This session marks the third flight of our original Skysurfer. As it had been already tested and trimmed in previous sessions, we were expecting to do 1 or 2 warm up flights before mounting the node for tests.Unfortunately, Singapore being just 1 degree north of the equator means that weather conditions vary greatly from day to day. Below was footage of the second flight of the day that ended in a crash

The crash here was entirely due to pilot error: as the white plane was flying against white backdrop of clouds, the pilot over-banked the plane during a turn, resulting in it going inverted. Not realising this, he pushed the elevator forward with the intention of gaining altitude but that caused the plane to nose down instead because of its inversion, hence leading to a crash.

Note that the plane was also rather wobbly throughout the flight. This was because the last time the pilot flew was about 3 weeks prior and had zero stick time on the simulator due to school commitments.

As this crash occurred at significant speed (evidenced by the loud thud), it broke the plastic gears of the elevator and rudder servos, causing quite a bit of downtime as we replaced them (we brought spares, thankfully) and taped up broken pieces of foam.

Second Attempt…and Second Crash

After patching up the plane and waiting out a passing drizzle, we were ready for another go. We noticed this time that the plane’s CG had shifted: it could no longer fly with the battery right up in the nose as this would lead to a 737MAX nose down at launch. We suspect that this might be due to the additional layers of fibre tape used to hold the nose together after the first crash.

Shifting the battery eventually got the plane to a flyable state. Here is some onboard FPV footage(please excuse the choppiness; it was downscaled from 60 to 30fps):

For reasons unknown, the onboard camera stopped recording before the crash occurred.

Here is footage from the ground:

From the footage, it seems as though this crash was caused by the battery shifting forward mid flight, throwing the plane’s centre of gravity off balance. We deduced this to be the likely cause because as the plane pitched down towards the end of the video, the pilot reported having zero control despite having the control stick pulled all the way back.

After this, the plane no longer could fly as well as before. besides the CG being off, the wind was also picking up speed, which led to it being significantly more difficult to control the plane in flight. This, coupled with the fact that on the field for quite some time already (around 3 hours), we decided to call it a day, having about 2 out of 4 batteries left untouched.

With that, we strapped the collector node to a quadcopter belonging to Kanesh since the plane was no longer flyable. Below are two graphs – the first is of the data actually saved to the SD card of the ground node. The second is of the data transmitted from the ground node to the collector node (on a UAV). It worked as expected (after dealing with a timezone display issue where the timezone was assumed to be GMT+0730) – the sampled data matched the collected data!

Data from Ground Node

Data from Collector