That’s a tall order for a hand‑built thing. First, get a stable base—an aluminum plate that won’t flex, bolted to a concrete slab. Then use a fine‑tuned optical bench: a beam splitter, a high‑resolution CCD, and a laser with an ultra‑stable wavelength. Mount the whole assembly on a vibration‑isolated table, and wrap the room in sound‑proofing so air currents don’t shift the optics. For the actual tracking, you’ll need a piezo‑actuated mirror that can move in micro‑nanometer steps, controlled by a feedback loop from the CCD. All the wiring has to be shielded, and you’ll have to keep the whole system in a temperature‑controlled chamber. It’s a lot of small parts, but if each one is built to tight tolerances, the whole thing can resolve a star’s wobble. The trick is to keep the design as simple as possible; the fewer moving parts, the less drift you’ll get. And remember, a lot of that precision comes from how well you isolate the whole setup from the world, not just the hardware.

You’re right, the sky isn’t a quiet room. I’d add a reference star in the same field, so the photometer can see the same scintillation and subtract it. Locking the laser to an atomic clock is a good call—keep the wavelength steady to the picometer. And yeah, every extra mirror can twist the phase, so I’d use as few optics as possible, maybe a single high‑quality beam splitter and a compact imaging system. The trick is to build a tight feedback loop that can keep the whole thing in sync with the atomic standard while still rejecting the cosmic hiss. It’s a balancing act, but that’s what makes it interesting.