Wow! Wasn't really counting on it being that simple. I was going to suggest as a next step to do the 500r pot thing, because ultimately, the computer justs wants to see a voltage from the sender/resistor divider network. If that worked and you figured the voltages that generate a particular tank level (which you did!), I was then going to suggest RC filtering the PWM and following it with a opamp buffer, but it appears the computer is happy enough with the PWM. Could be there is a passive or active low pass filter at the computer's ADC input. Good design practice since cars are a noisy environment.
I am a bit confused though as I thought you'd need an inverted PWM to work.
100%SOC=100% PWM (high from U6 pin 13) drives Q1 low all the time which should be empty tank for you. Maybe I'm missing something here.......well OK, I'm SURE I am
