Anyone can count seconds. The hard part is knowing exactly where the car is, thirty times a second, when the sky is not cooperating and the rear tires have given up.
A stopwatch does not know where the start line is. This does.
Everything downstream depends on this. Lap times, the trace on the map, the sector you were in when the oil temperature spiked. If the position is wrong, all of it is wrong, and the usual answer to a poor signal is to shrug and show you a jumping dot.
The cluster is in a better position than a phone on a windshield mount, because it is wired into the car. It knows the speed of each individual wheel, the angle of the steering rack and the rate the body is rotating, all straight off the bus. That is enough to keep tracking the car with no satellites at all.
Confidence is not decoration. It is a single number combining fix quality, satellite count, dilution of precision, how fast fixes are arriving and how hard the body is accelerating, and it gates real decisions. Below a threshold, automatic lap timing simply refuses to arm. The screen also distinguishes a true fix from an estimate and shows you the age of the last real one, so you are never quietly looking at a guess.
Most timers ask you to pick a circuit from a list, or to stand at the wall with a phone. This one asks you to drive.
A lap time you cannot trust is worse than no lap time, because you will chase it. So each lap is stored with the conditions that produced it: whether you marked it or the gate did, the navigation confidence at the moment of crossing, how fast fixes were arriving, the dilution of precision, and how far to the side of the line you actually passed.
A lap taken with a weak signal and eleven meters of cross-track error is visibly not the same thing as one taken at full confidence through the middle of the gate. The cluster shows you which is which instead of averaging them into a number that flatters you.
Starting the lap timer starts a recording. Not a video, and not a proprietary blob: an MCAP file, the open logging format from the robotics world, with five typed channels each carrying a published schema. It opens in standard analysis tooling, today and in ten years, whether or not we are still here.
Writing runs on its own thread at low priority, so recording never costs you a frame on the display. Only signals that matter for track analysis are captured rather than the whole bus, which keeps sessions small and readable. Pull the key mid-session and the file is closed out cleanly on the way down.
Acceleration and braking runs use the same position pipeline, which is why they are worth anything. Each test is individually switchable, so only what you care about is armed, and rollout can be applied the way a drag strip applies it: your car moves the first foot before the clock starts, exactly as it would over the beams.
Reserve a place in the first batch. No payment today. We’ll confirm fitment for your model year before anything is charged.