HANDLING PASS 06 / 14 October 2026

ACCEPTED BASELINE
On 14 October 2026 the user confirmed that NB4 reverse works correctly and the
wider track feels much better, and selected v0.6 as the baseline for future work.
The supplied screenshot shows v0.4 with power 90%, punch 100%, grip 1.93,
steering 26 degrees, steering softness 3%, rear weight 58%, high-speed push 6%,
and lift-off drag 30%. These are now the startup and Restore baseline values.
The forward motor boost also uses the screenshot's v0.4 3-7 m/s taper, preserving
the accepted response rather than silently retaining the later v0.5 curve.
NB4 channel calibration and stored controller profiles are unchanged.

REVERSE
First brake press always brakes. Release to neutral, then press brake again
near a stop to engage reverse. Holding brake through the stop never reverses.
A second press while still moving continues to brake; release and press again
after stopping. A 50 ms neutral interval and separate press/release thresholds
reject trigger jitter. The same behavior works with the keyboard's Down / S.
Reverse uses rear-wheel torque and the existing tire model, with a gentler motor
curve that settles near 9.78 km/h. Neutral slows reverse to rest. Forward throttle
first brakes backward motion and then drives forward near a stop. Pause, reset,
and controller loss clear reverse permission. The Live input panel shows the
selected direction and when reverse is ready. Keyboard pedal release is immediate
so the neutral step is not hidden behind input smoothing; NB4 input is unchanged.

WIDER TRACK
Lane width increases from 1.5 m to 2.1 m (40%). Waypoint spacing increases 20%,
and a blend of interpolating and rounded curves opens the tight bends. All carpet,
barrier, and shadow outlines remain free of crossed rails. Circuit length is now
47.616 m. Car dimensions and physics units are unchanged; the camera fits the
larger floorplan. Collision geometry and painted lane width use the same value.

VALIDATION
All 43 checks pass: 9 baseline, 5 historical tire response, 7 steering/contact/
geometry, 7 accepted drivetrain comparisons, and 15 ESC/reverse checks. Against
the accepted v0.4 screenshot setup, full-trigger 0-6 m/s remains 0.768 s, neutral
6-4 m/s remains 0.728 s, and top speed remains 34.556 km/h at 90% power.
The new reverse tests include returning to forward, mirrored steering, refusing
reverse while still sliding, and backing 0.54 m away from a head-on barrier in
0.6 s. Browser checks verify first-press braking, second-press reverse, returning
to forward, saved setup values, and the wider rendered track. The user has now
confirmed the reverse sequence on the physical NB4 and accepted the wider track.

PREVIOUS HANDLING PASS 05 / 14 October 2026

DRIVE RESPONSE
The requested next step was more punch or passive deceleration. Pending a more
specific preference, this pass strengthens both. Punch rises from 60% to 100%,
with full extra torque held to 4 m/s wheel speed and tapered smoothly to zero at
7.5 m/s. This gives useful rolling acceleration as well as a harder launch.
High-speed motor output remains the original curve. Lift-off drag rises from
18% to 30%. Steering, tire-force coefficients, controller mapping, and barrier
response are unchanged. The extra torque can use more of the rear tires' grip;
this is a high-grip tune, and lower-grip experiments may need less punch.

MEASURED IN THE MODEL (baseline car on open ground)
Test                              v0.4        v0.5
0 to 14.4 km/h                     0.484 s     0.400 s
0 to 21.6 km/h                     0.994 s     0.800 s
14.4 to 21.6 km/h, full trigger     0.516 s     0.408 s
14.4 to 21.6 km/h, half trigger     1.136 s     0.808 s
Neutral: 21.6 to 14.4 km/h          1.046 s     0.728 s
Settled top speed                 33.624 km/h 33.624 km/h

All 27 checks pass, including the nine power/grip terminal-speed combinations,
no launch overshoot, and forward/reverse neutral stopping. Active brake commands
above 30% exceed neutral drag; corner tests compare coast with 40% and 60% brake
and verify progressive response without abrupt yaw. drivetrain-results.csv now
labels the comparison baseline for every row. These remain model measurements,
not measured race-footage performance.

The in-app preview was observed still displaying v0.3 during this pass. Refresh
the driving browser and verify v0.5 before comparing the new response. Existing
NB4 calibration is kept in the browser where it was saved.

PREVIOUS HANDLING PASS 04 / 14 October 2026

DRIVER FEEDBACK
v0.3 steering and barrier contact were confirmed better. The next request was
snappier acceleration and passive deceleration, keeping the current top speed.

DRIVE RESPONSE
Acceleration punch defaults to 60% extra motor torque at low wheel speed. It
smoothly tapers between 3 and 7 m/s of tire surface speed, reaching zero extra
torque above 7 m/s. The original high-speed motor output and tire forces remain
unchanged. Extra drive still shares the rear tires' available grip with cornering.
There is no hard speed clamp or extra aerodynamic resistance.

Lift-off drag rises from 6% to 18% of maximum brake torque at a neutral trigger.
It fades out over the first 12% throttle, as before. Active brake torque is
unchanged. Neutral braking remains dissipative in forward and backward motion;
it does not command reverse. Stronger drive and lift naturally increase weight
transfer without changing the steering settings or tire-force coefficients.

Quick setup adds Acceleration punch and expands Lift-off drag to 0-30%.
To compare with v0.3, use punch 0% and lift-off drag 6%. Restore baseline returns
to punch 60% and drag 18%. NB4 calibration and controller mapping are unchanged.

MEASURED IN THE MODEL (open ground, baseline car, straight driving)
Test                              v0.3        v0.4
0 to 14.4 km/h                     0.748 s     0.484 s
0 to 21.6 km/h                     1.352 s     0.994 s
0 to 28.8 km/h                     2.490 s     2.130 s
Neutral: 21.6 to 14.4 km/h          1.860 s     1.046 s
Neutral: 21.6 to 7.2 km/h           4.344 s     2.256 s
Settled top speed                 33.624 km/h 33.624 km/h

All nine power (35/78/100%) by grip (1.00/1.65/2.00) combinations converge to
their previous top speed with no launch overshoot. The historical tire tests
below explicitly hold the original drive settings to isolate the unchanged
tire model. New drivetrain checks cover the current defaults separately.
node tests.mjs passes 26 checks in total; node drivetrain-checks.mjs --report
recreates drivetrain-results.csv with the old/current response comparison.
The UI was checked for adjustment, baseline restoration, and keyboard operation.
These response values are tuning choices guided by driver feedback, not fitted
measurements from carpet-racing footage. NB4 driving is the next feel check.

PREVIOUS HANDLING PASS 03 / 14 October 2026

DRIVER FEEDBACK
The v0.2 speed/lift response was closer to the desired feel. Mid-slide grip was
confirmed good. This pass changes steering input and barrier response only;
the tire forces, load transfer, motor, and braking coefficients are unchanged.

STEERING
Baseline travel is now 18 degrees instead of 24 (25% less full-lock travel).
A 35% cubic steering-softness blend reduces small corrections without adding
servo delay. Half transmitter wheel now commands 6.64 degrees instead of 12.
Full wheel still commands the selected travel, and left/right remain symmetric.
Quick setup exposes both Steering travel and Steering softness. Transmitter
calibration and channel mapping are unchanged. Refresh to load v0.3.

BARRIER RESPONSE
The old collision response multiplied all velocity by 0.88 on each contact,
potentially 500 times per second. This erased motion along the edge and caused
sticking. Contacts now use normal impulses at a rounded chassis footprint, with
angular response and low friction bounded by the normal impulse. Glancing hits
carry the car along the barrier; direct head-on hits still stop it. No automatic
track-following or heading snap is applied. Reset remains available with R.

VALIDATION
node tests.mjs passes 20 checks: 9 baseline, 5 tire-handling, and 6 steering/contact.
The tire tests explicitly hold the original physical 2.4-degree wheel angle;
the v0.2 speed, lift, and brake results below remain unchanged at that angle.
Contact checks cover mirrored glances, head-on stops, repeated contact, sustained
wall pressure, and 2,000 impacts around the circuit at varied headings/velocities.
Those impacts conserve or dissipate kinetic energy and keep the chassis inside.
The UI was checked for new defaults, adjustment, and baseline restoration.
These are simulation checks; subjective NB4 feel still needs a driving pass.

PREVIOUS HANDLING PASS 02 / research and tire-model reference

DRIVING TARGET
At a held steering input, the car should follow a wider arc as speed rises. Lifting
should let the nose tuck in; moderate braking should produce more corner-entry
rotation. The transitions should be gradual, with rear stability under power.

IMPLEMENTED
Front/rear cornering stiffness was rebalanced to produce a positive understeer
gradient. Front-wheel angle still follows the transmitter and servo response.
There is no speed-based reduction of the transmitter command or rack travel.
The speed-dependent reduction in curvature comes from the tire forces.

Tire cornering stiffness now changes with normal load, as the tire force limit
already did. Acceleration unloads the front tires; coasting and braking load them.
The existing 56 ms acceleration filter makes that transfer progressive. Gentle
neutral motor drag (default 6% of maximum brake torque) strengthens lift-off feel.
Throttle input fades that drag out. Rear drive/braking still share the rear tires'
finite grip with cornering, so excessive braking can upset the rear.

Lateral load transfer is now limited per axle while preserving the axle's total
weight. Differential coupling reads both wheel speeds at the same instant to avoid
a left/right bias caused by sequential wheel updates.

QUICK SETUP
High-speed push: default 65%. More gives a wider fast arc; less gives more bite.
This adjusts front tire stiffness relative to rear stiffness, not steering travel.
Lift-off drag: default 6%. More adds deceleration at neutral and helps entry bite.
Try changing one control at a time. Motor power and the calibrated NB4 endpoints
are unchanged. Refresh the app to load v0.2; existing transmitter calibration stays.

REPEATABLE SIMULATION CHECK
At a constant 2.4-degree wheel angle, open-ground steady tests produced:
Speed        Radius       Curvature relative to the 7.2 km/h result
7.2 km/h     7.52 m       100%
14.4 km/h    9.51 m        79%
21.6 km/h   12.97 m        58%
28.9 km/h   17.99 m        42%

Starting from the SAME settled 21.6 km/h corner, after 0.25 seconds:
Full power: 23.6 km/h, 17.49 m radius, front axle load 5.75 N.
Coasting:   20.4 km/h, 10.66 m radius, front axle load 6.74 N.
30% brake:  18.9 km/h,  8.18 m radius, front axle load 7.19 N.
The entry comparisons include both speed loss and tire-load redistribution.
These are model measurements, not measurements from footage or a real car.
handling-results.csv contains the numerical output. Run node handling-checks.mjs
--report to reproduce it. node tests.mjs runs all baseline and handling checks.

RESEARCH BASIS AND LIMITS
Mark Pavidis, Basic Suspension Tuning: distinguishes initial turn-in from corner
exit response and discusses smoothing steering on high-bite surfaces. This guide
concerns nitro buggies, so its adjustment directions are context, not numerical
coefficients for this 1:10 2WD model. No camber-link or shock geometry was copied.
https://site.petitrc.com/reglages/BasicSuspensionTuning/

Associated RC10B4 manual, printed pages 21-22 (PDF pages 19-20): recommends calmer
steering options on high grip, distinguishes corner-entry and exit steering, and
describes weight-placement effects with exceptions at extreme grip. This older
2WD buggy is a closer structural reference than the nitro guide.
https://apps.associatedelectrics.com/aereferencelibrary/manuals/9037_B4RS_manual.pdf

MathWorks vehicle model documentation: front/rear cornering stiffness, tire slip
angles, vehicle speed and yaw are coupled. Its tire documentation shows lateral
force curves changing with normal load. This supports the model structure, not
the RC-specific values selected here.
https://www.mathworks.com/help/ident/ug/modeling-a-vehicle-dynamics-system.html
https://www.mathworks.com/help/vdynblks/ug/tiremodel.compute.html

Footage reference: EURO RC SERIES, 2WD Modified Buggy A3 EOS 2025 Daun, GER.
Visually spot-checked early-race frames, including the straight/transition around
2:03-2:08. This is useful for appearance and motion context; it includes camera
panning, occlusion and jumps. No ground-plane trajectory or control inputs have
been measured from it, and it does not establish the numerical steering curve.
https://www.youtube.com/watch?v=PfZJUWRHFbQ
Official class/chassis context:
https://www.eos.racing/2025/01/31/team-associateds-kaerup-wins-2wd-at-eos-2025-in-germany/

Front stiffness, its load exponent, and drag amount remain tunable assumptions.
This pass addresses the requested steering behavior. It is not validation of a
specific Tamiya or modern carpet buggy, and still has no vertical suspension or
traction-roll simulation. A calibrated fixed-camera clip with track dimensions
and, ideally, steering/throttle traces is needed for quantitative video matching.
