# The gear chart & the total-gear-ratio math channel

> Plot rpm against speed and every sample should fall on one line per gear. The total-gear-ratio math channel turns that into a stepped trace that reveals gear usage, wheelspin, lock-ups and poor down-shift blips.

Area: Gearbox & Gearing · Starting level: L1 Foundation · Updated 2026-11-17
Source page: https://race-engineer-academy.vercel.app/lesson/gearbox-gear-chart-total-ratio

## L1 · Foundation

**The gear chart** (speed–rpm chart) draws, for each gear, the straight line $n = v\, i_{tot}/(0.377 r)$. The engineer uses it to see:
- the speed range of each gear between the post-shift rpm and the shift rpm,
- the rpm drop at each upshift,
- which gear a corner of a given speed falls into.

**See it in the gearbox.** Below, a six-speed sequential gearbox is cut open: only the lit gear pair carries the drive, the dog rings and selector barrel choose which. The chart is the gear chart (engine rpm against road speed, one straight line per gear, shifting at 8500 rpm) and the bars show engine, output-shaft and wheel rpm. The figure uses its own ratio set (3.17–1.08, final drive 3.60, 0.345 m rolling radius), close to the example that follows.

**Try it:**
- In **4th**, move the road speed: the three rpm bars keep fixed ratios – divided by **1.48** in the gearbox and by **3.60** in the final drive. At 8500 rpm, 4th gives **208 km/h**.
- Select **5th** at that speed: road speed stays the same, the engine drops to about **7260 rpm** – the bold drop on the sawtooth. From 1st to 2nd the drop is much bigger (to about **6135 rpm**).
- Pick a low gear at high speed: the box steps down one gear at a time and brakes first, so the engine never over-revs – the way a sequential box is driven.

**Takeaway:** each gear is a straight line through the origin on the gear chart, and the ratio spacing sets the rpm drop. Gaps get smaller in the higher gears so the engine stays near peak power where the car accelerates least.

Example, a GT3-like set (r = 0.335 m, shift at 7500 rpm):

| Gear | $i_{tot}$ | Speed at 7500 rpm | rpm after upshift |
|---|---|---|---|
| 1 | 9.6 | 99 km/h | 5470 |
| 2 | 7.0 | 135 km/h | 5890 |
| 3 | 5.5 | 172 km/h | 6140 |
| 4 | 4.5 | 210 km/h | 6330 |
| 5 | 3.8 | 249 km/h | 6320 |
| 6 | 3.2 | 296 km/h | – |

**From data** (Segers §6.3): make an X-Y plot of engine rpm vs wheel speed using only samples with positive longitudinal acceleration. Ideal data falls on straight lines through the origin. You immediately see:
- **which gears are used** (Segers' example: 1st never used; 3rd most dense),
- the **maximum shift rpm** per gear,
- **scatter** off the lines — wheelspin (points to the high-rpm side of the line, in low gears especially) or wheel-speed sensor location effects in corners.

**The total-gear-ratio channel** inverts the relation sample by sample:

$$i_{tot} = 0.377\, \frac{r\, n}{v}$$

Plotted against distance it is a staircase, one step per gear. **Sign convention:** in this app $i_{tot}$ is always the positive value above, so higher gears give *lower* steps. Segers multiplies the channel by −1 so that his staircase rises with higher gears; on his plots every up/down reading is therefore inverted relative to ours. A gear-position sensor shows the selected gear more cleanly, but the calculated ratio shows **what the driveline is really doing**.

## L2 · Practitioner

**Diagnostics with the total-ratio channel.** Any deviation from the ideal step value means engine speed and road speed are not in the expected relation. Positive convention ($i_{tot} = 0.377\,r\,n/v$; a downshift steps *up* to a higher value):

| Pattern on the ratio channel | Meaning |
|---|---|
| Ratio rises above the step on throttle (calc. from undriven-wheel speed) | driven wheels spinning — compare with slip ratio |
| At a downshift the trace stays below, or dips under, the new (higher) gear's value before settling | engine below synchronous at engagement — insufficient blip; the driven wheels must spin the engine up, a braking spike on the driven axle (rear destabilised in RWD) |
| At a downshift the trace overshoots *above* the new gear's value | engine above synchronous — over-blip (over-rev risk, momentary drive pulse on entry) |
| Ratio dropping below the step under braking (undriven-wheel or GPS speed source) | driven wheels slower than the car: rear lock-up from the brakes or from engine braking |
| Long transition between steps | slow shift or clutch slip |
| Value between two steps in a constant gear | wrong radius or slipping clutch |

Use the **undriven-wheel speed** (or average of both fronts in RWD) for the speed input when you want to see slip; use **driven-wheel speed** when you want a clean gear detection.

**Theoretical rpm in other gears (Segers Eq. 6.4).** For every gear make a channel

$$n_k = 2.653\, \frac{i_{tot,k}\, v}{r} \quad (v \text{ in km/h})$$

and plot them against distance. At any corner you read the rpm the engine *would* have in the gear above and below. Segers' example: a 113 km/h corner taken in 3rd at 5390 rpm would be 6890 rpm in 2nd and 4460 rpm in 4th. Then ask: in 2nd, would the driver hit the limiter before the corner exit ends (forcing an upshift on the exit)? In 4th, is 4460 rpm inside the torque band? This turns the endless "which gear for turn 7?" debate into a quantified discussion.

**Practical setup:** gear detection = round $i_{tot}$ to the nearest known ratio within a tolerance (e.g. ±4 %), only when throttle > 10 % and clutch engaged. Keep the effective radius updated (new vs worn tyres) or the bands will drift.

## L3 · Advanced

**Calibrating the effective radius from data.** In a high gear on a straight at moderate throttle (low slip) compute $r_{eff} = i_{tot}\, v_{GPS}/(0.377\, n)$. Do it for each tyre set and against speed: the increase of $r_{eff}$ with speed quantifies tyre growth; the decrease over a stint tracks wear (≈ 1–2 mm). Feed the radius into TC, speed calculations and the gear chart.

**Slip from the ratio channel.** If the speed source is the undriven wheels, the relative deviation of the measured ratio from the nominal value *is* the driven-wheel slip ratio:

$$SR = \frac{i_{meas}}{i_{nom}} - 1$$

That gives you a slip channel even when regulations restrict wheel-speed sensors on the driven axle (Segers' "backdoor" in §4.2). Accuracy is limited by the assumed rolling radius (see Level 4), by driveline wind-up during torque *changes* and by the differential (which averages the two driven wheels). A steady twist does not alter $n/v$, but a changing torque produces a wheel-to-engine speed error $\Delta\omega = \dot T/k_\theta$: a few % of apparent slip for 50–100 ms after a cut or a tip-in in the low gears.

**Shift quality metric.** For each upshift, measure: (a) time from cut start to ratio reaching the new step within 1 %; (b) the dip of $i_{tot}$ below the new (lower) step, i.e. engine speed undershooting synchronous on engagement → torque spike; for each downshift, (c) the area of the ratio trace below the new (higher) target (insufficient blip) or above it (over-blip). On Segers' −1 channel these signs flip. Histogram per driver and per gear across a stint: degradation in shift quality late in endurance stints often precedes a gearbox failure (dog wear, actuator pressure).

**Pitfalls.** Wheel-speed sensors mounted on one side see corner-radius effects (an outside-wheel sensor reads ≈ $t/(2R)$ high versus vehicle centreline speed: 1.6 % with a 1.6 m track at R = 50 m, and 3.2 % versus the inside wheel). Use the average of left and right where possible. Low-speed quantisation of Hall sensors (few teeth) makes the ratio noisy below ≈ 30 km/h — gate the channel on speed.

## L4 · Expert

**The identification problem: radius vs slip.** The ratio channel measures $n/v$. A deviation from nominal can come from the driven tyre's effective radius changing (growth with speed, wear, pressure, load) or from slip, and **the two are indistinguishable at a single instant**: under drive, effective rolling radius and slip ratio enter the equation as a product. Any slip channel built from it is only as good as the radius you assume. Practical consequences:
- At 250 km/h a slick may have grown 1–3 mm (≈ 0.3–1 %) relative to the 100 km/h calibration. Uncorrected, that shows as 0.3–1 % "slip" at top speed, the same order as the real driving slip there.
- Driving slip is not zero even at moderate throttle: a tyre carrying 2–3 kN of drive force needs ≈ 0.5–1 % slip. A radius calibrated "at moderate throttle" therefore already contains some slip.

**Estimating $r_{eff}$ properly.** Treat it as a slowly varying parameter. Two workable methods:
1. **Least squares on gated data.** Take samples with near-zero drive force (coasting in gear with the clutch engaged and throttle at the zero-torque point, or a constant-speed cruise) and fit $r_{eff}(v) = r_0 + c\,v^2$ against GPS or undriven-wheel speed. The $v^2$ term is the centrifugal growth.
2. **Kalman filter.** State: $r_{eff}$ (random walk with small process noise) plus slip. Measurements: engine speed, undriven-wheel speeds and GPS. Drive force from the engine-torque model provides the slip prior ($\kappa ≈ F_x/C_\kappa$). The filter separates the slow radius drift (wear, temperature) from the fast slip transients. That is essentially what a good TC system does internally.

**Undriven-wheel speed is not "vehicle speed" either.** Front tyres under braking slip and change radius too, in corners they run at different radii (yaw rate × half-track), and under combined slip their longitudinal speed is not the CG speed. For the cleanest reference, fuse GPS (absolute speed, low rate, latency ≈ 50–100 ms), IMU long acceleration (high rate, drifts) and the front wheels with a complementary filter.

**Driveline wind-up and backlash.** Between engine and wheel there are several degrees of torsional wind-up at the wheel (≈ 5–15° per side at the low-gear traction limit, less in the high gears) and some backlash. In the steady state this does not change $n/v$, but every torque reversal (lift, cut, blip) makes the engine run ahead of or behind the wheels for tens of milliseconds. Short spikes after a shift are therefore partly driveline dynamics, not slip or blip errors. Judge shift quality from the integrated area and settling time, not from the peak of a single sample.

**Limits.** Below ≈ 30 km/h (quantisation), during clutch slip (pit exits, launches) and with an open/locked diff in a turn (the engine sees the average of the two driven wheels), the channel cannot be interpreted with a single nominal ratio.

## Key formulas

**Total gear ratio channel (Segers Eq. 6.3)**

$$i_{tot} = 0.377\, \frac{r\, n_{engine}}{v}$$

- r: rolling radius (m)
- n_engine: rpm
- v: speed (km/h)

**Theoretical rpm in gear k (Segers Eq. 6.4)**

$$n_k = 2.653\, \frac{i_{tot,k}\, v}{r}$$


**Slip from ratio channel**

$$SR = \frac{i_{meas}}{i_{nom}} - 1$$


## Trackside takeaways

- rpm vs speed (accelerating samples only) shows gear usage, shift rpm and wheelspin at a glance.
- Total ratio = 0.377·r·n/v; a staircase that exposes spin, lock-ups and poor blips.
- On the positive ratio channel, a downshift dip below the new step = too little blip; an overshoot above it = over-blip (on Segers’ −1 channel the signs flip).
- Theoretical-rpm channels per gear quantify "which gear for this corner".
- Keep the effective rolling radius calibrated per tyre set.

## By car category

### GT3

With homologated ratios, the gear chart is used to choose corner gears, check limiter hits in the tow and diagnose shift quality; GT3 loggers usually provide gear position, but the calculated ratio still reveals blips and spin.

### FIA F3

Spec data system with limited channels; the total-ratio channel from rpm and front wheel speed is the main tool to see rear wheelspin when driven-wheel sensors are not available.

### FIA F2

Spec gearbox and ratios; drivers differ in corner-gear choice and shift points — overlay theoretical-rpm channels to settle debriefs.

### One-make Cup

Identical cars and ratios: the gear chart shows driver differences (gear choice, minimum rpm) directly.

## References

- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §6.3 The Gear Chart (X-Y rpm vs speed)
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §6.4 Total Gear Ratio Channel (Eq. 6.1–6.3, −1 sign convention)
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §6.6 Determining in Which Gear to Take a Corner (theoretical rpm per gear, Eq. 6.4, 113 km/h example)
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §6.2 Down-shifting (ratio spikes from insufficient blipping)
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §4.2 Traction and Longitudinal Slip (driven-wheel speed from rpm, Eq. 4.14)
