Hi everyone

I have read a couple of discussions on this already but it seems no one has asked about this for over a year and I wondered if there have been any advances that would make this easier since then.

So last year my school sent a weather balloon to space with a payload containing various sensors and a camera, and after a lot of searching with a yagi we managed to find it! This year we want to do something a bit more interesting, and so plan to take another balloon up with a glider, and have it autonomously glide down to a predetermined location. We have a skywalker x-5 that we've assembled and got flying, and decided on an APM. (We decided against the pixhawk because it's outside of our budget). We are ordering an neo-6m gps unit for it, and aren't sure about whether we should get an airspeed sensor or not. Should we get one? And will we need to modify any of the code to make the APM do what we want or can we just put it into RTL mode on takeoff and have it glide down once it is detached from the balloon?

Any help is greatly appreciated.

Views: 3777

Reply to This

Replies to This Discussion

APM needs to be armed on the ground, standing still, and then set to STAB/manual mode..

Or switched on in "Stab" mode ?

Airspeed sensor is highly recommended for flying in windy conditions.

You somehow manually or automatically have to command APM to go into RTL mode at release height.

Most auto type modes would engage the throttle on selection.

Maybe a PIC or Arduino switching a pulse train on the "Mode" pin from "Stab" to "RTL" upon release?

And a small relay to switch "Mode" pin back to your RC RX for landing.

Thank you! I'm ordering the APM, neo-7m gps, and analog airspeed sensor. We'll set up a small picaxe chip outputting to ch5 of the apm to tell it which flight modes to go in to. Do you know if we'd be able to set one of the APM's outputs to high when we reach a certain altitude, in order to break the string between the plane and the balloon? Thanks Gustav :)

Not too sure if that can be easily done, look into setting up a "Geo Fence" function.

Ambitious project but, a good one to discuss.

Before making the big leap, why not test your vehicle at intermediate altitudes?

With the proper lifting plane, you could release at .5, 1.0, 1.5, ... km to dial in the performance and function.

If no budget for a lifting plane, tethering the release balloon might be an optional approach.

With that data, you might find a timed return profile would fit your project.

A manager processor (picaxe in your case), stepping through the return profile would also monitor for external override command from your GCS.

Scaling up to the big leap is cheaper than an 'all or nothing' approach.

I love projects. Plan on!


If you can program a picaxe then you can program an arduino. You can make your APM do what ever you please (within the capabilities of the hardware) - including arming, changing modes, setting a given pin high at a given altitude and much more... All this without adding any additional hardware at all. Check out http://dev.ardupilot.com
There are plenty of folk around that can point you in the right direction when it comes to finding your way around the code and implementing some custom feature.

Thanks Gustav, I'll look into using that so that when it gets to our altitude it'll try and return to ground level. I think what we'll need is a kind of dormancy mode while it waits for the altitude to reach say 25000m OR notice a considerable decrease in altitude (as in the balloon burst early) and then have it have a pin high to melt the string. Then it'll need to go into an auto mode and say glide down to 2000 metres. At that point we could do with it deciding which landing spot to choose based on which it's closest too (could we program a list of GPS coordinates before takeoff?) and have it auto land. I'll ask the guys on the dev site, thank you James!

At the moment we've got a setup with a motor and prop so that we can do some test flights and work out the glide slope easily. Then we'll try and do some RTLs with the motor. Then after that we plan to fly it to say 500m in the air and see if it can accurately glide down without the motor to where we launched from and go from there. Because of the distance it'll travel we won't be able to manually land it, so we'll have to pass that over to the flight controller! I'll keep this thread updates as we make progress :)

Niall Glad to see your interest in this field. Yes there has been a great deal of effort going into this by our team Team Prometheus and others such as LOHAN and NTNS.

What we have been working on the most is long range telemetry and video. We feel getting telemetry and video at a range of 250 miles for a single ground station is a key factor in a working spaceplane mission.

Our goal is to launch a plane into space using a booster rocket from a high altitude balloon.

Some of the issues with the APM code are speed density which needs to be worked out for high altitude flight.

Our system includes cold gas thrusters for stabilization during the trip threw the high atmosphere where there is very little air for flight. 

If you attempt flight at very high altitudes you will need a LOT of speed to be successful. You will not have control without some type of stabilization or enough speed to maintain attitude.

Another issue could be saturating the INS with spins and flips during the decent.

Also you need to decrease the G sensitivity from the standard APM to a higher value.

We are now about to begin construction of the X12A which is a continuation of the experiments we have done with Chris Courtis and his X11E which is an all aluminum larger version of the same aircraft.

You can see progress https://www.facebook.com/pages/Team-Prometheus/213171526223?ref=br_tf or at teamprometheus.org (which I update far less)

You can also use X-Plane once you have built yourself a model to do simulation flights I suggest you do that if you can first.

Good Luck! and Welcome!

There is much work still to do with the APM code. Do not underestimate the need for telemetry and video at long range. I suggest you get that worked out on your balloon flights to come.

Here are some highlights.

This is an ongoing project of several years work.

 You can just use the APM as is for a flight IF the INS does not loose it's mind on the tumble from altitude and it will start to fly around 50,000ft (change the G sensitivity)

 You can set the geofence to release the plane if you don't want it to get too far away. You can use a simple dowel threw the CG of the plane and use a servo to pull a pin from the dowel to release the plane.

 You can use RTL

 If successful what you will get is a video with a LOT of tumbling and spinning until you get to about 50kft. The video will most likely be unwatchable until you get to the lower altitude.

 The trip up is great! So that video is nice.

 So if I where you I'd get a Skyfun because I know these fly fairly well. Learn how to fly it. Learn how to use the APM with it. Use the 3DR 433 radios or the RFD 900, get a good CP antenna for both ground station and plane. Forget about FPV (Unless you go for a light 5.8 ghz to use at landing and you get good enough at tracking your balloons and can be at the landing site) and just use a gopro on the plane or something similar.

 If you make it back to launch you would have been very lucky with winds that day it is possible but unlikely. So be prepared to give chase!

 Don't mess with the code accept that you will tumble from high altitude. And go for it!

 There's a little more to it than that and that part you will learn launching balloons and retrieving them.

 One last thing get you a Big Red Bee APRS tracker on 2 meters (requires a license in the US which is easy to get or have a ham friend use his) or something similar depending on where you are located.

 When you loose contact with the plane via telemetry that way you can give chase and regain your signal. The new antenna tracking feature using an apm or pixhawk could work for you better but I have not tried that yet (we optically track with a telescope)

 A hail Mary would be just the apm, RTL and a APRS tracker and hope for the best. You could set up an LZ in MP and perhaps do better.

 Setup your plane heavy with a fast glide slope and don't try to fly until you you hit 50k and you'll have plenty of fun!

Ahh yes one more thing the wings of the skywalker are not ridged enough. Your plane will hit at least 200mph if not more during the first part of the fall even with a parachute perhaps even more.

You need a plane with ridged wings like a skyfun or a EDF Jet.

The NTNS EDF hit 498mph on the way down.

Yeah, but in very thin air, so its not going to cause the same structural problems as flying near sea level. 

Reply to Discussion


© 2020   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service