Hi,
I'm actually being paid to lift an rc plane as high as I can using a balloon and then release the plane to fly home on it's own.
I have my Permission for Aerial Work and an exemption from the UK CAA for this adventure, plus insurance, so the legal side is sorted.
I am a professional aerial cameraman with a few years multicopter experience for my day job and FPV racing. I used to fly fixed wing rc for fun without autopilot but now do large land surveying using fixed wings.
I have built the prototype plane using a Skywalker X5 and Pixhawk. The delta shape was chosen because it has very few sticky-out bits for the balloon lines to get caught on. The X5 was chosen because I can replace the frame relativity cheaply after "teaching" it it auto-land. It also has a big payload area.
Obviously I've not done this before so I'm hoping that people here will be able to help out. Here's a few questions to start with:
Recovery from stall: Once the plane has been released from the balloon lines, it will fall for a few seconds before trying to recover horizontal flight. I'm not going to try for a long range rc control so the pixhawk will have to fly it home. What is the best flight mode to have it recover? I'm more interested in what mode to use in testing since I will fly up to 100m or so and then force a stall before allowing the pixhawk to recover for me.
The full size plane will probably be either an X8 or a Buffalo, anyone have a better airframe which is easily available in the UK?
I've already sent up a balloon to test the pixhawk, cameras and batteries. I was going to use a Raspberry Pi and Pixhawk to run the mission and make decisions when to move the release mechanism servo, etc but the Pixhawk 2.1 with Edison module looks a neater solution with less things to go wrong. I'd prefer the more tried and tested original pixhawk but I'm happy to give it a good bashing and go back to the pi if the Edison is not up to scratch.
Any other helpful comments or ideas would be most welcome.
I'd like to get this right because it's going to be landing live on TV. :-/
Thanks, Tim
Replies
It's on the todo list :-)
Going to follow these instructions: http://diydrones.com/profiles/blogs/tutorial-programming-your
Got a balloon launch in the next few weeks to test and I'll get back to you with the results.....
Hans H. said:
nice MK :)
what about the GPS problem?
is there any possibility to get a m8n working ?
I'll just point out a few things for you right quick. Metal gear servos (remove the grease and use light oil) The big secret is balance and CG you want perfect balance on the CG front to rear AND side to side. You will hit speeds near 500 MPH on the descent so aileron, wing and stabilizer flutter needs careful consideration. If you can't balance the plane properly and use drag tabs for stabilization during decent your going to spin spin for a major portion of the fall.
If you use a standard RC plane you will not gain much control until about 50,000ft you won't have to worry much about flying until then unless you have a specially designed plane and it's really no big deal as long as your craft can survive the fall.
And actually the original ardupilot with the thermopiles works better for this kind of flight because it can be reset mid flight and re calibrate it's self (something you can not do with any of the others) Which is something I want to add back into Pixhawk at some point but have not had much luck getting Tridge interested in :)
A helical tracker and a CP antenna is the only way to go for FPV at these distances and high power transceivers for telemetry.
I also recommend using X-Plane on the ground station in real time because for one it's just so cool and 2 if you can't get FPV you can fly on telemetry.
The release is the only other big deal. Use a servo to pull a pin in a rod attached to the balloon string. When the servo is activated the pin is pulled and the rod stays with the balloon string.
You can set geofencing to insure the plane does not get too far away before release regardless of altitude just activate the release servo with geofencing.
I strongly suggest using an APRS tracker and you need someone with an amateur radio license so while your at it go get yourself one because it is very easy and you will thank me for it later.
Ah contact me at teamprometheus.org I can help you.
Monroe King
Thanks David. There's a significant price jump in brushless but it's probably worth paying. I've no experience of brushed servos (motors yes, but not servos) any comment on these please?
http://www.kingslynnmodelshop.co.uk/p9223/Futaba_BLS151_Servo_-_Bru...
Tim, Nice project. For servo bearing and gear grease at high altitude the Dupont Kytox GPL203 grease might work for your application. Brush type servos have very short lives at high altitude. You may want to use brushless servos as well.
Oh yes! I forgot about that one. Thanks Chris. The Reg seems to have been running that project for about 4 years with not much happening in the last year or so. Unless I've missed something they've been held back by permissions from FAA et al.
I've got exemption in principle from the CAA as long as it's dropped over a UK Danger Zone. I need to finish the final air frame before I get written permission. My plane will initially be powered by gravitational potential energy and not a chemical rocket, like LOHAN, so permission is easier to get.
Still not sure what flight mode is best to recover from a stall though. I'm planning to test the X5 with a stall at 120m above long grass on Yorkshire moorland, good cushioning for a dodgy landing but not for a full spiral. So far planning to fly up there on Manual, stall nose up and switch mode to Cruise. I'll upload the video if the camera survives!
See the LOHAN project.
Thanks for your reply Hans. I've got lots of experience with Pixhawk, fix wing and multicopters - I know my elevons from my rudders! I fly them for a living.
There's not just ONE try, it's four more balloon launches to test different stages. I've been involved with a few high altitude balloon launches and done one myself, specifically as a feasibility test for this.
There will be two GPS modules, one set to the usual Pedestrian and the other to Flight mode. I'm hoping that the Pixhawk will pick the most accurate, i.e. Pedestrian for altitudes below 12,000m or so and the other module for higher. The next balloon launch will test. The M8N is rated to 50,000m in Flight mode.
Good point about the temperatures, I've tested the Pixhawk and batteries inside an EPO box and the internal temperature started at 7C at ground level and peaked at 27C, while the external temperature dropped to -51C. I must find a grease that will cope with the low temperatures. I'll also need an external test servo or two on the next balloon (another launch with box and no plane) so it can measure the effect. Maybe a servo attached to elastic and a camera to monitor. I wander if this can also be used to "steer" the box so it does not spin so much......
The air pressure was 3% of ASL at 36,000m so I plan for the plane to drop for about 10 seconds to gain speed and therefore enough airflow to get a little control. I pretty sure the airspeed sensor will not work and the Flight mode GPS is only accurate to 100m so GPS speed will be very inaccurate. I'm hoping that the IMU will be enough to get the plane to nearly horizontal. :-/ Maybe.
On the last balloon launch, one of the cameras (Runcam 2) did overheat and stop working but it was completely embedded in the foam and was plugged in to a BEC to charge from a bigger lipo.
Thanks Lyn too, we've done lots of googling to find anyone with experience but only found one paper on dropping plane from a balloon to the surface of Mars and another guy to tried to fly remotely from near space.
https://www.youtube.com/watch?v=rpBnurznFio
and these guys
https://www.youtube.com/watch?v=uGeYNaCl6Go
who managed to release a plane from a balloon but there not much detail apart from it took off and it flew back. Not sure if it was automated or FPV. Sent an email as a request for help/info but no response yet. I'd really like to know what they used to release the plane, probably a timer since their telemetry failed.
Whatever happens, it's exciting and fun :-)
Hi Tim.
I remember many years ago some guy's in Canada spent a long time experimenting with this with various rates of success. They had a really good web site tracking their progress. I have tried to find reference in my PC archives but as yet I haven't found it. I will keep looking. But I do remember they had many failures before realizing any success. Hopefully other members here may remember the web site. Perhaps Rob Lefebvre being a fellow Canadian and a long time member will remember:-).
Good luck with this endevour and I hope you will keep us updated.
Lyn.
-
1
-
2
-
3
-
4
of 4 Next