A differential could work, but you’ll have to pin down the exact load distribution first. The key is to split the torque evenly so the teeth never fight each other. If you can get a bevel‑gear pair that locks only when the input is overloaded, you’ll keep the speed constant while letting the slip mechanism absorb the excess torque. Just remember to size the pinion teeth for the high‑speed side and the ring gear for the low‑speed side—any mismatch and you’ll introduce more friction than you’re trying to cut. Also, keep an eye on the lubrication path; a clogged line will turn that elegant design into a sticky mess. Try a small test run and watch the wear pattern; if you see the outer gear starting to flatten, tweak the gear ratio or add a roller‑bearer to spread the load. It’s a bit of trial‑and‑error, but that’s the fun part.

Sounds like a solid plan, but I’d double‑check the helix angles before you tighten the tolerances. Even a 0.5° mismatch can throw off the contact ratio and cause early flank wear. Also, make sure the lubricant path follows the actual force vector; if it runs perpendicular to the thrust, you’ll get a sludge build‑up. Keep a micrometer record of every dimension, and when you test, log the torque ripple at 10k rpm—those tiny oscillations can tell you if the differential is truly balanced. Once the data’s in, tweak in 0.1% steps; that’s where the real breakthroughs happen. Good luck, and keep the test sheet handy—failure logs are as good as success reports for me.

Sounds like a rock‑solid workflow. Just remember to run a quick test at 2k rpm first—low speeds catch alignment errors before you hit the 10k mark. Good luck, and keep that micrometer in one hand, the torque meter in the other.