Gear ratio fundamentals & tractive effort curves
The gearbox keeps the engine inside its power band across the speed range. Tractive-effort curves per gear, compared with the ideal constant-power hyperbola, show what each ratio costs and gains.
Gear ratio fundamentals & tractive effort curves: L1 Foundation explanation
An engine only produces useful power over a band of rpm (for a racing NA engine maybe 5500–8500 rpm, for a turbo GT3 4000–7000 rpm). The car must cover 60–300 km/h. The gearbox bridges this: each gear maps road speed to engine speed through a total ratio
and
with the rolling radius in m and in rpm. A ratio is often quoted as tooth counts: a 13-tooth pinion driving a 34-tooth gear (written 13/34) gives a ratio of driven/driver = 34/13 = 2.615.
Torque goes the other way: wheel torque = engine torque × × η. Low gears multiply torque (big wheel force, low speed range), high gears trade force for speed.
Tractive-effort diagram. Plot wheel force against speed for each gear. You get a family of humped curves. Overlay the ideal curve — what an infinitely variable gearbox would give at peak power everywhere. The gaps between the gear curves and the ideal hyperbola are the acceleration lost to having only six ratios; closer ratios fill the gaps, at the cost of more shifts (each one costs time) and a narrower speed range per gear.
Also overlay resistance: drag plus rolling resistance. Where the top-gear curve meets resistance is the theoretical top speed. And overlay the traction limit (≈ μ × driven-axle load): any gear curve above it is unusable force — that is why 1st gear in a powerful car mostly produces wheelspin.
Typical numbers: driveline efficiency 0.88–0.93 (spur-gear sequential gearbox + bevel/hypoid final drive); GT3 6-speed with total ratios from ≈ 9–10 in 1st to ≈ 3 in 6th; rolling radius ≈ 0.33–0.35 m (GT3 rear), ≈ 0.30–0.33 m (formula).
Gear ratio fundamentals & tractive effort curves: L2 Practitioner explanation
What engineers actually change. In open categories you choose individual gear pairs and the final drive; in homologated ones (GT3, F2, LMP2…) you may change little or nothing — but you still need the math to choose shift points, judge corner gears and diagnose data.
Rolling radius is not a constant.
- Loaded radius < unloaded radius; the effective rolling radius for speed calculation lies between, typically 2–3 % below unloaded on a slick.
- Tyres grow with speed (centrifugal growth: a few mm at 250+ km/h on bias-ply slicks), so the ratio-speed relation drifts at the top.
- Tyre wear reduces radius by ≈ 1–2 mm over a stint; pressure changes radius a little. → Calibrate the effective radius from data: compare engine rpm and a GPS or front-wheel speed in a stable high gear.
Equivalent mass. Engine, flywheel, clutch and gearbox input inertia appear at the wheels multiplied by : . For a racing car ( ≈ 0.05–0.1 kg·m² for engine + carbon clutch, four wheel/tyre/disc assemblies ≈ 5 kg·m² in total) the rotating parts add ≈ 7–12 % in 1st and ≈ 4–5 % in top gear. The wheels alone contribute ≈ 3–4 % in every gear; gearbox shafts and gears add a little more. Road cars with heavy flywheels and a 1st-gear total ratio of 12–15 reach ≈ 40–60 % in 1st (Gillespie's empirical mass factor , also used by Segers Eq. 4.6). That formula is calibrated on road cars and overstates a racing car's figure: at = 9.6 it would give +27 %, against the ≈ 8–9 % computed from racing-car inertias. A short 1st therefore gives less acceleration advantage than the torque multiplication suggests.
Reading the tractive-effort chart for decisions:
- Crossover of adjacent gear curves = optimum upshift speed (where the next gear gives more force).
- Gaps: if the force drop at the shift is large in a gear used on a long straight, that gap costs time on every lap.
- Traction ceiling: gears whose curve sits entirely above the traction limit in the corners where they're used are "too short" — the extra torque only makes the driver work harder (or TC cut).
Efficiency and heat. At 400 kW, 7 % driveline loss is ≈ 28 kW of heat into gearbox and diff oil — that is why gearbox oil coolers exist and why gearbox temperature is a vital sign.
Gear ratio fundamentals & tractive effort curves: L3 Advanced explanation
Ideal ratio spacing. If the engine had a perfectly flat power curve between and , a geometric progression with step keeps the engine in band in every gear. Real power curves are humped, and two effects argue for progressive spacing (smaller steps in higher gears):
- Time-weighting: the car spends more time per gear at high speed, and power deficit there costs more time because drag absorbs a larger share of power — net acceleration is small. For a GT3 at 250 km/h a 5 % power dip costs ≈ 10–15 % of net acceleration, and 20–30 % within the last ≈ 20–30 km/h before terminal speed.
- Traction-limited low gears: in 1st–2nd the engine's torque curve shape barely matters because the tyre limits force; larger steps cost little.
Example of a GT3-like progression (step = ratio of consecutive total ratios): 1.37 / 1.27 / 1.22 / 1.18 / 1.19. With a 7500 rpm shift, the drop is 2030, 1610, 1360, 1170, 1180 rpm respectively.
Power-weighted gap metric. For each upshift, the drive energy lost to the ratio gap is the force deficit between the ideal hyperbola and the gear actually used, integrated over the distance the car spends accelerating at full throttle through that speed band:
(The area on a force–speed chart, , has units of N·m/s = W and is only an area index, not an energy.) In practice: bin the force deficit by speed, multiply each bin by the full-throttle distance the track spends in it (a distance-based speed histogram) and sum. Comparing across candidate ratio sets is a quick way to pre-screen sets before running a full lap sim. For a time estimate, divide each distance bin by the net propulsive power at that speed, (J/W = s), and multiply by the decay factor from the exit-speed lesson ( = speed at the end of that straight): . A force deficit costs most time where net acceleration is small, unless it occurs right at the end of the straight.
Rotational inertia and shift dynamics. On upshift the engine speed must drop by ; the energy released, , is partly delivered to the car as a torque spike (clutchless dog engagement) — visible as a longitudinal-g spike after the cut. On downshift that energy must be supplied (blip). Lower engine inertia (small flywheel, carbon clutch) makes both faster and is why racing engines have rotating inertias typically 2–5× lower than road engines (road engine + dual-mass flywheel + clutch ≈ 0.15–0.3 kg·m²; GT3/formula engine + carbon clutch ≈ 0.04–0.1 kg·m²).
Limits of the chart. The tractive-effort chart is quasi-static: it ignores shift time, driver behaviour (part throttle in low gears), tyre slip (actual wheel speed > vehicle speed when traction-limited) and aero load changes on traction. Use it for intuition and pre-screening; decide with lap simulation and track data.
Gear ratio fundamentals & tractive effort curves: L4 Expert explanation
The driveline is a spring, not a rigid gear train. The quasi-static picture () hides a torsional system: engine + clutch inertia , reflected through , connected by the half-shaft and gearbox torsional stiffness (wheel-referred, ≈ 5 000–15 000 N·m/rad per side for a GT3, so ≈ 5–15° of twist per side at the low-gear traction limit) to the wheel inertia. The wheel is coupled to the road through the tyre: a tangential carcass spring plus the longitudinal slip stiffness, which acts as a speed-dependent damper ( gives ), so tyre damping of shuffle falls with speed. Reflected to the wheel, the engine inertia is ≈ 5–9 kg·m² in 1st (as much as all four wheels together) and < 1 kg·m² in top gear, so the driveline's first torsional mode ("shuffle") falls with gear: typically ≈ 4–10 Hz in the low gears and higher in top gear.
- Why it matters: a step in engine torque (end of the upshift cut, dog engagement, a downshift blip) excites shuffle. You see it as a decaying oscillation at that frequency on longitudinal g, engine rpm and driven-wheel speed straight after a shift. On a loaded exit the oscillating wheel torque modulates rear slip and can be what the driver calls "the rear steps out on the upshift".
- Calibration tools: torque ramp-in after the cut (the ramp time should be comparable to, or longer than, half a shuffle period), active anti-jerk damping in the engine torque model, and clutch slip in the driveline as a deliberate damper.
- Data method: band-pass rpm and wheel speed around the expected frequency and measure the logarithmic decrement per gear. A drop in frequency over a season points at softer shafts or worn splines (more backlash).
Efficiency is not a constant. Mesh losses depend on load, speed, oil viscosity and temperature: spur-gear pairs run ≈ 98–99 % each at high load, but churning and windage losses depend mainly on speed (and oil temperature) and hardly on load. Part-throttle and coast efficiency is therefore much lower than the 0.9 used on the tractive-effort chart, and a cold gearbox can lose several kW extra. For lap-time and fuel models, use a loss map (torque, speed, oil temperature) rather than one η; for top-speed predictions, remember the gearbox-oil and diff temperatures at the end of a long straight.
Limits of the equivalent-mass idea. assumes all inertias accelerate together. That holds during steady acceleration, but not during shifts (the engine changes speed while the car hardly does) or wheelspin (the engine and driven wheels accelerate while the car does not). In those transients the full equations of motion per inertia are needed, which is exactly what a driveline model in a lap simulator does.
Key formulas
- r
- rolling radius (m)
- n
- engine speed (rpm)
Trackside takeaways
v [km/h] = 0.377 · r · n / i_tot — know it by heart.
Tractive-effort curves vs the ideal P/v hyperbola show what your ratio gaps cost.
Calibrate the effective rolling radius from data; it is 2–3 % below unloaded and changes with speed and wear.
Engine inertia appears at the wheel × i²: low gears are heavier than they look.
Upper gears usually need closer spacing than lower gears (progressive ratios).
On your car
6-speed sequential transaxle (Xtrac, Hewland, Ricardo, manufacturer units). Ratios are homologated — usually one set — so ratio maths is used for shift points, corner-gear decisions and diagnostics rather than for choosing gears.
Check yourself
Rolling radius 0.33 m, 6th-gear total ratio 3.0, engine at 7500 rpm. Road speed?
- Segers – Analysis Techniques for Racecar Data Acquisition — §6.4 Total Gear Ratio Channel · speed–rpm relation, Eq. 6.1–6.3
- Segers – Analysis Techniques for Racecar Data Acquisition — §4.1 Torque and Horsepower · mass factor and rotational inertia
- Other — T. D. Gillespie, Fundamentals of Vehicle Dynamics (SAE, 1992), Ch. 2 · empirical mass factor
- Milliken & Milliken – Race Car Vehicle Dynamics — Ch. 20 Driving and Braking