You don't need to do two flights to get accurate fuel consumption data. The PA28 series has a L - R - OFF fuel selector. Start with the tanks filled to a known, easily seen point eg the tabs or the filler neck or a convenient scratch or whereever. Start, taxi, take off & climb to your typical cruise altitude (better slightly higher than lower) then set *leaned* cruise power on one tank. Note the time and change to the other tank and cruise for an hour or, better, two. Note the time & change back to the start tank for descent & landing.
Refuel to the original level & note what each tank used. You now have the data to calculate fuel consumption from start to landing excluding cruise and, separately, leaned cruise. If you apply the calculated leaned fuel rate to the entire flight duration you will have a figure that represents how much fuel would be used if the aircraft had started above the airfield at cruise altitude/configuration/power setting, and arrived above the destination the same way. Compare that extrapolated fuel burn with the total fuel burn for the flight to find the additional amount of fuel that is used from start to top of climb and top of descent to landing.
If you have multiple tanks eg mains and auxillaries and are without an accurate fuel flow system it's possible to compensate - slightly - and have at least two points where the amount of fuel on board is known accurately. First point is prior to departure with full tanks or filled to a known level. The second point is running the alternate tank dry or very nearly dry. The principle is that if you have two or more containers (instead of a single large container) you will have a known quantity remaining in the remaining, unused containers if consumption is confined to a single container at a time until it is empty.
Provided the time from start to when you switch to the alternate tank is relatively short then what's left in the mains will be very close to the calculated amount after subtracting estimated fuel burn. Over such a short period of time the margin of error is small.
If you have a good idea of cruise fuel flow eg the proving fuel consumption flight discussed above, then the time to empty on the Aux can be anticipated to within a few minutes. At the point where the Aux runs dry or you decide it's as close to empty as you're willing to go, you have a pretty accurately known quantity of fuel on board in the remaining tank(s). If not run dry there will be a margin of error but it doesn't affect the accuracy of the known fuel in the other tanks.
If you have to switch between pairs of aux's to maintain lateral balance then for this purpose you would consider the pair to be a 'single' tank and known fuel on board accuracy will be when both aux. tanks are empty.
Last edited by Tinstaafl; 22nd April 2009 at 16:43.