Same truck, same trailer, same driver, same load, same road, north then south. The pump and the engine's own fuel meter agree on every leg except the northbound night run, where eleven litres go missing in the same forty miles.
I run the transport side of a small haulage firm. Fourteen artics, mostly food, mostly nights, because the roads are empty and the delivery windows are at four in the morning.
My whole job is pence per mile, and fuel is most of it, so I want to explain how a fuel figure is actually arrived at, because it is not a man writing down a number.
Two measurements, completely independently.
The first is what went in. The truck fills on our own bunkered diesel at the yard, and the pump is a metered commercial pump with a card reader. It records litres against a vehicle and a driver and it is calibrated annually because it has to be for trade use.
The second is what was burned. Every engine built in the last twenty years reports fuel consumption on its own data bus, and the telematics box reads it and sends it to me. The engine knows how long it held each injector open and how many times, and that is not an estimate, that is the thing that actually happened.
Those two should match over a tank. They do, across my fleet, to within about one and a half per cent, which is the error you expect from temperature and the dregs at the bottom.
Unit 6 is a four year old tractor unit on a fixed night trunk. Loaded out at nine in the evening, forty one miles north, tip, reload, forty one miles back, in the yard by half one.
The same road in both directions. A dual carriageway with one long climb and one long descent, so the two legs are mirror images of each other and the fuel burn should be lopsided in a completely predictable way: more going up, less coming back.
It is lopsided in a completely unpredictable one.
Southbound, the engine's figure and the pump agree. Northbound, the engine reports about eleven litres less than the arithmetic needs.
I need to be careful about what that sentence means, because it sounds like the truck is saving fuel and that is not what is happening.
Over a full night the pump says the truck used, for argument's sake, a hundred and fifty two litres. The engine's own total for the same night says a hundred and forty one. Eleven litres are in the tank's accounting and not in the engine's.
Split by leg, the southbound half matches. All eleven of them are missing from the northbound leg.
Eleven litres in forty one miles is not a rounding error. It is about seven per cent of the night and it is the difference between a profitable trunk and one I would cancel.
What I checked, in the order any transport manager would.
Theft. This is the first and only thing anybody in haulage thinks about, and we have had it before on another unit, and it looks completely different: a tank level that drops while the vehicle is stationary and the engine is off. The level sensor on unit 6 is flat every time it is parked, all night, every night, for two years of logs.
The driver. Same man for three years, and I will say plainly that he is the best I have and the least likely to be at anything. I also put a different driver on it for a fortnight. Same eleven litres.
The truck. Swapped the trailer, which changes the weight and the aerodynamics. Same. Put unit 6 on a different trunk entirely for a month: the pump and the engine agreed on everything, all month, both directions.
Put a different unit on the night trunk for a month. Unit 9, older, thirstier. The pump and the engine agreed. Both directions. No eleven litres.
So it is not the truck and it is not the driver and it is not the trailer. It is unit 6 on that run, northbound.
Then the manufacturer got involved, because that is what a dealer does when you go in with a fuel discrepancy and a warranty.
They put their own logger on it for six weeks and took the data away and came back with a report I still have. It says the engine's fuel reporting was verified against a flow meter fitted in the feed line and found accurate across the whole range, and that no fault was found.
Which means the engine is telling the truth about what it burned.
And the pump is calibrated and stamped and tells the truth about what it dispensed.
And the tank is not leaking, because a leak is eleven litres on the ground and there is no ground in forty one miles of dual carriageway that has eleven litres of diesel on it twice a week for two years.
The part I cannot get past.
I went out with him. Twice, in the passenger seat, which a transport manager doing his job properly does anyway and which I had not done in four years.
Both nights the engine did the work. You can feel the climb and you can hear it and the gauge moves the way it should move. It is an entirely ordinary forty one miles with an entirely ordinary lorry under you, and at the top of the hill you come over the crest and the road drops away and you can see the lights of the town fourteen miles off.
The second night I had the diagnostic laptop on my knees, plugged into the port, watching the instantaneous fuel rate because I wanted to see where in the forty one miles the shortfall was.
It is not anywhere. The instantaneous rate is sensible for the whole leg: high on the climb, low on the descent, nothing odd, nothing zero, nothing negative. I integrated it myself afterwards on a spreadsheet and the total comes to the engine's total, which is the short one.
So the engine burned a hundred and forty one litres, continuously, plausibly, the whole way, and a hundred and fifty two went into the tank, and the difference is not on the road and not in the tank and not in the engine.
What I do with it is nothing, and I am aware of the shape of that.
The trunk is profitable on the pump figure, which is the figure we bill and budget against, so the eleven litres cost me money and I have absorbed them for two years. Eleven litres twice a week is about fourteen hundred pounds a year and I have stopped raising it, because raising it means another six weeks of somebody's logger and another report saying no fault found.
One thing has changed. I used to tell the lad to take his break at the top, in the services, because it is the sensible place and the food is not bad.
He mentioned, once, in passing, that he had started taking it at the bottom instead, before the climb, and when I asked why he shrugged and said he would rather do the hill in one go.
I said that was fine. It is fine. It makes no difference to the hours or the delivery window.
I have not told him that the eleven litres go missing on the climb, because I do not know that, because the instantaneous trace says they do not go missing anywhere at all.