# Sequential gearboxes, paddle shift & what homologation leaves you

> Racing sequential gearboxes use dog engagement, a selector barrel and electronic shift control (cut + blip). In GT3, F2 and similar categories ratios are homologated — but shift calibration, maintenance and monitoring remain the engineer’s job.

Area: Gearbox & Gearing · Starting level: L1 Foundation · Updated 2026-11-16
Source page: https://race-engineer-academy.vercel.app/lesson/gearbox-sequential-homologation

## L1 · Foundation

**How a racing sequential works.**
- **Dog rings instead of synchromesh**: gears are engaged by dogs (typically 3–6 large teeth) sliding into slots on the gear. No synchroniser cones → much faster engagement, but speeds must be roughly matched and the dogs unloaded.
- **Selector barrel (drum)**: a cylinder with tracks that move the selector forks. Each barrel step = one gear; you can only go up or down one at a time — hence "sequential".
- **Actuation**: a lever pull/push or, in most modern cars, **paddles** on the steering wheel driving a pneumatic, hydraulic or electric actuator on the barrel.
- **Gear-control unit (GCU / ECU function)**: on upshift it requests an **ignition cut** (tens of ms) to unload the dogs so they can disengage without the clutch; on downshift it commands an **auto-blip** to raise engine speed. It also handles over-rev protection, neutral selection and fault handling (e.g. a missed shift retry).
- **Clutch**: carbon multi-plate, small diameter, low inertia — used for starting and pit stops only.

**Not every racing car has a dog box.** Road-derived GT4 and one-make cars often keep a **dual-clutch transmission** (Porsche 718 GT4 RS Clubsport: PDK; Ferrari 296 Challenge: DCT), and some TCR cars use DSG-type boxes. A DCT has two input shafts, each with its own clutch. The next gear is pre-selected on the idle shaft, and the shift is a clutch-to-clutch handover, so drive is almost never interrupted. There are no dogs to wear, but clutch temperature and wear become the life items.

**What homologation means.** In GT3 the gearbox, its ratios and usually the final drive are fixed by the car's homologation. In F2/F3/F4, LMP2 and one-make series the gearbox is spec. So most race engineers **never change a ratio**. What remains:

| Still yours (typically) | Usually fixed |
|---|---|
| Shift-light rpm per gear | Gear ratios, final drive |
| Corner-gear choice & shift points (driver coaching) | Gearbox internals, dog design |
| Differential preload/settings within allowed range | Cut/blip base calibration (often homologated or spec) |
| Gearbox maintenance (oil, dog inspection, life tracking) | Actuation system |
| Oil cooler configuration (if allowed) | |

Always check the series' technical regulations and the car's homologation form — some GT3 homologations include alternative ratios or final drives for specific high-speed events, and some series allow choosing among a small set.

## L2 · Practitioner

**Monitoring a sequential gearbox.** Channels: barrel position (potentiometer), gear, shift request, cut time, blip %, actuator pressure (pneumatic/hydraulic), gearbox oil temperature, shift time.

- **Gear position voltage**: a barrel sitting between nominal positions = incomplete engagement (dog "on the edge"). This damages dogs quickly and can jump out of gear under load.
- **Actuator pressure**: falling pressure across a stint (pneumatic leak, pump problem) slows shifts and causes missed shifts; set an alarm.
- **Gearbox oil temperature**: typical 80–130 °C; rises in traffic and hot conditions; sustained high temperatures accelerate wear and reduce oil film strength.
- **Missed shifts / false neutrals**: the GCU logs these; frequent occurrences in one gear point at worn dogs on that gear or a selector-fork issue.

**Life management.** Dog rings and gears have a life in km; endurance teams track gear-change counts and torque history per gear. Inspection intervals (crack testing of dogs/gears, barrel wear) are part of the engineer's job sheet. On **lever-operated** sequentials, drivers who pre-load the lever before the cut (or upshift without a cut or lift) side-load the forks and dogs and shorten their life. On **paddle** cars, upshifting flat out is the designed procedure: the paddle triggers the GCU cut before the barrel moves. The risks there are repeated refused or queued downshifts and over-aggressive downshifting at high rpm.

**Driver-adjustable features** (where allowed): neutral button, pit-limiter, launch mode, sometimes a choice of upshift-cut aggression or auto-upshift. Make sure the driver knows the procedure for a missed gear and for the "gear stuck/failsafe" mode — a common race-ending event that can sometimes be recovered via a reset.

**When shift calibration is allowed** (non-homologated or test cars):
- Cut time per gear: too short → dogs don't disengage (missed shift); too long → wasted time. Typically 30–60 ms.
- Torque ramp-in after the cut: a gentle ramp reduces driveline shock and rear instability on loaded exits.
- Blip amplitude per gear and per rpm: check with the total-ratio channel (no spikes) and rear wheel speed during downshifts.

## L3 · Advanced

**Dog engagement mechanics.** On upshift the cut makes driveline torque pass through zero; the old dogs release; the new dogs must catch the slots while relative speed between dog ring and gear is non-zero (the engine is still decelerating). Engagement succeeds if the dog can travel its engagement depth before contacting the slot face — a time window governed by relative speed, number of dogs and backlash. Higher relative speed → more "clatter" and impact loads; that is why large ratio steps in low gears are harder on dogs and why cut duration is gear-dependent.

**Back-cut (undercut) angle** on the dog faces holds the gear engaged under torque. Too little → jump-out under load; too much → hard to disengage, requiring longer cuts. Worn back-cuts are the root cause of the "jumps out of 3rd on the bumps" complaint.

**Why seamless shift is usually banned.** Seamless dog boxes (two gears momentarily engaged, one-way mechanisms; F1 since the mid-2000s) remove the torque gap. As shown in the shift-analysis lesson, each upshift gap costs roughly 0.2–0.6 of its own duration in lap time, so against a 40–80 ms cut-type box the gain at 15–25 upshifts per lap is several tenths per lap, at large cost. Most purpose-built racing categories ban them to control costs; a spec or homologated sequential with ignition cut is the norm there. The exception is road-derived cars that keep their **dual-clutch** gearbox (GT4, one-make Challenge, some TCR). They get near-seamless shifts by a different route, and in mixed fields a DCT car gains a little on every upshift relative to a dog-box car. BoP has to absorb that, so engineers should know it when they read straight-line speed traces against competitors.

**Homologated ratios as a design constraint.** With fixed ratios, the car's competitiveness across tracks is partly determined by how the manufacturer chose the set: a long top gear for Le Mans/Monza costs acceleration on twisty tracks. BoP compensates for overall pace but not for track-to-track ratio suitability; engineers then adapt with **drag level** (wing) — moving the car's terminal speed so it sits at the right rpm in top gear — and with corner-gear choices. Understanding where your ratio set is weak (e.g. a big 2–3 gap that hurts a specific hairpin exit) is valuable knowledge for setup priorities and for predicting your performance at each track before you get there.

## L4 · Expert

**Dog engagement as a timing window.** Model one engagement as a ring with $N$ dogs approaching a gear with $N$ slots. The useful angular window is the free angle between a dog and the slot it must enter, $\theta_w$ (slot width minus dog width, i.e. the designed backlash, often 15–30° on 3–6-dog rings). The dog needs time $t_{ax} = h/v_{ax}$ to travel the engagement depth $h$ at axial speed $v_{ax}$. In that time the parts rotate relative to each other by $\Delta\omega\,t_{ax}$, so a clean entry needs

$$\Delta\omega\, \frac{h}{v_{ax}} < \theta_w \quad\Rightarrow\quad \Delta\omega_{crit} = \frac{\theta_w\, v_{ax}}{h}$$

Example: $\theta_w$ = 20° (0.35 rad), $h$ = 4 mm, $v_{ax}$ = 0.5 m/s gives $t_{ax}$ = 8 ms and $\Delta\omega_{crit}$ ≈ 44 rad/s ≈ 420 rpm of relative speed. Above it, the dog "skates" over the dog faces with repeated impacts (the clatter you hear on a bad shift) until the speed difference falls. Below it, engagement happens at the next window, but the expected **waiting time** grows as $\approx \pi/(N\,\Delta\omega)$. At $\Delta\omega$ → 0 a dog sitting face-to-face can block completely.

There is therefore an **optimal band of relative speed**: enough to bring the next window round quickly, low enough to drop in cleanly. This explains several things trackside:
- A "perfect" synchronous blip can make downshifts slower or blocked. Calibrations deliberately aim slightly above synchronous.
- More dogs shorten the waiting time but narrow $\theta_w$ and the dog section (strength). Fewer, larger dogs with more backlash engage more reliably but clunk harder at every torque reversal.
- Faster barrel actuation (higher $v_{ax}$) raises $\Delta\omega_{crit}$. That is why actuator pressure loss shows up first as missed shifts in the gears with the largest relative speed (the low gears on upshift).

**Impact energy and damage.** Once engaged, the remaining speed difference is removed by the dog impact. The energy is of the order of $\tfrac12 J_{red}\,\Delta\omega^2$, with $J_{red}$ the engine-side inertia reduced to the dog shaft. That energy is absorbed in dog-face plastic deformation and driveline wind-up. Damage scales with $\Delta\omega^2$, so the low-gear upshifts (largest steps) and badly blipped downshifts dominate dog wear. Logging $\Delta\omega$ at engagement (from rpm and output-shaft speed) per shift, and accumulating $\sum \Delta\omega^2$ per gear, gives a far better dog-life predictor than a plain shift count.

**Limits.** The model ignores back-cut geometry (which pulls the dog in under torque), oil drag, fork compliance and barrel detent dynamics. Use it to understand trends and set inspection intervals, not to design dogs.

## Key formulas

**Relative dog speed at engagement (upshift, dog ring on the output shaft)**

$$\Delta \omega = \omega_{e}\left(\frac{1}{i_{g,n+1}} - \frac{1}{i_{g,n}}\right) > 0$$

- \omega_e: engine speed at the start of the cut (rad/s)
- i_{g,n}: gear-pair ratio of gear n only (no final drive)

## Trackside takeaways

- Dog-engaged sequential boxes need torque relief (cut) on upshift and speed matching (blip) on downshift.
- In GT3/F2/LMP2/one-makes ratios are fixed — the engineer manages shift points, corner gears, monitoring and life.
- Barrel position between nominal values = incomplete engagement: inspect dogs.
- Falling actuator pressure and rising shift times precede missed shifts.
- Check homologation forms and BoP for any alternative ratio/final-drive options per event.

## By car category

### GT3

Homologated 6-speed sequential with paddle shift; ratios fixed per homologation (occasional alternative sets for fast tracks). Diff preload often still adjustable within limits.

### GT4

Mix of dog-box sequentials and road-derived dual-clutch boxes (e.g. PDK in the Porsche 718 GT4 RS Clubsport). DCT cars have no torque gap and no dogs; monitor clutch temperatures and TCU fault codes instead.

### FIA F2

Spec 6-speed sequential with paddle shift and spec ratios — no ratio choice; engineering is shift-light setup, corner-gear coaching and reliability monitoring.

### FIA F3

Spec gearbox and ratios; same approach as F2 at lower power.

### Formula Regional

Spec sequential with paddles and fixed ratios; good training ground for shift-point and corner-gear analysis.

### One-make Cup

Porsche Cup and Super Trofeo: sealed/spec sequential dog boxes; Ferrari Challenge: road-derived DCT. Teams must not open them in some series, so monitor and report.

### LMP2

Spec Xtrac paddle-shift gearbox; ratios constrained by class rules; gearbox life tracking matters in endurance.

### Super Formula

Spec sequential paddle-shift gearbox; strategy around the Overtake System, not ratios.

## References

- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §6.1.2 Shift Duration (sequential vs H-pattern, power-shift systems)
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §6.2 Down-shifting
- Other: Series technical regulations & homologation forms (SRO/FIA) (ratio freedom per category)
