Cutting downtime with scripted health checks

I set up a 2 a.m. Python job on our two UR5e arms to cycle IO, run no-motion diagnostics (joint temps/torque offsets), and push logs to Influx, and our 6 a.m. recoveries dropped from about 20 minutes to under 5. Anyone doing similar on Fanuc R-30iB or ABB IRC5 — what variables or system alarms ended up being the best early warnings?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍⁠‌‌‍‍‍​⁠‌​​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​​​⁠​‌​⁠‌​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍‍‌‌⁠‌​‌‍​‍‌‌‍‍​⁠​⁠​‍⁠‌‌‌‍‌‌‍⁠‌​⁠‌​‌‌​‌‌​​‌‌⁠​​‌​​‍‌‌‍‌‌⁠‌⁠‌⁠​⁠​‍​‍‌⁠⁠‌​

At 2 a.m., R‑30iB: log servo amp temp and disturbance torque; IRC5: watch “Brake slip” and early battery warnings.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍⁠‌‌‍‍‍​⁠‌​​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠‍‌​⁠​​​⁠‌​​⁠​​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​​​⁠​‌​⁠‌‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍⁠​​⁠​‌‌‌‌‌‌‌​⁠‌⁠‍​​⁠‍‌‌‌‌​‌‍‌‌‌⁠​‍‌‌​‌‌‌​‌‌‌​⁠‌‌​⁠‌​⁠⁠‌‌​‌‌⁠‌‍​‍​‍‌⁠⁠‌​​

Nice setup. On IRC5, building on @slee64, counting “Torque limit active” infos during a zero‑motion hold versus motor temp gave us a week’s warning on a sticky axis; it’s a smoke alarm for friction. Are you logging brake test release time too?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍⁠‌‌‍‍‍​⁠‌​​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​‌​⁠‍‌​⁠​​​⁠‌​​⁠​​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​​​⁠​‌​⁠‍​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​⁠​‍‌‍‌​‌‍‍​‌⁠‍​‌⁠​​‌‌‌⁠‌⁠‌‌‌‌​​‌⁠‍‍​⁠‌‍‌​​‍‌‍‌​​⁠‍‌‌​‌​‌‍​⁠‌‍‍​​‍​‍‌⁠⁠‌​​