Trailer Load & Thermal
The other pages on this site follow energy toward the trailer. ProPower System covers how the truck makes it; The Measured Case covers what it costs to move it. This page covers what happens to it once it arrives — which is the part that decides whether a given source is useful or merely present.
The short version: a trailer is not a passive battery. It runs a refrigerator, a 12 V system, and a thermal-management plant that operates whether or not you are charging. Those loads are served first. The cells get the remainder.
Subtraction, Not Percentage
House load is not a proportional loss like a conversion inefficiency. It is a fixed draw removed from whatever the source delivers, and what accumulates in the pack is the difference.
That distinction is the entire subject. A fixed draw taken off a large source is a rounding error. The same draw taken off a small source can be most of it.
It also means the familiar way of comparing charging sources — Level 1 is slow, Level 2 is faster, roughly in proportion to their ratings — is wrong for a trailer. The proportionality breaks because the subtrahend does not scale.
How Big Is the Subtraction?
This is the softest number on this site, and it is presented that way deliberately. What follows separates what was measured from what was observed from what cannot currently be measured at all.
Across Campaign 1 the truck-side converter held three distinct output states: 0.000 kW with the umbilical unplugged, 0.938–0.951 kW connected but not bulk charging (six independent legs), and 5.69–6.06 kW delivering (seven legs). Those magnitudes are firm and repeatable across 3,719 instrumented miles. Campaign 2 (September 2026, 46 PIDs at 1 Hz) reproduced the delivering state at a 5.85 kW mean and, for the first time, metered the inverter's input side: 22.0 A at 272 V in against 5.85 kW out, an inverter ratio of ≈ 0.98 at bulk. The rate never derated with heat in either campaign.
They are measurements of truck output — what left the inverter, not what the trailer consumed. And the unplugged state is not zero: with the output channel reading 0.000 kW, the input side still carried ≈ 0.13 kW — the truck's own bed and cab loads on the same circuit (EVSE standby, satellite terminal, router, refrigerator, device charging), nothing on the trailer.
With trailer loads shed as far as they can be shed, the connected-not-bulk-charging delivery has been observed as low as 200–300 W. On Campaign 1's hot-weather legs it ran up to roughly 900 W. That range is a truck-side reading of what the umbilical carried. It is not the trailer's house draw, which is larger: with the air conditioner running across the whole drive the house draws on the kilowatt-hour-per-hour scale (about 3 kWh/hr observed at camp with the canopy up; lower in motion with it down), and the pack supplies the balance — about 3–4 % of the pack per leg in motion, which is what drives the trailer's charge-limit cycling on hot legs.
The reading that the middle converter state tracks demand is operator judgement, not an instrumented result, and it is under test. Two things argue against it: 0.938–0.951 kW is a very narrow band for something supposedly following a variable house load across different ambients and altitudes, and the Campaign 2 input-side channel disagrees with it outright — see the flag below. An earlier revision of this page stated flatly that it tracks demand and is not a fixed baseline. That is withdrawn to attributed, under test.
| Condition | Draw | Basis |
|---|---|---|
| At camp, hot ambient, canopy up | ≈ 3 kWh/hr | HVAC plus house; Campaign 1 stops at 100 °F and above |
| Overnight, lived in | 1–2+ kWh/hr | HVAC alone; arriving near full is the practical priority |
| In motion, canopy down, HVAC running | ≈ 3–4 % of pack per leg | Drives charge-limit cycling on hot legs (10–73 min cycles), trailer battery-management logic |
| 12 V rail base load, between charge cycles | 7–14 A | Operator-added 12 V battery monitor, hand-read; roughly 90–190 W at 13.3 V (arithmetic, not a reading) |
Three limits keep this from being pinned down with the rest of the campaign data.
The instrument has a floor. The truck's ProPower output channel does not register loads at roughly 200 W and below. Campaign 2 put numbers on it: the output channel reads 0.000 kW while the input channel shows 0.455 A flowing, and a deadband state reports 0.106 kW out against 0.177 kW in. The bottom of the range is therefore partly inferred from behaviour rather than read off a channel.
The two truck-side channels disagree in the connected-not-bulk-charging state. At bulk charge the output and input channels agree to about 2 %. In the middle state the output channel reports 0.940 kW while the input side carries only 0.720 A at 277 V — 0.199 kW. A line through the two trustworthy states predicts about 0.08 kW at that current; the reported figure is more than eleven times it. This is not a sampling artefact: windows mixing the two states interpolate correctly on both channels. Three explanations are live — the 0.940 kW is a status or setpoint value rather than measured power; the input channel under-reads in that state; or a second inverter stage exists that the input channel does not see. The fuel channel cannot adjudicate, because the on/off legs available so far sit on different grades and the grade term is about twice the signal. If the first or third holds, the load-support line elsewhere on this site is wrong in magnitude and possibly in destination, and this page's subtraction argument is being made with a number that may not be a power.
There is no logged trailer-side channel. The OBD stream carries what the truck sent, not what the trailer consumed or stored. The program's only trailer-side instrument is a 12 V battery monitor the operator added — state of charge and battery current, read by hand, not logged — and nothing sees the high-voltage pack. Attributing the difference to house load remains operator judgment supported by circumstance.
Working figure until that is closed: house base load is somewhere between unmeasurable and about 1 kW, condition-dependent — and the 1 kW end now carries its own question mark. Any number quoted more precisely than that is quoting one leg's weather.
Why It Is Not Controllable
The obvious response to a standing draw is to turn things off. The trailer's own circuit layout explains why that does not work, and it is worth walking through because it converts the base load from a mystery into an inventory.
The AE.1 carries two 12 V distribution boxes, 47 protected circuits between them. Sorting those circuits by whether an operator can switch them off produces a lopsided result.
| Category | Circuits | Operator control |
|---|---|---|
| Always-on — control modules, monitoring, safety | 21 | None |
| Thermal & battery management — automatic | 14 | None |
| Discretionary or event-driven | 12 | Partial |
Thirty-five of forty-seven circuits have no operator control at any point — not in the trailer's app, not on a panel, not anywhere. And the split is not incidental: the thermal block contains essentially every high-current circuit on the trailer.
The panel shows three separate coolant pumps, two radiator fans, and two interior blower fans, alongside coolant reversing valves, expansion valves, a coolant valve, four pressure transducers, a dedicated thermal-system controller, and two battery-monitoring units.
All 12 V. All thermostatically commanded. None switchable. This is the mechanism behind the observed variance — the devices drawing meaningful current on the 12 V rail are cooling equipment, and cooling equipment runs when it is hot. Base load rises with ambient because the things riding on it exist to move heat.
Reversing valves plus expansion valves plus four pressure transducers describe a reversible refrigerant circuit — a heat pump, not a one-way air conditioner. Three separate coolant pumps imply at least three thermal loops (battery pack, power electronics, cabin) with valve-commanded cross-connection between them.
The practical consequence: there is no state in which the thermal system is simply off, because pack conditioning and cabin conditioning share hardware. Reducing cabin cooling does not stop the pumps if the pack still wants conditioning. This mirrors the tow vehicle's own arrangement, where the compressor also serves battery chiller duty and cannot be fully commanded off.
Strip out the circuits that only draw momentarily — waste valves, canopy actuators — and the drive-motor circuits that only exist while towing, and the entire operator-controllable standing load is USB ports and LED lighting.
That is the finding. Turning things off in the trailer's interface addresses a rounding error, because everything with real current is either always-on or thermally commanded. Note also that the onboard charger and the DC/DC converter are themselves in the always-on block: the charging hardware is part of the base load.
The trailer's 12 V house battery is a 100 Ah lithium pack, charged from the high-voltage pack through a bidirectional DC/DC converter on a 40–70 % state-of-charge schedule at a measured 68.1 A — about 920 W while it runs, roughly half an hour at a time, every three to five hours. Between charges the 12 V base load cycles 7–14 A as pumps, valves and electronics come and go. The same converter runs the other way to harvest solar: the three solar controllers sit on the 12 V bus, and surplus is converted up to the high-voltage pack — but only after the 12 V battery is satisfied to 95 %.
Two consequences. The 12 V charge cycle is served from the high-voltage pack, so it is not a direct load on the umbilical. And the whole 12 V side is weather-driven: with sun on the panels the DC/DC is suppressed and the cycle disappears; after dark it runs on the pack. Solar itself is nameplate 1.8 kW with an observed ceiling of about 1.1 kW — the cells are bonded down with no air behind them and derate as they heat, so the shortfall is widest exactly when insolation is highest.
68.1 A at a lithium charge voltage of about 13.8 V is 940 W — the connected-not-bulk-charging figure to three figures, and a constant-current charger would explain the suspiciously narrow 0.938–0.951 kW band far better than a demand-follower does. But the 12 V charger is fed from the high-voltage pack, not from the umbilical, so the match is most likely coincidence. It remains possible that the trailer pulls about 1 kW from the umbilical to offset that drain, which would make the load-support state real and timed to the 12 V charge cycle. That is exactly what the test in § 06 is designed to see.
What This Does to Charging Rate
Carry the observed range through to what actually accumulates in the pack, and the asymmetry is the point.
| Source | To the trailer | Net at 0.2 kW base | Net at 0.9 kW base | 10 kWh takes |
|---|---|---|---|---|
| Level 1, 120 V | ~1.2 kW | 1.0 kW | 0.3 kW | 10 to 33 hr |
| V2V, measured on this rig | 5.8 kW | 5.6 kW | 4.9 kW | 1.8 to 2.0 hr |
| 240 V export from an EV truck | 7.2 kW | 7.0 kW | 6.3 kW | 1.4 to 1.6 hr |
Source figures are measured or rated; base-load figures span the observed range; the subtraction and resulting times are arithmetic. Charging is treated as linear, which it is not near the top of the pack, so these are best-case durations at any given base load. No conversion losses beyond the delivered figure are applied.
Read the last column rather than the middle two. At 5.8 kW the entire uncertainty in house load moves the answer by about ten minutes across two hours — a detail. At 1.2 kW the same uncertainty is the difference between an overnight top-up and something that runs for a day and a half.
This is also why the connected-but-not-bulk-charging state would be worth having rather than switching off — if the attribution holds. On that reading it is not filling the pack, but it is holding the house loads off it, which matters most in exactly the conditions where the house load is largest and the pack can least spare the energy. If the Campaign 2 channel disagreement in § 02 resolves the other way, the state is a status flag rather than a transfer, and this paragraph goes with it.
The Edge Case That Is Not Rhetorical
Everything above is a question of convenience until the pack approaches empty. Then it becomes a question of whether the trailer can be recovered in the field at all.
The failure is a cascade rather than a single event. The high-voltage pack depletes; the battery management system declines to close its contactors; the DC/DC converter that feeds the entire 12 V system loses its source; the 12 V system goes flat. At that point nothing remains to wake the onboard charger, and the trailer cannot accept a charge even with a perfectly good source standing next to it. It is not discharged. It is unresponsive.
Nor is a public fast charger a way out: DC fast charging is not enabled on delivered units (LightShip; no date given), so every recovery runs through the AC onboard charger. That is a service call, and by the argument on the Charging Infrastructure page it happens where service calls are hardest to obtain.
| Tow vehicle export | Recovery capability |
|---|---|
| 120 V only, ~1.5 kW cap | Marginal at best. Net to pack may approach zero in heat. |
| 240 V from an EV tow vehicle | Real, but drawn from the traction pack that is also your range. Recoveries are a limited resource. |
| 240 V from an onboard generator | Real, and refills in ten minutes at a pull-through lane. Bounded only by fuel. |
The distinction in that last row is not efficiency. It is whether the recovery capability is itself limited. An EV tow vehicle rescuing a flat trailer is choosing which half of the rig to strand; a generator-equipped vehicle is not choosing at all.
Whether the trailer's onboard charger engages at all from a 120 V source, and at what minimum current, has not been established by this project. That answer determines whether a 120 V-only tow vehicle is a weak lifeboat or no lifeboat. It is a bench test, not a campaign. It has not been run as of this revision (September 2026).
Also unresolved: recovery attempts happen disproportionately in heat, and the one confirmed temperature-driven charging stop in this project's data occurred on a V2V session, not at a public charger. A lifeboat that derates is still a lifeboat, but the recovery plan should assume it might.
What Would Close This
The base-load range is wide because it has never been measured directly — only inferred from the truck side, through an instrument with a blind spot at the bottom of the range and, as Campaign 2 showed, two channels that disagree in the state that matters. That blind spot is a tow-vehicle artifact and has nothing to do with the trailer's own wiring, which means it can be stepped around entirely.
- Whole-rail baseline. The operator-added 12 V battery monitor now gives 12 V current and state of charge, which is most of what a clamp on the main feed would have given. What is still missing is a log: readings against ambient across a warm day would convert “200 to 900 W” from a range into a curve with a mechanism attached. The high-voltage side of the house load — the compressor, the onboard charger's own standby — is still unseen.
- Per-circuit isolation. Measuring at individual fuses would confirm or refute the hypothesis that the three coolant pumps and two radiator fans drive the variance.
- Matched charging windows — retired as designed. The planned test was twenty minutes in Camp mode against twenty in Storage mode with the umbilical connected, to see whether delivery followed the house load. It cannot work. The 12 V charge cycle runs about 920 W for half an hour every three to five hours, so a twenty-minute window samples one phase of a process longer than the whole test. A simulation dropping such a window at a random phase of that cycle returns a mean near 230 W with a spread of about 180 W; a paired Camp/Storage difference would carry roughly 250 W of noise against a plausible 200–400 W signal. It would return a number that looked like a measurement and was a coin flip. Interleaving ten-minute blocks for four hours gets the noise down to about 70 W — a rest day spent on the campaign's lowest-priority open question — and sun on the panels would still shift the answer within the day.
- The replacement: watch the 12 V charge cycle, after dark. Parked, umbilical connected, settled in the connected-not-bulk-charging state, log three to five hours of both truck-side channels and read the 12 V monitor every half hour. The 12 V charger is a free stimulus of 700–920 W several times a night; at the bus voltage that is +2.5 to +3.3 A on the input channel against a 0.7 A baseline, unmissable if it is there. If the output steps to 0.94 kW and the input steps with it, timed to the 12 V charger, the load-support state is real and more specific than claimed. If the output steps and the input does not, the output channel is not measuring power. If the output sits at 0.94 kW regardless, it is a status value and the attribution retracts. After dark, because sun suppresses the DC/DC and removes the stimulus.
The retired test is the second time this program specified a measurement window shorter than the process it was measuring; the first was a thirty-minute state-of-charge capture that would have returned less than one display quantum. Same root cause both times, and the rule that comes out of it is short: before setting any window length, state the period and amplitude of the slowest process that contaminates it.
The target output is a single base-load-versus-ambient curve. That curve would let any owner of any electric trailer compute net charge rate for a given source in given conditions — which is the question this entire page is circling, and the one no manufacturer currently publishes.
Everything on this page is reconstructed from the outside — from truck-side telemetry, factory panel layouts, and inference. The builder of any electric trailer holds the parasitic-load inventory and the thermal control logic directly, and could publish a load-versus-ambient curve without instrumenting anything. It is unglamorous documentation, and it would materially change how buyers evaluate tow vehicles.
If you have questions or comments, please visit the LightShip RV Owners Club on RVForums.com and post them to this topic. I monitor that site regularly and would enjoy your feedback.
