# The beginner’s logging kit and sensor plausibility checks

> Learn what the six basic signals and damper pots look like on a good lap, and how to verify sensors with physics – pit-lane steady-state checks, sign conventions and cross-channel consistency.

Area: Data Acquisition & Analysis · Starting level: L1 Foundation · Updated 2026-11-16
Source page: https://race-engineer-academy.vercel.app/lesson/daq-beginners-kit-sensor-validation

## L1 · Foundation

**Know what normal looks like.** Before you can spot a problem you must know the healthy shape of each trace:

- **Engine RPM**: a sawtooth on upshifts (shift RPM roughly constant), sharp upward blips on downshifts. A ragged trace out of a slow corner means the car is fighting for traction.
- **Speed**: the *results* graph. Rising = accelerating, falling = braking; minimum corner speeds and top speeds are the first numbers to compare. Experienced engineers navigate the lap by the speed trace, not the map.
- **Throttle (%)**: 0 = off, 100 = flat. Meaningful only alongside other channels.
- **Steering angle**: SAE convention is positive for right turns (check your car – sign conventions vary).
- **Lateral g**: positive for right turns in SAE convention. Peak value depends on grip and speed ($a_y = V^2/R$): a GT3 car reaches ~1.6–2.2 g in fast corners, a Super Formula or F2 car 3–4 g.
- **Longitudinal g**: negative when braking (peaks of about −1.4 g from low speed to −1.9 g from high speed for GT3 with ABS – Segers' Nogaro GT3 reaches −1.95 g – and −2.5 to −3.5 g for high-downforce single-seaters at high speed). Acceleration falls with speed because tractive force is power-limited ($F \approx P/V$) and drag rises with $V^2$; upshifts show as short dips.
- **Damper position**: bump is wheel moving up relative to chassis (positive by convention). The raw trace mixes slow body motion (roll, pitch, aero) and fast road input.

**Sensor checks before every event:** calibrate and zero potentiometers, steering, accelerometers and brake pressure on the setup pad, and verify sign conventions.

## L2 · Practitioner

**Plausibility via physics.** The aggressive environment (vibration, heat, water) means sensors drift and fail. A robust check compares each measurement with an expected value from a simple calculation. The best moment is **steady straight-line running in the pit lane**: limiter on, constant speed.

In that window (identified by a condition such as pit-limiter active, $|G_{lat}|$ and $|G_{long}|$ near 0, steering near 0):

| Channel | Expected |
|---|---|
| Longitudinal g, lateral g | ≈ 0 |
| Yaw rate | ≈ 0 |
| Steering angle | ≈ 0 (else offset or alignment issue) |
| Damper positions / ride heights | static setup values, corrected for current fuel load |
| Wheel loads (if load cells) | corner weights, corrected for fuel |
| All four wheel speeds | equal within tyre circumference tolerance |

**Fuel correction**: setup corner weights and ride heights are measured at a reference fuel load. With extra fuel, the expected ride height at each wheel drops by
$$\Delta z = \frac{\Delta F_{wheel}}{k_{ride}}, \qquad k_{ride} = \frac{k_w\,k_t}{k_w + k_t}$$
with $k_w$ the wheel rate, $k_t$ the tyre radial rate and $\Delta F_{wheel}$ the fuel weight shared by tank position (≈ $\Delta m\,g/4$ for a central tank). Laser ride heights see spring *and* tyre ($k_{ride}$); damper pots see only the suspension part ($\Delta F_{wheel}/k_w$, times MR for damper travel). For a GT3 ($k_w$ ≈ 160–180 N/mm, $k_t$ ≈ 320–340 N/mm, so $k_{ride}$ ≈ 110 N/mm), 40 kg of fuel is ≈ 1 mm. Small – but the same size as the sensor errors you are hunting, so always correct. Assume linear stiffness at pit-lane speeds.

Compute the average offset between sensor and expected value over the pit-lane window, and flag channels outside tolerance. Segers' example from BHAI Tech shows a rear ride-height laser reading 6 mm off – clearly beyond tolerance – while two others sat at the edge.

**Cross-channel checks:** $G_{lat}$ vs $V \cdot r$ (yaw rate) in steady corners; $G_{long}$ vs the derivative of GPS speed; differentiated damper pot sign vs vertical accelerometer; brake pressure front/rear ratio vs the balance the driver selected.

## L3 · Advanced

**Building a validation pipeline.** Top teams automate this as an out-lap report:

1. Detect steady windows automatically (pit lane, constant-speed straights, static garage).
2. Compare each channel against expected values held as session constants (updated with every setup change).
3. Show offsets with traffic-light thresholds; allow one-click re-zeroing for channels where an offset is legitimate (e.g., a steering-angle sensor re-installed after a column or rack change). Before re-zeroing steering, confirm on the setup pad: a toe split or rear thrust angle also makes the driver hold a few degrees of lock to drive straight, and re-zeroing would hide a real alignment fault.

**In-lap vs out-lap.** On the in-lap the tyres are hot: higher pressures and different radial stiffness change ride heights and loaded radius. Formula car tyres change much more than GT tyres. Use the out-lap for ride-height and load checks unless you have a tyre stiffness model as a function of pressure and temperature.

**Accelerometer mounting.** An IMU not mounted at the CoG reads extra terms: $a_{measured} = a_{CG} + \dot{\omega}\times r + \omega\times(\omega\times r)$. With high yaw acceleration in a quick direction change, a sensor 0.5 m ahead of the CoG adds noticeable lateral g. Also, body roll and pitch tilt the sensor so it picks up a component of gravity: ~0.017 g per degree. A GT3 pitches only ~0.6–0.8° under heavy braking (formula cars less than 0.5°), so 0.8° adds ~0.014 g to long g. Two larger terms are often forgotten: a 1° IMU mounting tilt (0.017 g, present all the time) and **track gradient** – a 3 % slope is 0.03 g and Raidillon at Spa is about 0.17 g. For drag or brake-efficiency work from long g, correct the gravity term with GPS altitude or a track elevation map.

**Motion ratio and sensor geometry.** Ensure the pot measures damper travel, not a lever with its own ratio; otherwise correct in the calibration. A wrong motion ratio corrupts every derived channel (wheel travel, roll angle, ride height).

## Key formulas

**Gravity component from body tilt**

$$\Delta a \approx g\,\sin\theta \approx 0.0175\,\theta_{deg}\; [g]$$

- \theta: pitch or roll angle of the sensor
- \Delta a: error in measured acceleration (g)

**Steady-state lateral acceleration check**

$$G_{lat} \approx \frac{V\,r}{g}$$

- V: speed (m/s)
- r: yaw rate (rad/s)
- g: 9.81 m/s²

## Trackside takeaways

- Learn the healthy shape of each basic trace before hunting problems.
- Zero and sign-check sensors on the setup pad before every event.
- Use pit-lane steady running to compare sensors with expected values corrected for fuel load.
- Cross-check channels with physics: lat g vs speed × yaw rate, long g vs d(speed)/dt.
- Tilt and mounting offset add ~0.017 g of gravity per degree; track gradient is often the bigger term.

## By car category

### Formula 4

An AiM-type logger with internal accelerometers and GPS: mounting orientation errors are common in student teams – check that the g channels read zero on the flat pad.

### Formula Regional

Small fuel tanks mean fuel corrections are small, but tyres change a lot with temperature – validate ride heights on out-laps only.

### GT3

Large fuel loads (100+ litres, ~75 kg) change expected ride heights by ~1.5–2 mm, more at the axle nearer the tank; that is the size of a real sensor error, so always correct pit-lane checks for fuel.

### Hypercar (LMH/LMDh)

Automated out-lap validation reports are standard; dozens of channels are checked before the car completes its first flying lap.

### TCR

Check steering-angle zero after any alignment change – a small offset biases understeer-angle comparisons for the whole weekend.

## References

- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §3.5 The Beginner’s Data Logging Kit
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §3.6 A Possible Approach to the Testing of Sensor Readings (contributed by G. Callea, BHAI Tech)
