3D Robotics

T3-Round 8: The egg-drop round!


Inspired by the Outback Challenge's waterbottle drop challenge, we're happy to now announce the start of the 8th round of the T3 Competition.

Your mission this time: have your UAV autonomously drop an egg as close to your home/launch position as possible (heads up!).

(Note: Our contest judge, Gary Mortimer, reminds us that it is illegal to drop any solid object out of an airborne vehicle in some countries, including the UK, so please check your local regulations to determine if this is legal in your area.)

How you have your UAV carry the egg and what mechanism you use to drop it is up to you (maybe a good time to use the built-in relay on your new ArduPilot Mega board?). The only requirement is that the drop be AUTONOMOUS--you need to set your autopilot to initiate the release when it detects that it's the right distance from the home location. The aircraft must be in forward motion with a speed of at least 15mph at the time of drop, and at least 50ft high (ie, no unfair advantage for quads!)

Because this first "T" in T3 stands for "trust", we're going to trust you to mark your home position and measure the distance the egg landed from that--no need to strap a GPS logger to the egg. The path your UAV takes before and after the drop doesn't matter, as long as it was under autonomous control during the drop part of the run. You will get EXTRA POINTS for an unbroken egg. How you achieve that (parachute, whatever) is up to to you, but please document your method with pictures.

Please submit the following in the comments as your entry: KML track of your UAV, with drop point, egg impact point, and "home" marked. Distance measured and reported, along with autopilot type. Please include a picture of your egg after it's landed, broken or not.

Scoring will be as follows: competitors will be ranked in closeness of egg to home. You get a 10m bonus for an unbroken egg.

Deadline: about six week from now--Sunday, September 5th at 12:00 midnight PST.

Have fun!

E-mail me when people leave their comments –

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

Join diydrones


  • T3
    I added a little fudge factor for release delay, but it was a guess for the first drop. The release responds pretty quick though-probably less than 200mS to open the hatch.

    The plan was to tweak the delays after learning from the first drop, just ran out of time.

    You and Brakar had some nice runs!
  • T3
    Hi Brian. Good effort despite all the problems.
    I had to include a ~1 second delay to allow for the release mechanism to react and actually release the egg. Did you include any delay for this?
  • T3
    KML File Upload didn't work. Lets try this.

  • T3
    The gods were conspiring against me on this one, but persistence finally prevailed. The results are below, but first I will share my tale of woe and triumph.

    It all started last week when I decided to put the plane in my truck for a drop attempt after work. I looked the plane over before launch and everything seemed OK so I let her fly. Suddenly she rolled over and nosed into the ground. What the heck happened?? Rebuild number one.

    A few days later I'm ready for attempt number two. I throw everything into the truck and head out to the flying area knowing that I still have about a week to learn a few things from this drop and roll it into another attempt. I throw the bird into the air and hit the throttle. Everything seems fine and I gain some altitude then suddenly she rolls over hard and heads for the ground. I have time to cut the throttle to minimize the damage and much to my surprise I have control again. I level out and decide to get some altitude and keep going. Throttle up and she quickly rolls over again and heads for the ground. No time to recover. Rebuild number two.

    I notice as I'm picking up the carcass that the motor is skewed. I grab it and find that the glue that used to hold the motor in place is all soft and gooey. I throw the pieces into the truck and head back home. The temperature had been well over 100 Deg F for the past few days, and when I got back to my air-conditioned shop I noticed that the glue holding the engine was solidified again. Apparently when the glue got too hot it would soften up and no longer hold the motor in place. At least I know what was happening now, so I epoxy the motor into place and glue/tape and generally force the smashed battered foam back into something that resembles an airplane. I also noticed that the receiver case had popped apart. Looks OK. Probably should do a range test to make sure, but I'm sure it's fine.

    A few days later, out for attempt number three. I'm running low on time but should still be able to learn some things from this drop and roll it into a final attempt with time to spare. No Worries. The wife is also dropping comments about finishing up the patio project that unfortunately has been competing for my time along with the T3 contest for the last several weekends. Given my wife's general lack of interest in all things hobby/electronic/airplaneish and her strong desire to have a new patio, I think it was her that summoned the gods to smite me hoping that I would have given up by now. But I don't give in. Bring it on.

    So back to attempt number three. I release the plane into the air knowing this will be a glorious flight culminated by the release of the golden brown egg hitting inches from the target. She seems to be flying true. I'm at 3/4 throttle and she's not rolling over. She's slowing sinking so I gently pull back on the stick. She's still sinking. Pull back. Still sinking. More throttle. Still sinking. Pull back...PULL BACK...PULL BACK....Rebuild number three.

    Looks like the gods have kicked it up a notch, but I'm not out yet. I decide the battered airframe is too far gone, but fortunately I have a shiny new Floater kit that arrived from Hobby King a few days earlier. It'll take a couple nights to build it up and I'll be down to my last weekend, but it should still work out if I stay focused and convince my wife progress is still being made on the patio project.

    The plane is near completion when I notice the carpet in my office is all soggy around the water heater. Oh no. Not now. A leak in the water heater steals away several of my precious few remaining hours. Curse you gods.

    Finally the plane is done. I lay a few more tiles to appease the wife and I'm off to my flying spot. My heart is pounding in anticipation as I launch the plane. She gracefully and effortlessly climbs out. I had forgotten how much better she flies when the foam isn't all battered and smashed with four pounds of tape and glue holding it together. I flip the switch to autonomous mode and she banks towards the first waypoint. When it gets to what looks like the first waypoint the plane begins to circle and drop in altitude. I flip back to manual mode but I have no control. I notice when I flip to manual the plane levels out and begins to fly away from me though. Back to auto and it looks like she's headed for the next waypoint but then she starts to circle and descend. I remember that the fail-safe mode is to circle and descend, but why is it in that mode. My radio can work at ten times this distance.....Oh crap, maybe I should have done that range test when I found the receiver case split apart. Back to the ..dam gods...problem at hand. Auto mode = descend and circle, Manual mode = straight and sort of level. The plane is getting pretty far away so I need to do something quick. The plane descends too quickly when in fails safe mode (note to self - Fix Fail Safe mode) and I have no control when in manual mode but at least it's level. I decide to let it circle and descend until just before impact then flip to manual and let it level out and land with hopefully minimal damage. Other than clipping a bush and flipping around, the plan seems to work.

    When I get to the plane I notice that there's no egg in the holder. I should note that after every previous crash in addition to finding egg goo all over the housing, the servo that controls the release mechanism needed to be replaced because of stripped gears. I used very cheap micro servos and fortunately bought several. This time there was no egg goo and the servo seemed to be fine although the hatch seemed a little loose. Hmm something new to ponder.

    I head back home and install a few more tiles on the patio to claim progress then start mulling over the software to see if I can figure out the strange behavior when switching between auto and manual mode. I decide that since I'm only using one of the RC channels to determine when I have valid contact with the transmitter that maybe the failsafe mode would behave that way if I was getting partial reception. I decide to swap out the receiver and try again. I'm on my last day so it's now or never. Some things obviously need to be looked at in more detail and fixed but their simply isn't time. Off I go to the flight area.

    Into the air goes attempt number five, quite possible my last attempt - The wind kicks up every day about noon and the wife is at her limit for patio progress. Into the air she goes everything looks good. She climbs right out without effort and I hit the auto switch and cross my fingers. She heads to waypoint one no problem. Waypoint two looks good. Three..Four..Five..Six..everything looks good. Here comes the plan back to the launch point to drop it's cargo. I can hear the release servo move as the plan passes over, but no egg drops. What now. I land the plane in despair thinking that was my last chance. The gods have prevailed and foiled my plans at last. I shut down the ground station and start to pack up the camera and antenna when I decide to have another look at the egg release mechanism. I take a second look at the loose hatch that I noticed before and can see that the servo case is actually broken so it can't support the weight of the hatch and egg. It probably popped open enough to let the egg squeeze out during one of the turns. At this point I'm desperate. No more doing a test drop and learning from the experience to refine my technique. I just want to drop something to get on the scoreboard. I decide to put some tape on the hatch to hold the weight of the egg and hope the servo will have enough force to pull the tape off when it's time to drop.

    I load up my last egg (my other spares had broke on the bumpy ride out to the flight area), start the cameras rolling and launch the plane. Everything goes smoothly as the plane flies it's pattern and heads back to the launch area for release. As the plane approaches I watch intently with the gaze of a cat about to pounce on it's pray. My senses are heightened and time seems to slow as the plane approaches the drop point. I know the plane should drop 30 meters before the target point based on the ballistic calculations I ran back on my computer, but no drop. An eternity seems to pass and my hopes for a successful drop start to fade when suddenly I hear the sound of a servo whirring and groaning. It stops momentarily but there is still no egg. A split second later the servo groans again and I let out a deep sigh of relief as I witness an egg drop away from beneath my plane. I watch in slow motion as the egg tumbles through the air followed by a satisfying "thump" as it hits the ground and explodes in a cloud of shell and slime. The wind had blown the plane about 16 yards off to the side and the late release meant I landed long by about 20 yards but I finally had a successful drop and I was happy.

    Back home I went, knowing I still needed to reduce the data, make a movie and post the entry, but first I needed to finish the patio and prepare for guests arriving at 6:00. As I toiled away setting the one inch Travertine tiles into the sand with my rubber mallet I taunted the gods and gloated that I had prevailed and they had lost despite their best effort to deny me my rightful place as an entry into the T3-8 competition. As if to acknowledge my victory the gods left me with a parting gift. As I was bringing the rubber mallet down upon the travertine tile with all the force a great warrior - akin to Hercules or Perseus who dared to challenge the gods - could muster, I felt the slightest nudge against my arm, just enough to throw my aim off slightly. Down on the edge of my thumb came the mallet, splitting the flesh and driving the nail deep into my finger. As blood began to flow from the wound I quickly lost my smug persona and headed for the bathroom for a bit of self-administered first aid

    Not the best run, but it will have to do.

    Airframe: AXN Floater (Easystar clone)
    Autopilot: PicPilot
    Wind: East 8 mph gusts
    Drop Speed: 25mph
    Drop Altitude: 183ft AGL
    Miss Distance: 26M

    KML File

    Bombs away!!!


    I did learn several things that I'll roll into my next attempt at this. Most importantly is that the lag in GPS data appears to be what caused the late release. The autopilot thinks it dropped the egg at exactly the right time based on the logged data, but I clearly witnessed the egg dropping about 15 meters late - about the right distance for a 1 second GPS latency at the speed the aircraft was moving.

    Time is growing late and I'm in need of rest after sparing with the gods, so no video just yet, but here are a few pics to commemorate the odyssey.

    Patio + Zack+ Reaver

    Good to Go!.
  • T3
    Hi Michael,
    What will the UAV OBC Committee do with it?
  • Well done guys, I have passed this on to the rest of the UAV OBC Committee.
  • Moderator
    NIce one Brakar
  • Eggdropping isn't exactly in the midst of my interests. I did however do one run as part of testing of trigger-settings. I attached two boxes with eggs to the plane and set the two triggers to release 10m apart. (I am a bit lazy, so I didn't bother to do complicated math. Did only do Newtons formulas).

    So for the results; one egg landed 26m away, the other one 31m. Crazy thing is both eggs survived the drop.

    I didn't find out how to upload the kml-file, so here is a plot instead:

    And here is a pickture of the plane, fully rigged:

    As can be seen from the pickure, I had attached a FlyCam One2 to the nose ot the plane to film the drop from above. Unfortenatly the cam didn't work this time (either), so no video.
  • "HOWEVER nobody cares if the bottle will land intact! Nobody states: the bottle must stay visible after landing. Nobody states: you must not hit closer than x metres to the person, they say 'you must not hit him directly' (this is same as saying 'you must not hit a nearby scorpion on the left leg')."

    I know this is a bit late, but the rules of the Outback Challenge do say that 500mL of water must be recoverable from the bottle after landing, so logically this means that the bottle does have to survive the landing intact.
  • T3
    Hi Krzysztof,
    I was not expecting such accuracy so a bit suprised too. As you see from the video, the egg almost hit me! (one hit the car). The launch/home position was on the grass directly in front on the other side of the concrete pavement.
    In answer to your questions:
    - I had 6 eggs so I had 6 tries. Repeatability was quite good but 1 metre was the best.
    - The Start and Release waypoints are calculated and updated twice based on, amongst other things, both the wind speed and wind direction.
    Hope this helps, Mark
This reply was deleted.