Calibration Test · Controlled A/B Loop

TrekDrive Calibration — State A vs. State B

Two matched loops near Broomfield, CO: Run 1 ProPower ON / TrekDrive OFF (State B), Run 2 ProPower OFF / TrekDrive ON (State A). Direct head-to-head fuel comparison under near-identical route, distance, and elevation.
Test location (approx.): Broomfield, CO — as of 2026-07-14, Run 1 13:47–14:36 MDT / Run 2 14:45–15:47 MDT
01

Run Summary

Run 1 track — State B (ProPower) — 1,453 pts, TrackLogger 1 Hz
Run 2 track — State A (TrekDrive) — 1,411 pts, TrackLogger 1 Hz

This is a purpose-built calibration test, not a distance leg: two loops driven back-to-back over closely matched course and elevation near Broomfield, CO, each isolating one system. Run 1 held ProPower on with TrekDrive off for the full loop (State B); Run 2 held TrekDrive on with ProPower off for the full loop (State A). Both loops start and end within ~0.03 mi of the same point and cover nearly identical elevation range (5,015–5,365 ft), so — per your instruction — no wind correction was applied; the controlled loop design substitutes for it far better than a wind model would on a course this short. Ambient conditions were hot both runs (OAT 93–102°F).

02

Fuel & Charge Accounting

Run-level totals
MetricRun 1 — State B (ProPower)Run 2 — State A (TrekDrive)
Window (MDT)13:47:31 – 14:35:5514:45:13 – 15:46:55
Distance36.67 mi34.43 mi
Elapsed duration0.807 hr (48.4 min)1.028 hr (61.7 min)
Avg moving speed45.0 mph43.6 mph
Total fuel2.765 gal1.486 gal
Fuel economy13.26 mpg23.18 mpg
Fuel rate (elapsed basis)3.428 gal/hr1.445 gal/hr
ProPower AC delivered4.670 kWh (95.7% of moving time on, mean 6.03 kW when on)0.060 kWh (effectively off)
Mean engine power47.23 hp24.86 hp
Mean engine load61.9%31.7%
Idle rows filtered (gear 70/128/130)17 rows (~34 s)460 rows (~15.2 min, 4 blocks — see §08)
Watch

Run 2's elapsed duration includes two extended stops late in the loop (5.4 min and 9.2 min, both with engine off / near-zero fuel burn — see §08). These stops don't materially affect the mpg or gal figures above since idle fuel burn during them was negligible (mean 0.22 gal/hr, mostly zero), but they inflate the "elapsed-basis" gal/hr figure relative to a pure-driving rate. A moving-time-only rate is given in §04.

03

ProPower Incremental Fuel — State B (Run 1)

ProPower ran continuously for 95.7% of Run 1's moving time — the same underdetermined-regression situation flagged on prior legs (e.g. Prescott→Salida): with essentially no ON/OFF pairing at matched speed and grade, this run cannot independently re-derive the 7.70 kWh AC/gal baseline. It's used below only as a consistency check against that baseline, and as an input to backing ProPower's share out of Run 1's total fuel for the Run 1 vs. Run 2 comparison in §04.

Run 1 fuel, applying the KB baseline
QuantityValueBasis
ProPower AC delivered4.670 kWhmeasured (DC/AC converter output × Δt)
Implied ProPower fuel share0.606 gal4.670 ÷ 7.70 kWh/gal — KB baseline, modeled
Implied propulsion-only fuel, Run 12.159 gal2.765 − 0.606, modeled
Implied propulsion-only mpg, Run 116.99 mpg36.67 ÷ 2.159, modeled
Implied propulsion-only rate, Run 12.675 gal/hr2.159 ÷ 0.807 hr, modeled
Nightly incremental-GPH regression — attempted per protocol
TermValue
ProPower coefficient (b)0.178 gal/hr per kW
Implied rate at 5.8 kW1.03 gal/hr
0.27
Mass basismeasured — 6,949 kg (CAT-scale tare + independently weighed cargo, per instruction set)
Flag

R² = 0.27 is below the ~0.4 reliability threshold, and the implied 1.03 gal/hr at 5.8 kW sits above the 0.4–1.0 gal/hr physically plausible band. Root cause: with ProPower on 95.7% of the time, the regression has almost no ON/OFF contrast to separate ProPower's effect from grade and speed — exactly the underdetermined case the protocol warns about. Not presented as a corroboration of anything; the measured round-trip baseline (0.64–0.65 gal/hr, range 0.55–0.75) remains the reference figure.

04

TrekDrive Fuel Savings — State A (Run 2)

Run 2 held TrekDrive on with ProPower fully off for the entire loop, over a course and elevation profile matched closely to Run 1. This is the calibration's core deliverable: a direct, controlled comparison of total drivetrain fuel burn with vs. without TrekDrive assist, on a loop distinct from (and slower/more stop-and-go than) the highway-cruise conditions the KB model and 62-mph baseline are built on.

Run 1 (backed-out propulsion-only) vs. Run 2 (TrekDrive), gross comparison
MetricRun 1 propulsion-only (no TrekDrive)Run 2 (TrekDrive)Δ
Fuel rate2.675 gal/hr1.445 gal/hr−1.230 gal/hr (−46%)
Fuel economy (propulsion)16.99 mpg23.18 mpg+6.19 mpg (+36%)
Fuel per mile0.0589 gal/mi0.0432 gal/mi−0.0157 gal/mi (−27%)
Mean engine power47.23 hp (Run 1 total, both propulsion+ProPower)24.86 hp−47%
Mean engine load61.9% (Run 1 total)31.7%−49%
Flag — net calc not possible this run

Gross only. The protocol calls for gross fuel saved paired against the trailer battery cost TrekDrive drew to achieve it, expressed as a fuel-equivalent net. This OBD dataset has no LightShip/trailer SOC or TrekDrive draw-kW channel — that telemetry lives on the trailer's own CAN bus, not the truck's. Net savings cannot be computed from this data. Recommend capturing LightShip Atlas SOC (start/end, or ideally a screen-record) on future calibration runs so a battery-kWh-spent figure can be paired against the gross fuel gal saved.

Modeled context

The KB TurboAssist model (highway 62 mph basis) gives State A 1.73 gal/hr and State B 4.28 gal/hr (ProPower-inclusive). This loop ran ~44 mph with two long stops, not sustained highway cruise, so it is a different operating regime, not a re-confirmation of those figures — Run 2's 1.445 gal/hr and Run 1's 2.675 gal/hr backed-out propulsion rate are both lower than the highway-basis numbers, consistent with lower average road load at lower average speed. The directional finding — TrekDrive substantially cuts engine load and fuel rate relative to unassisted towing — holds and is, if anything, cleaner here than in the highway model, since this is a genuine two-run A/B rather than a single-leg estimate.

Moving-time-only comparison (excludes Run 2's two long stops)

MetricRun 2, elapsed basisRun 2, moving-time only
Duration1.028 hr0.775 hr (61.7 − 15.2 min idle)
Fuel rate1.445 gal/hr≈1.92 gal/hr

Idle-period fuel burn during the two stops was negligible (near-zero, engine off), so total gal consumed is essentially unaffected — only the rate denominator changes. The moving-time rate (≈1.92 gal/hr) is the more representative one for comparing driving effort; the elapsed-basis rate (1.445 gal/hr) is what a trip-level mpg/gph figure would show. Still well below Run 1's backed-out 2.675 gal/hr either way.

05

Heat-Enrichment Fuel Penalty — Measured λ

Filtered to sustained rich λ (commanded λ < 0.98, load > 50%) to exclude transient tip-out/restart excursions. Both runs show small, comparable enrichment penalties — this test wasn't run in the sustained high-load, high-IAT conditions (e.g. mountain grades, hot plains at speed) where the effect is largest.

Enrichment penalty by run
RunRows λ<0.98 (sustained)Peak IAT2 during episodesMean enrichmentEst. penalty
Run 1 — State B66 rows / 132 s143.6°F3.7%≈0.005 gal
Run 2 — State A76 rows / 152 s145.4°F5.5%≈0.006 gal
Note

Run 2 additionally shows a much higher rate of brief (transient, load < 50%) rich-λ excursions overall — 50.9% of rows vs. 18.0% in Run 1 — consistent with the frequent auto start-stop engine restarts during Run 2's idle blocks (each restart briefly commands rich, cold-start-like enrichment). These are excluded from the sustained-penalty figures above by the load filter and are not a TrekDrive thermal effect; they're a start-stop-cycling artifact of this run's stop pattern.

06

Thermal & Trim Watch

07

Instrumentation Check

08

Flags & Carry-Forward

RIG: Ford F-150 PowerBoost + LightShip AE.1 · INSTR: OBDLink MX+ (42-PID, 2.0 s) + TrackLogger Pro 1 Hz.
Location approximate, city-level, from GPS track; precise coordinates not published.
Measured values green, modeled amber, flags rust — modeled numbers are never presented as measurements.
Source CSVs: CSVLog_20260714_134700.csv (Run 1) · CSVLog_20260714_144451.csv (Run 2) · TrackLogger 2026-07-14_13_47_05.csv / 2026-07-14_14_44_49.csv · generated 2026-07-14.