[MetroMPG]

There's a post on GasSavers that member krousdb wrote (he's a member here too, inactive now). I'll see if I can find it.

In a nutshell though: he connected his multimeter to the fuel gauge sending unit on his car and measured its resistance, if I recall.

Effectively: he made a digital fuel gauge with vastly better resolution & accuracy than the analog OEM gauge.

He started from an empty tank, then recorded the readings as he progressivlely added a small measured amount of fuel (quarter gallon at a time?) until the tank was full.

He made a plot of multimeter reading vs. fuel volume from empty to full.

Since he always parked in exactly the same place each night (tank in the same attitude as when he calibrated it), he could then fairly accurately measure daily fuel consumption from his commute by comparing the multimeter reading to the plot he made from the calbration exercise.

The main thing that comes to mind that would mess it up (slightly) would be thermal effects on fuel volume.