Unpowered Glider from Space

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.

You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –


  • So far Gary has had the best flight from 100kft or better. Using the APM and a Stinger 64 EDF plane. NTNS 4 in October 2013 in Texas.

     As you can see roll control with this planes aerodynamics is very difficult at the beginning of the flight. Some manual control was attempted that gave an indication the flight might have been attainable sooner than 54kft where the pull up maneuver was initiated.

     The APM was set up with a geofence to initiate the drop when the fence was breached.

     One of the problems my team has been focusing on is getting live video feed at these slant ranges which can be 100 to 200 miles quite easily.

     The other solution we are using is to run X-Plane using the telemetry which also requires long range communications as a back up to video (FPV) flight.

     If you watch the ARES video you'll notice the planes aerodynamics right away the very thin wing sections and contours of the fuselage are very smooth.

     The air molecules have a hard time following the contours over the wing to the control surfaces. You can even get reverse control action (which is what I believe Gary experienced with NTNS 4)

     The altitude of the pull out of the ARES plane is also deceiving because in just a few seconds from 100kft your going so fast your really loosing a lot of altitude.

     Modeling is difficult and the best tools for this are using both X-Plane and Flightgear if you don't have access to megabuck software.

     X-Plane helped us the most with the X-11e to get the proper sizing of the control surfaces.

    3701917494?profile=original You can't just wing it and hope it flies at altitudes this high, it takes real aerospace engineering to succeed on a project like this. 


    •  A couple of other things we do is we use ADSB for avoidance of other aircraft and we use live satellite weather data. We also use GPS prediction software to determine the satellite constellation coverage during the planned launch window. These are safety features the FAA appreciates and they are useful tools others could use to satisfy the FAA your concerned about airspace. (In order to get a waiver to fly legally)  

  • I'm having a hard time understanding something and perhaps this question will help others, as well....

    How could destructive forces be applied to the airframe before flight can be achieved (other than critical mach related)? How would a tumble increase in severity to become threatening upon an increased dynamic pressure? I feel I am missing some fundamental understanding of the aerodynamics of a ultra-thin atmosphere. Would the speed inflicted damage not be caused by the same rise in dynamic pressure that would simultaneously allow static stability to be regained?

  • I think everyone is making this a bit more complicated than it needs to be.  Having problems getting their head around true air speed, indicated airspeed, forgetting mach number, and really not considering the G loading you may get.

    First off, I regularly operate an aircraft in the 40,000 foot range I do have a simple understanding of high altitude aerodynamics though I am not a aeronautical engineer and don't claim to be. 

        If you release at 100,000 feet you will tumble if you have an aircraft shaped air frame with fixed CG and fixed wing geometry. Stuff such as parachutes or streamers hanging off the back could only tangle around the air frame causing a tumble all the way to the ground.  You may find it, and any mission like this should have a cell phone tacking system in place for backup.  

         Shifting the CG way forward (lawn dart) will assist in getting a nose down pitch quickly to attempt to get the wings flying.  Then shift the CG rearward to get the aircraft in a controllable configuration let AMP fly it.  G loading will become a problem when trying to recover to controlled level flight.  Here is why.  You release at 100,000, get nose down at 70,000. You plane gets an indicated airspeed of 50kts that is a true airspeed of 247kts.  The wings feel 50kts and fly well the air frame feels 247 and will load up with G's if you attempt to pull out of this dive normally.  If you reduce the rate of pitch change when you get down low, you will have an aircraft slow to correct for pitch changes when the 50kts indicated equals 50kts true. You could cope with this by limiting the max/min pitch.  By 50,000 feet 50kts would yield a true airspeed of 135kts this is a much easier speed to deal with.  Yet if you keep the nose down from 100,000 feet though as you pass through 50,000 you will be at the terminal velocity of the aircraft in a dive. maybe 250kts indicated?  This is 680kts true, mach 1 at 50,000' is 573kts. You will have a mach stall on the top side of the wing, you probably will not be flying unless you have chosen a supersonic airfoil design.  Here is where the problem lies with trying to fly above 60,000'  with the needed indicated airspeed to fly you will be above Mach one.  Because a few seconds after release you will be going fast, 10 seconds 220mph and climbing. I don't think this is much of an option.

    A huge glider think U2, you would try flying at high altitude right away. Two issues.  One is the difference between the critical mach number and stall speed may close. With a straight wing the critical mach number hangs around .75, at 90,000' that is somewhere just short of  40kts indicated airspeed.  Easy to make a glider fly below 40kts still maybe a 15kt flyable envelope.   The other is air frame failure if it does tumble with long skinny wing you could easily loose them.

    Variable geometry, think Spaceship one. This is great if you can keep it light and simple.  It will tumble, self orientate and then unfeather to flight mode and let AMP fly it.  My thought would be to use a second board (arduino maybe a second ardupilot) to operate the feather system and the primary to fly once configured in flight mode. Seems complicated, maybe heavy. If properly done I think this is a good idea.

          No one has mentioned lifting body.  This is a perfect use for it, except that it will be a fast ride.  After drop it will tumble, but as the atmosphere density increases the AMP will start to get control.  Being a lifting body you could design an air frame that could take a lot of G's.  Lifting bodies have a tendency to be unstable, but the AMP seems to handle poor stability quite well in my experience.  Learn to fly it using the APM fly by wire mode and let it rip.  I think this is a practical solution for a high altitude glider. 

          You could also do what has been proven to work, let a delta wing tumble till the AMP can control it.  NASA liked the delta wing design for a reentry vehicle.  It's swept wing handles higher mach numbers well.  The structure can be made strong enough to handle a lot of G loading by adding some carbon fiber spars and keeping the wingspan short.  Again the only downfall is a fast approach speed.  You probably will not be flying till under 50,000', a swept wing can handle a mach number of up to .85 you will not have a critical mach problem.The delta wing is what I am building for my balloon project.  It will have a 30" wingspan with fuselage underneath the wing, and a long short vertical stabilizer. No rudder, only elevons.  Our hopeful launch month is April.    


    • Rather silly question but why is everybody talking about getting control so high up? Wouldn't things be much simpler if, instead of, say, 50000 ft, it was pulled out at, say, 3000 ft or something? Wouldn't it have slowed to a crawl (relatively speaking) by then, even in a nose down dive? If that works, perhaps try again higher and higher?

      • We've launched balloon in the past, with a data package attached to a parachute. You chase it and hope to find it, and find it landed on some random lake 25 miles from where you expected it to land. You end up hoping to find it.  With a return glider you can launch and designate a landing spot, or plan is to launch the balloon downwind, then have it return to a guys backyard (40 acre corn field.)  If we waited for the glider to descend down to 3000' even with a 8:1 glide ratio you'd have a 4.5 mile glide distance. If you start flying at 50,000' you have a lot more glide distance.


        • OK, but there is a lot of space between the "so high that we are dropping at relativistic speeds" and "so low we are skimming the trees". 50k ft is above cruising altitude for an airliner. What I 'm saying is why not try to perfect the pullout at low altitudes where terminal velocity is not too insane and then progressively go higher. My totally unqualified guess is that you should be able to go to 15k ft without having to do anything too extreme to the airframe and that gives you 13.5 miles glide distance in your example. That's not too bad for a start and a hell of a lot better than a random landing under parachute.

          • The jetstream winds can be hundreds of mph, so between the ascent and descent it can move a long way away from the launch point.  The aim of getting it to fly as early as possible is to be able to recover the position, so there's no need to waste time getting it flying.

            On top of that, a tumbling airframe will have some pretty random, possibly fatal, forces applied to it compared to flying.  The pull-out doesn't have to involve huge g-forces at all.  In fact, the higher the g-loading, the more chance you'll stall and lose control of it.  It's an issue of adapting the control system to fly in a low density environment.  Don't forget that APM et al has been largely developed to fly in an environment where density doesn't change much and IAS == TAS for the most part.  The non-linearities of high altitude flight can be mastered

            I take your point that you could start the drop at 15k feet for initial testing though, however there's no need to send the craft to 100k feet to do that.  Small, cheaper balloons can get it to 5, 10 or 15k feet to establish the pull-out algorithm before heading higher.

            • You will only stall first if you are under maneuvering speed.  The trick is to make your plane so the maneuvering speed high enough it will stall before it breaks.  

              • You will stall above the critical angle of attack.  The problem is that this angle reduces fairly significantly with reducing Re.  Might be 16° at sea level, whereas could be less than 10° up high.  Adding control deflections exacerbates the problem, so you need to limit them based on density altitude.  For small aircraft, Reynolds effects are a large part of the issue.  TAS/IAS divergence and the effects on controller gains is the remainder.

This reply was deleted.