Ultrasonic fuel level

[RossM]

I thought that I would try and install a fuel gauge so that the fuel level could be monitored while flying (as apposed to try and look at the sight glass along the side of the seat tank while flying).
I do use a timer, check fuel level before flying, and allow for 1/2 hour reserve fuel by the time I return.
I purchased an ultrasonic transducer for less than $30, and an Arduino (from China) controller.

When testing using a couple different plastic jugs, this worked.
When I connected to the fuel tank on my gyro, I am not getting a reading. As the base of the seat tank is not horizontal, I thought that maybe as the signal is going through the tank and bouncing off the fuel level, the signal will not be redirected back to the transducer. So I cut a bit of aluminium bar at an angle to make a wedge so that when I stuck the transducer to it, it would be perpendicular to the fuel level.
Still no luck.
I thought maybe rotational moulded tanks may not work, so I attached to another seat tank stored in my shed. This worked.

Not sure why not working on the tank on my gyro.

[RossM]

the other photos

[RossM]

If I go ahead with installing a (accurate) fuel gauge, may have to go with a capacitive fuel sender.
With a capacitive (rather than a sender with reed switches that will make changes in step on the gauge every time the float passes a switch) you can code the microcontroller for non-linear tanks, and has a resolution of (less than) 1 mm.
The cheapest sender I can source is AU $ 250.
Unless I want to buy at least 20 of.
Would also be cheaper if of a size less than 600 mm long, but need about 850 mm for the seat tanks.

Rossco

[4CA]

Hi RossM,

I’d be very interested in seeing the transducer as a project.. the GG seat tank with sight gauge on the side isn’t the most user friendly I must admit.. I remember seeing a project somewhere that used a laser pointer running down along a bunch of light sensitive diodes and when the laser hit the fuel it bent toward the diode which gave a voltage figure depending on how high it was up the pole. From memory the pole had to be set at an angle for the refraction to be enough yo hit the diodes and of course it could only work in opaque fluids 😃. Cheap set up though..

Paul

[RossM]

I had a piece of 40 mm PVC pipe and stuck an end cap on. Then put the transducer on the cap. This worked as a test piece, even worked when tilting the pipe over a bit. The pipe would also have the added advantage of 'holding' the signal.
Should do a sketch to explain!

The transducer does not work in the range less than approx 5 cm, so if going down the path of using a piece of 30 to 40 mm tube (aluminium would be my preference) one could have the bottom below the level of the fuel tank to allow for this. Just would have to mount the pipe vertically.

I have not tested or repeatability, hooking a small hose to the side of the tube at the bottom and changing the (water) level.

I can buy a 700 mm capacitive fuel level sensor for US$ 100 delivered. The manufacturer has in stock from a previous order, but 700 mm would end at about 10 litres of fuel.

Rossco

[RossM]

[RossM]

I decided to check the consistency of the readings on filling and emptying a tube.
I have used as a test a piece of 32 mm PVC pipe, with a cap on the end (bottom). I attached a hose barb to near the bottom of the pipe and added a sight tube.
Marked the tube every 10 mm, though I only checked every 50 mm of water.
I added enough water (which had roughly equated to the 7.5 mm mark on the centre piece of my rain gauge).
So I tipped in 7.5 mm of the rain gauge level (roughly equal to 50 mm of PVC pipe) in at a time and recorder the reading output from the Arduino.,
I then checked the readings by removing 50 mm of water head from the PVC pipe.
Checked the readings again, filling and emptying the water inside the PVC pipe.

The recorded values were fairly consistent.
This was not done in a lab environment, my pen marks on the side of the PVC pipe would be roughly accurate, and I did not fluffy about trying to get the meniscus of the water level in the site tube spot on.

[RossM]

Recorded values every 50 mm, first column is the marks on the side of the PVC pipe, next column is the Arduino value filling the pipe next column is difference.
Next column is the Arduino value emptying the pipe, then the difference.
Next columns are a repeat of filling and emptying.
Checking readings_2.pdf

So the readings from the transducer are fairly consistent, it would be easy to map and correct the values.
Also map to suit a non-linear fuel (seat) tank.

[waraton]

I got one of these to go into the smaller goodwin seats but it doesn't screw into the fuel cap hole. The threads are supposedly the same so not sure why it wont go. The ultra sonic idea sounds good if you can get it working. Any thoughts on budget fuel senders, the mgl style and red cube style are up to $500 each and I think you need two for a rotax with one on the feed line and one one the return line.

[RossM]

I gather that that sender has reed switches. There may be X amount of switches evenly spaced along the inner tube, and these switch when the magnetic float passes them.

[RossM]

With the ultrasonic transducer, the readings are different between water and petrol. The readings are still repeatable, but a bigger difference between indicated and value.
With water it was about 23 mm difference at 1200 mm, and a difference of 250 mm at 900 mm for ULP.

See attached pdf.
The blue line is height in mm
the red and green lines are values from the transducer of two readings.

When I get the chance, a gyro bother will tig up a bit of 25 mm ali tube with a hose barb at top and bottom, mount behind the fuel tank.
Then I shall fill the tank, litre by litre, get readings, re-code the Arduino, and connect to a fuel gauge.