I have been able to use a pool to test the glider and although there was not a lot of video footage, it provided a lot of information about how well the glider currently works and any flaws that the current design has.

The below video only shows the descent phase as the buoyancy engine was not working reliably enough at this point to demonstrate the transition between descent and ascent. Also, the IMU has yet to be integrated into the control software, so the glider does not have a PID algorithm to adjust the position of the mass; you can see that the mass does not move from its fixed position at the front of the rail. As the mass is at the front of the rail, the glider’s angle of attack is high so the glider’s glide ratio is quite low.

The exterior printed parts would fill with water through very small imperfections, this caused the glider’s ballast trim to become incorrect over time (to the extent that the glider was unable to become positively buoyant at all). I will tackle this by reprinting the parts with a 100% infill, so water would be unable to enter the prints.

Also, when the engine motor has insufficient torque to drive the plungers, the plungers will jam with the rotor still rotating. This causes the nut to dig into the drive screw and the nut wears away the thread. This then causes the plungers to jam more frequently. A peristaltic pump would potentially address this issue. I plan to design a peristaltic pump proof of concept within a couple of weeks.