In an earlier thread, we discussed existing low fuel alarm sensors, and I mentioned that I found an alternative that might work. It's an ultrasonic liquid sensor that detects the presence or absence of fuel by shooting high-frequency sound pulses through the wall of the tank and measuring reflections. I bought the sensor from this Amazon listing, which was the lowest price I could find, by half.

The sensor is 1.1" diameter x 0.75" thick. It has an 18" long 4-conductor pigtail with a 4-pin connector (JST XH-series). The four wires coming from the sensor are:

  • black (ground)
  • red (12V)
  • yellow (active high output)
  • white (active low output)

When the sensor detects liquid, the active high output provides 12V and the active low output pulls to ground. It comes with three small plastic vials containing acoustic gel for testing, epoxy resin for mounting and hardener for the epoxy.

DS1603NF.jpg SensorEpoxy.jpg

I plugged most of the bungs on a Kitfox aluminum header tank and filled it ~3/4 full of water, then I mounted the sensor to the back of the tank with acoustic gel and tape. During testing I tilted the tank along the plane of its back face to move water behind the sensor. The sensor is rated for a power supply of 10-36V. I powered it with 12V from a bench supply and it consumed 13 milliamps (0.013 amps).

It doesn't like to be mounted near the edge or corners of the tank; I'm guessing that the extra metal in those locations interferes with sensing. After some trial and error, I ended up mounting it ~2" down from the top edge, between two of the top-side bungs. In that location it detected changes as I tilted the tank very reliably, and I never saw it miss. Sensing a change of state takes about one second, at most.

From my testing, I'd say it's a winner. It's dirt cheap, it's small and lightweight, it mounts outside the tank and it doesn't require an additional tank, hoses or fittings. I'm going to go ahead and use it. I'll design a small interface board to make it easy to connect to ship's wiring. Once that's done and tested, I'll post details in case anyone else is interested.

One note: I don't know that the sensor location I tested is feasible once the tank is mounted to the airframe, so I'm planning to mark areas with no interference on the tank and re-test, when I get to that point.