Live Garage
This is my own validation car, running on the real T800 pipeline every morning. Not a mockup, not a demo — real CAN telemetry, with customer-facing monitors and validation-only monitors shown side by side.
259-Day History
The cards above are today. This is the run behind them: the same four systems, decoded from this car's own CAN bus every day since Jan 1, 2026, with no gaps. A used Tesla's real condition is a trend, not a test-drive — a single reading can't tell a healthy pack from a failing one, but 259 of them can. Data delayed — showing the last verified run.
Average brick CAC in amp-hours, one point per day: the mean computed capacity across brick groups (the weakest group and imbalance are tracked separately). The axis is zoomed to the data, spanning about 5 Ah, so the shape is legible; the decline is real but gentle. Capacity is one piece of evidence, read with cell balance, cycles and temperature — never a standalone "battery life %".
Pack-to-chassis isolation. The daily median sits at the reported ceiling of 5.16 MOhm essentially every day; the daily minimum swings through a healthy band. Across the whole run, exactly three single frames came back near zero — junk frames, not faults. A monitor that alarmed on the worst sample would have failed this pack three times; t800 gates on persistence, so it raised zero alerts.
Ride-height drift while parked, in mm per hour, common-mode removed so it isolates one corner sagging rather than the whole car settling with temperature. Day to day it is noisy; the solid line is the 7-day median. A developing air-strut or valve leak shows up as that trend rising on one corner and staying up, long before a dashboard warning. So far it has not.
DC-DC converter efficiency — how well the car tops up its 12V from the HV pack. This measures the conversion path that supports the 12V system, not the resting health of the 12V battery itself. A visible step-down begins in spring and holds; we are monitoring it, and this chart alone does not identify the cause. No validated resting-voltage series is shown here — t800 does not pass off awake charging voltage as a true sleep-state reading.
Full explanations, with the method behind each: battery capacity vs mileage · isolation false alarms · air-suspension leak detection · 12V warning signs.