# Damper velocity histograms, box plots and run charts

> The shock-speed histogram characterises each damper as installed on the car. Symmetry, zero-bin height, % time in low/high-speed bump and rebound, and statistics like standard deviation, skewness and kurtosis turn it into objective metrics you can track lap by lap.

Area: Dampers · Starting level: L2 Practitioner · Updated 2026-11-16
Source page: https://race-engineer-academy.vercel.app/lesson/dampers-histogram-analysis

## L1 · Foundation

### A picture of what each damper does over a lap
Take the damper velocity channel for one corner over a lap and count how much time it spends in each velocity bin (e.g. 5 mm/s bins from −300 to +300 mm/s). That's the **shock-speed histogram** (a method pioneered by Claude Rouelle and detailed by Segers ch11). Plot the four corners side by side.

What you look at first:
- **Shape**: ideally close to a bell curve (Gaussian), centred on zero.
- **Symmetry**: bump side (positive) vs rebound side (negative). For mechanical grip, Segers argues the ideal is symmetric – equal time and similar velocities in bump and rebound, keeping tyre load as constant as possible.
- **Width**: a wide base = lots of high-velocity movement (soft/under-damped or bumpy track); a sharp tall peak = stiff/well-damped.
- **Left vs right, front vs rear**: big differences between corners point at setup asymmetries, pressure differences or a damper fault. Left vs right compares directly; front vs rear only after correcting for motion ratio (shaft velocity = wheel velocity × MR, see L2).

**Key point:** the histogram is a **tool, not a target**. It doesn't show the ideal setup; it shows *how the car's suspension is behaving* on this track. Its shape changes with damping, springs, ARBs, mass, tyres, pressures and the track itself. Use it for comparison: car A vs car B, run 1 vs run 2, left vs right.

## L2 · Practitioner

### Reading the numbers
**Set up the histogram properly** (so histograms are comparable):
- Same bin width and same velocity limits for every corner and every run (e.g. 5 mm/s bins, ±300 mm/s).
- Sample pots at ≥ 200 Hz; filter consistently before differentiating.
- Include a zero bin (−2.5…+2.5 mm/s) or sum the two bins adjacent to zero.
- Express as % of samples (time), not counts.
- When zooming into a single corner, start and end at the same damper position – otherwise the extract is biased to bump or rebound.

**Metrics per corner (Segers Table 11.4 style)**
| Metric | Meaning |
|---|---|
| % LS bump / LS rebound (0…±25 mm/s) | body-motion time per direction |
| % HS bump / HS rebound (> 25 mm/s) | road-input time per direction |
| Avg bump speed / avg rebound speed | magnitude per direction |
| Zero-bin height | relative stiffness of the corner |

**Interpretation rules**
- More time in **LS rebound** than LS bump (and lower average rebound speed) → rebound is slow → relatively too much LS rebound (or too little bump). Segers' example: LF 3.5 % more time in rebound → reduce LF rebound or add bump; RR 5.1 % more time in bump → reduce RR bump or add rebound.
- **Front vs rear: correct for motion ratio first.** Shaft velocity = wheel velocity × MR, so with front MR 0.6 and rear 0.9 the same 35 mm/s wheel motion reads 21 mm/s (LS) at the front and 31.5 mm/s (HS) at the rear. Refer velocities to the wheel (divide by MR) or scale the 25 mm/s boundary by each axle's MR before comparing ends. Track input also differs front to rear (wheelbase filtering, axle loads).
- **Front vs rear HS percentages** that still differ by several % after that correction → the end with more HS time is under-damped at high speed (Segers: ~6 % more rear HS → add rear HS bump and rebound).
- **Zero-bin height** differing between corners of a symmetric car (Segers: LF 21 % vs others 14–15 %) → check tyre pressures first, then consider asymmetric damping.
- **Average shock speed** over a segment equals Δposition / T and must be ≈ 0 over a lap. If not, the segment start and end positions differ, the pot drifted (thermal growth, mounting), or the velocity math channel has a bias (filter/differentiator offset). A gain calibration error cannot cause it.

**Aero exception.** If you are deliberately using damping to hold attitude (e.g. heavy front rebound to keep the nose down), the histogram will *not* be symmetric – that's intended. Compare to your own reference shape for that car, not to a textbook Gaussian.

## L3 · Advanced

### Statistics, box plots and run charts
**Descriptive statistics** (spreadsheet or math channels):
- **Median** – should be near zero; positive = more samples in bump, negative = more in rebound. First indicator of asymmetry.
- **Standard deviation σ** – width; large σ = more HS movement. Doesn't capture asymmetry.
- **Skewness** – asymmetry of the distribution (sign depends on your bump-positive convention; check it with a synthetic test).
- **Kurtosis** (excess) – peakedness/tail weight. Race-car histograms are usually leptokurtic (> 0): a lot of time near zero plus long tails from bumps/kerbs. Higher kurtosis front vs rear means the front's motion is more concentrated at low speed.

Segers' two-configuration comparison: config 1 has the taller zero bin (14.2 vs 11.5 %), more LS time and the smaller σ (36.5 vs 43.3 mm/s), yet the *lower* kurtosis (5.7 vs 8.9). Config 2 is flatter and wider overall but has heavier tails relative to its σ. Lesson: kurtosis describes shape (tail weight relative to σ), not peak height.

**Box plots** (Segers' adaptation): box from 25th to 75th percentile, median line, whiskers at **15th and 85th** percentiles (not min/max, which are kerb strikes), plus average bump and rebound speeds. Advantages: objective (no bin choices), four corners comparable at a glance. A bigger box = flatter histogram. Unequal distances from median to the 25th/75th percentiles = asymmetry.

**Run charts.** Compute per lap the *time* (not %) each damper spends in LS bump, LS rebound, HS bump, HS rebound (integrate a 1 when velocity is in range, reset per lap). Plot across the whole weekend. Sessions, setup changes, fuel load, tyre degradation and track evolution all show up as steps or trends; it's the fastest way to verify a damper change did what you intended and to spot a damper failing (sudden HS increase on one corner).

**What the simple model says** (Segers' single-corner simulations): more damping → taller, narrower peak; softer spring → also sharper (the same damper is relatively stronger); more mass → lower peak, wider base. Use it to sanity-check a change: if you added damping and the histogram widened, something else changed (pressures, track, a fault).

## L4 · Expert

### Pitfalls and a more rigorous approach
- **Noise inflation**: differentiating a pot signal amplifies noise ∝ frequency. An under-filtered velocity channel adds a symmetric "fake" HS population and reduces kurtosis. Use the same filter everywhere and verify with a stationary car (velocity should be ~0 ± 1 mm/s).
- **Track content dominates**: histograms of the same car at Monza vs Zandvoort differ more than any click change. Only compare within track/conditions; for setup decisions prefer *segment* histograms (a bumpy braking zone, a kerb set) over full laps.
- **Symmetry isn't a law**: a damper with rebound > bump under random road input packs down; a perfectly symmetric histogram doesn't guarantee the optimum CPLV. Use histogram symmetry as a starting hypothesis, then confirm with lap time, driver feedback and frequency-domain metrics.
- **Velocity ≠ force**: two dampers with identical histograms can have very different forces (different curves). Combining velocity data with dyno curves gives a **damper force histogram** and **energy per lap** per corner ($\int F v\,dt$) – a more physical comparison and a check for damper overheating in endurance stints.
- **Mode histograms**: build histograms of modal velocities (heave, pitch, roll, warp) instead of corners. A roll-velocity histogram isolates transient balance; a heave histogram isolates platform/road input. This removes the cross-talk that makes single-corner histograms ambiguous.

## Key formulas

**Time fraction in a velocity range**

$$p_{[v_1,v_2]} = \frac{1}{T}\int_0^T \mathbf{1}\left(v_1 \le v(t) < v_2\right)dt$$

- T: lap (or segment) duration (s)

**Standard deviation**

$$\sigma = \sqrt{\frac{1}{N-1}\sum_{i=1}^{N}(v_i-\bar v)^2}$$

- v_i: damper velocity samples (mm/s)

**Sample skewness**

$$A = \frac{N}{(N-1)(N-2)}\sum_{i=1}^{N}\left(\frac{v_i-\bar v}{\sigma}\right)^3$$


**Damper energy per lap**

$$E = \int_0^T F_d(v)\,v\,dt$$

- F_d(v): force from dyno curve (N)

## Trackside takeaways

- Use identical bins/limits/filters for all corners and runs; express as % time.
- For mechanical grip aim for symmetric, bell-shaped histograms; aero attitude control legitimately breaks symmetry.
- More LS rebound time than LS bump → too much rebound (or too little bump) on that corner.
- Zero-bin height differences between corners → check tyre pressures before asymmetric damping.
- Box plots (15/25/50/75/85 percentiles) and lap-by-lap run charts make damper behaviour trackable across a weekend.

## By car category

### GT3

Histograms show big HS tails from kerbs; segment the lap (exclude kerb zones) to judge LS behaviour. Compare left/right to catch fuel-load or driver-side asymmetries.

### LMP2

Corner and third-damper histograms should be analysed separately; third-damper histograms reveal heave/pitch platform behaviour directly.

### FIA F3

Stiff platform → narrow, tall histograms; small changes in σ matter. Noise filtering is critical with small velocities.

### One-make Cup

Identical cars make histogram comparison against team-mates powerful – a damper with a different shape is often a damper fault.

### TCR

Inside rear wheel lift produces asymmetric rear histograms (droop-limited rebound); interpret with care.

### GT4

Few adjusters: use histograms mostly to choose between available settings and to spot worn dampers over a season.

## References

- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §11.3 Shock Speed Ranges
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §11.4 The Shock Speed Histogram
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §11.4.3 Statistical Analysis (zero bin, median, σ, skewness, kurtosis, box plots)
- Segers, Analysis Techniques for Racecar Data Acquisition, 2nd ed. (SAE 2014): §11.6 Shock Speed Run Charts
