Looks like Gary Casey beat me to the criticism of conventional
capacitive fuel-level sensors.
Like Gary, in my day job I design and build capacitive sensors. We
sell about 100 million units
per year. So we've seen failure modes that wouldn't show up in any one
person's experience.
If you really want to find out what can go wrong with a device, put
about a million of them
in the field operated by people who have no idea how they work. Some
of you are probably
helping test our products right now :-)

Like Gary, I took apart an "aviation-grade" capacitive fuel gauge and
was horrified to see how
badly it is engineered. One drop of water in the wrong place looks
just like a tank of gas to the
"aviation-grade" sensor.
Changes to the aircraft bus voltage also affect the fuel level reading.
So when someone says they "know" their sensor is accurate, I believe
that they knew it
was accurate the day they checked the calibration. With the clean fuel
that was in the tank at the time.
At that temperature. When their voltage regulator was working right.
And so on...

Like Gary, I know I can build a better fuel level sensor, but I
haven't built one for my aircraft, and probably
won't get around to it anytime soon. It would distinguish fuel
composition as well as level so that
water or alcohol would be detectable. It would operate on a few mA or
less at any voltage between 3V
and 30V, and provide digital serial data output. Data could including
warnings if water is detected.
If anyone wants a fuel sensor like that, let me know what you think it is worth.

The fuel gauge I trust most is the sight tube on my header tank. And
even that can give
erroneous readings. A clog at the top of the tube could cause it to
show full fuel even as the
actual level approaches empty. Cross-check with measured fuel
consumption through the
fuel-flow meter.

BTW my 235 carries 33 gal in three tanks. Practically all of it is
usable if the wings are exhausted
first, then the header. My practice is to climb on the header, then
run the wings to exhaustion at
high altitude. Any residue in the wings is pumped to the header. The
flight continues with some
certainty that all of the remaining fuel is in the header and can be
used with a complete electrical
failure. If the fuel selector valve broke, that would be a serious
problem, as that could leave me at 12000'
with 11 unusable gallons in the header tank. If that ever happens,
I'll be using my glider rating for
the rest of the flight. Depending on the flight plan, I take off with
at least 8 gallons in the header
and enough to get there plus reserves. I do not always take off full.

I also fly a Maule M5 that can carry 72 gallons (9 hours without reserves).
I rarely fill those tanks, except when tankering fuel into the
wilderness. Half full is plenty
for most flights with reserves. The forty gallons I'm not usually
carrying is good for 240 lbs
of payload capacity, or about 200-300 fpm more climb rate.




-bob mackey