Sometimes it's hard to resist the urge to fix something that's not broken. The stock fuel sensors on Van's aircraft are float-type resistive or possibly magnetic angle sensors (why is the word "sender" typically used? did someone mis-speak years ago and it stuck? weird.) These work ok, but have a few issues I'd like to address.
About 1/3 of the capacity cannot be measured. Because of the tank shape and wing dihedral, the float is held against the top of the wing for a portion of the tank's capacity.
Float sensors are not very accurate.
Ok, on to capacitive sensors. Every product on the market that I found for sale used a single capacitor to sense depth. If all you're going to use is 100LL this can work ok, but the moment you start using auto gas, especially with non-trivial amounts of ethanol, the calibration will be completely wrong. This is due to the low dielectric constant of gasoline relative to ethanol, especially if the ethanol has absorbed some amount of water.
To address this we will be adding a capacitor plate at the bottom of the tank to sense the dielectric constant of the fuel.
In order to be able to sense the top 1/3 of the tank that would typically just show as "full" we'll add a plate at the "top", outboard past the fuel cap.
This brings us to the circuitry. Dynon sells a capacitance-to-voltage converter for the now-unavailable Van's capacitive plates, but it doesn't have the compensation input for the plate sensing the dielectric constant. Oh well, I'll just make my own.
Below we have a custom 4-input capacitance sensor along with an AVR and UART-to-RS232 converter. This 28-bit resolution chip shows a repeatable change of about 140,000 counts using a test sensor submerged in vegetable oil which has a dielectric constant similar to gasoline.
More updates on the in-tank plates to come later...