Yestarday I wanted to make a first "real" range test of my DIY LRS system. "Real", because RX was supposed to be on a flying wing, but only as an passenger. Actual control was supposed to be happening via FrSky X8R. Crossbow RX was only to measure RSSI and check for failsafes.
It, well, did not ended up very well. Just watch the video.
I’ve described my Dropship Glider Project here. Previous weekend I finally tested its ability to glide when dropped from a drone 50m above the ground. And well… to be honest, I failed hard this time. Just see this short video from FPV camera:
When I saw this video from rctestflight I knew I will build something like that for myself. A voila, few months later it is done. Here is Dropship Glider.
It is 29cm long with 20cm wingspan. Weights 97g AUW and has 21g/dm^2 wing loading. So, in theory, should glide. Somehow… If I got center of gravity right. And did not made ailerons too big. Or…
Delta 6mm Depron "wing" is attached to 6mm carbon fiber rod and has some quite big dihedral: 15 degrees.
The biggest problem was radio link and mixer for ailerons. I could not use my Taranis: I need that for a carrier and only radio control link I had was EM-16 with PPM output only and no way to setup any kind of mixer. The radio just has no "features" like that…
So, took one Arduino Pro Mini and wrote short program that acts as PPM decoder and mixer for ailerons.
power is supplied by 1S LiPo taken from my Tiny Whoop
FPV AIO Eachine TX02 also taken from my Tine Whoop
5V is supplied from cheap, regulated, step-up converter
3rd servo is to release tether
First flight, or rather drop, tomorrow. There will be a video from the event of course…
Depron airplanes are easy to build, but also do not last long. Few weeks ago I decommissioned by Marabou Stork. It did not crashed, it was flying almost fine, but it had at least a few problems:
Motor was overheating. Just like that
Wing that I originally designed for something weighting around 450g, now had to work with twice that much weight. Take offs were, at least, problematic
It was big and slow and was not giving enough fun
So, to paraphrase Monty Python:
This is a late marabou stork
It's not a complete end of this design. Wing is still in pretty good condition and I have a plan to reuse it in experimental twin-motor design. But that later this year… In a mean time, last video footage.
One of the things that INAV was missing, was a decent support for Pitot tubes, or more generally speaking, airspeed sensors. Autonomous flight, or landing, without knowledge about airspeed can easily lead to a stall. Stall can lead to crash. A crash leads to rebuild. Rebuild of big airplane is a nightmare. Although, for some time now INAV was able to use digital PX4 Airspeed Sensors (I2C, based on MS4525), but they are quite expensive and airspeed was only reported in blackbox logs. Not very useful, right?
Now, this is changing. Next release of INAV (1.8 probably) will bring at least support for much cheaper, analog, APM Airspeed Sensor based on MPXV7002 chip. Although some simple additional electronics (2 resistors to be precise) will be required, but this pitot tube should be available for all flight controllers with free ADC input (Current or RSSI). Fancy ADC remapping will allow to use any ADC without built in dividers (Vbat has dividers so can not be used) as pitot input.
More than that, INAV 1.8 will (or at least should) bring PID scaling according to airspeed for fixed wings. This should result in better handling on both low and high speed.
As you can see on the picture above, APM Airspeed Sensor is already installed on my small flying wing and is waiting for first flight tests this weekend. Logging only for now…
It's official: next release of INAV (1.8 or maybe 1.7.2) will incorporate an automated landing procedure for fixed wings. I was already writing about it 2 weeks ago, but now new code has been merged back and will be released.
Bear in mind, that this is not "state of the art" landing yet. It's rather a simple solution that can be used in emergency situations that will not crash an airplane, but rather put it on the ground without crashing. Procedure is quite simple:
When Return-To-Home or Failsafe with RTH is engaged, go to Home position
When Home is reached, start to loiter with defined radius and descend. Descent speed is limited to nav_landing_speed when altitude is above nav_land_slowdown_maxalt. When altitude is below nav_land_slowdown_maxalt, vertical speed is scaled down to one fourth of nav_landing_speed at nav_land_slowdown_minalt. So, on using default values, vertical speed is between 2m/s and 0.5m/s
During descend, airplane is not allowed to raise throttle above nav_fw_cruise_thr when nose is up. This is to prevent airplane from gaining horizontal speed
When nav_land_slowdown_minalt is reached, ROLL axis is locked to 0 degrees, PITCH axis is locked to nav_fw_land_dive_angle (default is 2 degrees) and motor is stopped when MOTOR_STOP is used or put to IDLE when MOTOR_STOP is not used. This puts airplane into a shallow dive to the ground
That is all. Airplane should glide last few meters to the ground. Most designs should be able to do it without a problem. My testing platform did it like that:
Since there is no auto-disarm procedure yet, MOTOR_STOP is recommended to prevent propeller from breaking and motors/ESC from burning.
Those of you how subscribed to my YouTube channel should have noticed, that I got an interest in automated landing of fixed wings after RTH in INAV. And the sad truth was that, well, INAV up to 1.7 was unable to do it right. When landing after RTH was enabled (nav_rth_allow_landing = ON) and it was enabled by default, airplane usually started a 20 degrees dive to the ground. One does not has to be a prophet to figure out how it ended.
For example like this:
If not manual override, that would end up in a beautiful crash and probable full rebuild of an airplane.
Luckily, that motivated DigitalEntity enough to something about that, and yesterday I was able to perform (probably the first one ever) a controlled descend after RTH that ended up with an airplane on the ground without any damage. With enough optimism one can call it even a landing. This is how it looked like:
Current implementation is still far from perfect. Although it does not crash, it has a few small problems:
No disarm. Throttle is open all the time
It happily ignores speed. Both ground and airspeed
It also ignores wind, heading and so on
But, to be honest, this is a nice progress. Stay tuned for more changes here, since I’m planning to work on it in the near future.
Miniquads are fun, right? After all, there is a reasons most uf us flies 210-250 quadcopters. If so, small flying wings should be fun too! I tried that already in late 2015 and failed miserably! My design did not survived maiden flight. Well, things like that happens from time to time, so few months ago, after learning few new things, I’ve made a second attempt. And this time I’ve succeeded.