Full-day highway leg from the Des Moines West KOA (Adel, IA) to the EAA AirVenture grounds at Oshkosh, WI, via I-80 E / I-380 N / I-39 N / US-151 (Route Option 1). Single continuous OBDLink + TrackLogger recording, 8h 14min elapsed (7h 15min moving). Departed on 50A shore power the night before. This was a deliberate Phase A leg (per the Trip Plan's controlled test phases — TrekDrive OFF, ProPower OFF): the NACS umbilical was intentionally left unplugged for the whole leg to get a clean pure-towing baseline, confirmed at 0.0 kW ProPower output across all 14,809 rows. That makes this leg's 15.45 mpg / 27.18 gal over 420 mi a genuine no-generator-load reference point — see 02/03. Weather was warm and building through the day (OAT 72–95°F), consistent with the forecast humidity/heat-index flag for this corridor, but stayed well short of the Extreme Heat Warning conditions that ruled out Route Option 2. Arrival included an extended ~39-minute stationary period at the end of the log, consistent with AirVenture arrival-traffic/check-in congestion flagged pre-trip for this corridor.
420.0 mi / 27.18 gal, 15.45 mpg overall. Confirmed Phase A leg (driver-confirmed): NACS umbilical intentionally unplugged for a clean pure-towing baseline. ProPower AC output (DC/AC Converter A, Output 1 + Output 2) reads 0.0 kW at every one of 14,809 logged rows, consistent with that. State A/B duty is not computable this leg (or any leg) from truck-side OBD alone — the LightShip TrekDrive axle motor is not on this CAN bus; see 04. SOC below is the F-150's own hybrid traction battery (not the LightShip HV pack), shown for context only. Idle rows (gear 70/128/130) filtered from duty/fuel accounting per standing protocol.
| Metric | Value | Note |
|---|---|---|
| Total fuel | 27.18 gal | Trip Fuel PID; integration cross-check 26.99–27.21 gal (<0.7% spread) |
| Distance | 419.99 mi (OBD) / 421.88 mi (TrackLogger GPS) | 0.45% GPS/wheel-speed spread, normal |
| Overall fuel economy | 15.45 mpg | combined, not speed-filtered — pure-towing (Phase A) baseline this leg |
| ProPower delivered | 0.0 kWh AC | intentional — umbilical unplugged, Phase A test |
| State A duty | not available | no LightShip CAN telemetry on this rig (standing limitation) |
| F-150 hybrid SOC start → end | 44.2% → 37.6% | truck's own HV battery, not LightShip; context only |
Confirmed Phase A leg — no State B segment by design. DC/AC Converter A Output 1 + Output 2 Power read exactly 0.000 kW at all 14,809 rows across the full 8h14m log. Driver-confirmed: the NACS umbilical was intentionally left unplugged for this leg to get a clean pure-towing baseline (Trip Plan Phase A: TrekDrive OFF, ProPower OFF). This is not a data gap — it's the intended reference point. This leg's 15.45 mpg over 420 mi is now the trip's pure-towing baseline for comparing against ProPower-on legs — worth pulling into the end-of-trip synthesis alongside the measured V2V baseline (7.70 kWh AC/gal, 22.7% full-chain) as a "what towing costs with the generator off" reference.
Driver-initiated speed sensitivity test, generator off: 62 mph vs. 65 mph, this leg. Driver called out a specific cruise-speed comparison — the US-20 stretch (Eldora-area to Dyersville-area, IA) was deliberately held near 65 mph, above this project's established 62 mph nominal cruise target. Result below.
The 65 mph run was a genuine contiguous, no-traffic segment (see Section 01 track and the driver's own call-out): 143.2 mi at 15.17 mpg (Trip Fuel PID) / 15.13 mpg (integrated cross-check), 64.3 mph moving average, one incidental ~4 min stop mid-segment (fuel/engine-off, negligible fuel effect). No comparably long, comparably clean contiguous 62 mph segment exists elsewhere in this leg — US-20 was posted 65 mph for nearly its entire length here, with only one brief ~4 min dip to a 55 mph small-town zone (not 62). So the 62 mph comparison below uses the same speed-filtering method as the project's own June Prescott–Thermal SAE paper (bin by vehicle speed across the whole leg, integrate distance and fuel within the band) rather than a single contiguous stretch — this draws on far more data (77 min / 2,301 rows vs. a few minutes) and is the more statistically solid of the two, at the cost of not being one unbroken drive.
| Band | Avg speed | Time | Distance | Fuel | MPG |
|---|---|---|---|---|---|
| 60–64 mph (~62 mph nominal) | 62.00 mph | 76.7 min | 79.3 mi | 4.93 gal | 16.08 |
| 63–67 mph (~65 mph nominal) | 65.32 mph | 180.3 min | 196.3 mi | 12.63 gal | 15.55 |
| 64–66 mph (tight band, cross-check) | 65.37 mph | 147.0 min | 160.2 mi | 10.59 gal | 15.13 |
| US-20 contiguous segment (driver-called, for reference) | 64.3 mph | 137.1 min | 143.2 mi | 9.44 gal | 15.17 |
Band-matched result: 62 mph → 16.08 mpg vs. 65 mph → 15.55 mpg, a 3.3% mpg penalty for the ~3 mph step up. The tight 64–66 mph band (15.13 mpg) and the driver's own contiguous US-20 segment (15.17 mpg) agree with each other to within 0.3%, which is a good internal consistency check on the whole approach. Caveats, stated plainly: this is single-leg, single-day data — wind and grade are not isolated (no HRRR wind CSV this leg, see below), so part of the 62-vs-65 delta could be route/weather rather than pure speed²-ish aero effect. It's also not literally the same road for both bands (the 62 mph band pools brief low-band moments from ramps, small towns, and transitions scattered across the whole leg, not one steady 62 mph highway stretch), so treat this as a same-leg first look, not a controlled result. The Cd·A Method-2 speed-hold protocol (50/55/60/65/70 mph, ≥90 s dwell, flat segment, wind-corrected) is the right tool to nail this down properly — worth calling out explicitly on a future flat leg now that the mass basis is locked.
No incremental-fuel table, effective-yield comparison, or divergence-from-7.70 check applies this leg, by design — there's no State B segment to measure. Measured-vs-modeled comparison (Column A ÷7.70 baseline vs. Column B BTE rebuild) is skipped entirely rather than reporting a meaningless 0/0.
Nightly incremental-GPH regression: not run this leg. The regression requires real variation in ProPower on/off state within the leg (fuel_rate = a + b·ProPower_kW + c·grade_power + d·speed³proxy); with ProPower kW constant at zero for all 14,809 rows (by design, this being a Phase A leg), the b term is structurally unidentifiable. Per protocol, stating this and omitting the table rather than forcing a number. Trip ledger columns propower_gph_at_5_8kw_modeled and propower_regression_r2 left blank (not zero) for this leg.
Wind cross-check also skipped: with no ProPower segment there is no measured yield for a wind-adjusted expected value to be compared against. No wind CSV was uploaded with this leg's files either. Separately worth noting: this was a long, largely flat stretch across Iowa/southern Wisconsin — a plausible future Cd·A speed-hold candidate if a dedicated test segment is wanted on a similar corridor, now that the mass basis is locked (15,320 lb / 6,949 kg, CAT-scale measured). Not attempted here without a driver-called speed-hold protocol and wind data.
Not available this leg — standing instrumentation limitation, not a leg-specific finding. The LightShip TrekDrive axle motor, its draw/regen kW, and the trailer's own HV SOC are all on the LightShip's proprietary CAN bus, which this rig does not yet bridge into the OBDLink stream (the OWASYS + CAN bridge is still a funding ask, not installed hardware). Truck-side OBD has no channel that distinguishes State A from State B directly; it can only be inferred indirectly when ProPower is active (State B) — and this leg had none of that either (see 03). No gross/net saving table is presented rather than fabricate one from proxy signals.
Threshold: λ < 0.98 (excluding λ ≤ 0.5 dropouts) AND load > 50%, sustained ≥ 3 consecutive rows (≥ 6 s) to exclude single-sample tip-out/DFCO transients. 78 qualifying sustained episodes found across the day, totaling 588 s (9.8 min) of enriched running and an estimated 0.036 gal total penalty — small relative to the 27.18 gal burned. Episodes are short (median ~6–10 s) and scattered across the whole leg rather than concentrated in one stretch; they read as ordinary grade/load blips (short accelerations, brief climbs) rather than the sustained plains cruise-load enrichment the hot-plains protocol is specifically targeting. Episode OAT ranged 71.6–93.2°F; splitting at the 86°F line gives 40 episodes above (0.011 gal) and 38 below (0.025 gal) — no meaningful OAT-driven jump in penalty, consistent with these being load/grade events rather than temperature-driven ones.
| Time (elapsed s) | Duration | Peak IAT2 °F | Mean λ | Mean OAT °F | Est. penalty gal |
|---|---|---|---|---|---|
| 1,030 | 22.0 s | 109.4 | 0.845 | 73.4 | 0.0078 |
| 12,236 | 6.0 s | 116.6 | 0.958 | 84.2 | 0.0006 |
| 14,474 | 6.0 s | 120.2 | 0.955 | 92.0 | 0.0011 |
| 15,464 | 6.0 s | 116.6 | 0.957 | 87.8 | 0.0009 |
| 17,162 | 10.0 s | 118.4 | 0.956 | 89.6 | 0.0015 |
| 21,360 | 6.0 s | 116.6 | 0.971 | 91.4 | 0.0005 |
| + 72 further episodes ≤ 6–10 s each | 0.0297 (remainder) | ||||
No clean State-A-vs-State-B-at-temperature comparison is possible this leg — there is no State B (see 03) and State A isn't observable (see 04). This leg's λ data is therefore a baseline enrichment-under-normal-towing reference at up to 95°F OAT, not a TurboAssist-suppression demonstration. Separately: peak IAT2 this leg was 143.6°F, briefly exceeding the 140°F Tier-3 throttle-closure threshold — but see 06, this occurred exclusively during a stationary heat-soak at a stop, not while driving, matching the pre-trip prediction almost exactly.