Three-dimensional (3D) Ultrasound (US) can facilitate diagnosis, treatment planning, and image-guided therapy. However, current studies rarely provide a comprehensive evaluation of volumetric accuracy and reproducibility, highlighting the need for robust Quality Assurance (QA) frameworks, particularly for tracked 3D US reconstruction using freehand or robotic acquisition. This study presents a QA framework for 3D US reconstruction and a flexible open source platform for tracked US research. A custom phantom containing geometric inclusions with varying symmetry properties enables straightforward evaluation of optical, electromagnetic, and robotic kinematic tracking for 3D US at different scanning speeds and insonation angles. A standardised pipeline performs real-time segmentation and 3D reconstruction of geometric targets (DSC = 0.97, FPS = 46) without GPU acceleration, followed by automated registration and comparison with ground-truth geometries. Applying this framework showed that our robotic 3D US achieves state-of-the-art reconstruction performance (DSC-3D = 0.94 +- 0.01, HD95 = 1.17 +- 0.12), approaching the spatial resolution limit imposed by the transducer. This work establishes a flexible experimental platform and a reproducible validation methodology for 3D US reconstruction. The proposed framework enables robust cross-platform comparisons and improved reporting practices, supporting the safe and effective clinical translation of 3D ultrasound in diagnostic and image-guided therapy applications.