a secure and certifiable entropy source. The combination of ease-of-implementation, scalability, high security level and output-entropy, make our system a promising candidate for commercial QRNGs.

半设备独立 (Semi-DI) 量子随机数发生器 (QRNG) 获得了安全应用的关注，在安全性和生成率之间提供了极好的权衡。本文提出了一种基于准备和测量方案的原理验证时间段编码半 DI QRNG 实验。该协议需要两个简单的假设和一个可测量的条件：准备好的脉冲能量的上限。我们从能量界限和输入输出相关性中降低了条件最小熵，确定了可以证明的真正随机性的数量。此外，我们以半定程序的形式提出了一个广义优化问题，用于在多输入和多结果的情况下限制最小熵。该协议通过一个简单的实验设置进行测试，能够实现三元时间仓编码方案的两种配置。实验装置易于实施，包括电信波长的商用现成组件，提供安全且可认证的熵源。易于实施、可扩展性、高安全级别和输出熵的结合，使我们的系统成为商业 QRNG 的有希望的候选者。