A team racing large-displacement V8s in a closed-circuit series runs every engine on the same conservative map. This case follows one engine from a logged baseline to bounded, traceable changes, over six weeks and two race weekends.
Spark advance and fuel target, per operating region
Duration
Six weeks, two race weekends
Runs on
The team's own hardware
> The starting point
A Safe MapNobody Can Explain
All four engines run one calibration, written two seasons ago and adjusted since by experience. It is safe because it is conservative: spark advance sits well back from where knock would begin, on every engine, in every region.
Nobody can say how far back. The port logs sample a few times a second while the engine fires every 2.5 ms at 6,000 rpm, so the knock events that set the margin happen between samples. And a change is rarely written down next to the log that justified it.
How the decision is made today
Who
One tuner, between sessions
On what
Port logs at a few samples a second, and experience
How long
An evening per change, tried at the next session
Trail
Notes, when there are notes
# What could not move
Four LimitsThe Design Started From
Each constraint was settled before anything was built, and each one fixed part of the architecture.
No parts replaced
The series rules fix the hardware. Nothing is bolted on and nothing is swapped.
So the system works only through the calibration the engine already runs.
The engine reports its own limits
Knock retard, mixture and temperatures are the only authority on how far a change may go.
So the system proposes no change that a reported limit does not stand behind.
Closed circuit only
The work stays off the public road on purpose. Changing a road car's calibration can need an emissions approval, and that is a question per vehicle and per country.
So the system makes no road-use claim; that answer is given per vehicle, before a session.
Every change traceable
After the weekend, the team needs to know which change did what.
So the system records each change against the logged events it came from.
% The architecture
From Crank AngleTo A Signed-Off Change
Four layers, each with one job. The system decides what is worth proposing; the tuner signs off every change.
Read
Ignition, injection and knock windows, by crank angle
Manifold pressure and airflow as the load axis
Coolant and intake temperature as the boundary
Check
Each region's data against knock, mixture and temperature limits
Regions with too little data held, not guessed
Propose
Changes per operating region, in small steps
Headroom to the nearest limit beside each one
Record
Each change tied to the pulls it came from
The next session measured against the last
Runs onThe team's own hardware, at the circuit. Engine data does not leave it.
= Six weeks
Two Race Weekends,One Engine First
One engine goes through the whole loop before the next follows, so every change has a clean comparison.
Week0123456You leave with
BaselineWeek 0You leave with: A logged baseline, three pulls
The team
Dyno time, the current calibration, past notes
Ecarus
Logging set up at the rate each signal needs
First changesWeeks 1–2You leave with: The first recalibration log
The team
Dyno sessions, the tuner's sign-off
Ecarus
Bounded proposals for regions with clean data
Race weekend oneWeek 3You leave with: Changes kept, held or reversed
The team
Track sessions with logging on
Ecarus
Each session compared with the baseline
Second engineWeeks 4–5You leave with: Two engines, two calibrations
The team
The second engine on the dyno
Ecarus
The same loop, on that engine's own data
Race weekend twoWeek 6You leave with: A decision on the rest of the rotation
The team
The team's review
Ecarus
Results per region, a plan for the other two
Success criteria, agreed in week 0
Knock events per pull, against the baseline
Spread between repeated pulls on the same day
Share of changes traceable to the log behind them
Time from a log to a signed-off change
= The decision, as delivered
The Log The TunerSigns Off
Week two's output: one region moved, one held, and the reason for each beside it.
Recalibration log
Week 2
Engine
Large-displacement V8, naturally aspirated
Session
Three logged pulls, same fuel
Limits in force
Knock retard, lambda, coolant and intake temperature
Proposed changes, by operating region
Region
Parameter
Change
Headroom to limit
Bounded by
3,500–4,500 rpm, high load
Spark advance
+1.0° in two steps
Moderate
No knock retard across three pulls
4,500–6,000 rpm, high load
Spark advance
Held
Narrow
Knock retard seen in one cylinder
2,000–3,500 rpm, high load
Fuel target
λ 0.86 → 0.88
Wide
Intake temperature stable, no lean excursion
Idle
Cam phasing
No change proposed
Moderate
Insufficient data at this temperature
Why the high-rpm row is held
One cylinder pulled timing in the second pull above 4,500 rpm. A change that would need that margin is not proposed until a clean log says otherwise.
In the product, every line opens onto the logged events it was derived from. A change that would cross a reported limit is not proposed.
~ What gets measured
Against The Team'sOwn Baseline
Each outcome is measured against the team's own week 0 baseline. Results are published with the team's approval.
Measure
Baseline
How it is measured
Result
Knock events per pull
Three logged pulls in week 0
Same dyno, same fuel, same pull procedure
Published on approval
Power at the wheels
The week 0 dyno sheet
Same dyno and correction standard; no figure quoted before it is measured
Published on approval
Repeatability
Spread across the baseline pulls
Spread across the same number of pulls after each change
Published on approval
Time to a signed-off change
An evening per change today
From log to sign-off, per change
Published on approval
What stays with the team
Every calibration change, with the log behind it
The logging setup, at the rate each signal needs
A per-region record of where the headroom is, and where it is not
Close to yours if
You run several engines of one type on one calibration
Your logs cannot see what happens inside a cycle
Changes are made by experience and rarely recorded against data
You run on a closed circuit or a dyno
Your operation, in the same format
The technical briefing is where your specific constraints get discussed, including which of these four your operation resembles. Results are published with each customer's approval, never before.
