The meter is basically a little "box" with a fuel input
> and output on it and it will give an electronic "pulse" per every unit of
> fuel that flows trough it.  The gauge attached to it counts the pulses and
> determines fuel flow from the amount of pulses over time.

I'm not familiar with how the TDI engines control injection, but if it's
electronic and works like gasoline FI, then the best way to monitor fuel
usage is via injector pulse timing. Flow rate on the injectors ought to be a
fairly stable constant, so by taking the cumulative length of time the
injector is open and multiplying by the flow rate and then multiplying by
the number of injectors you can get a much more accurate measurement of fuel
consumption than trying to juggle the often minuscule differential between
two flow meter inputs. If you're already working with a microcontroller for
the instrumentation control, it's not too big a deal to whip up a little
subroutine in the main loop or an interrupt to add up FI pulse time. Even if
you don't know the flow rate of the injectors you can make up a calibration
routine that blindly counts total pulse time and, when you refill the tank,
takes the number of gallons/liters used and calculates the flow rate from
that.
I guess the real question is, how is the gauge built? If it's a totally
scratch-built unit and you or someone else have written the software,
modifying the flow sensor input routine should be a snap. If it's something
off-the-shelf and it's built to expect those flow meter pulses specifically,
you got a tougher row to hoe. If that's the case, you'd need to build some
sort of differential counter, and at that point you might as well just hook
an LCD to it and make your OWN gauge unit.

--
John Bange
'90 Vanagon - "Geldsauger"