1 hour 20 minutes - APM 2.5 takes the world record for Multi-Copter Duration of Flight powered by a rechargeable battery. [note: records also exist for hover that can occur indoors and take advantage of ground affect]
- Distance > 1 kilometer
- Minimum elevation from ground the greater of 2 meters or two prop diameters (no ground affect)
- Lands within 50 meters of Launch
- Minimum of two way points > 0.25 km apart
- Altitude Climbs: Two climbs > 100 meters each
- Ends before voltage drops below recommended minimum level for recharging
- Flight over ground that does not vary more than + 20 meters in elevation
- Distance: 1 - 2 km
- Min elevation 2 meters
- Ended at start
- Three way points with two 0.3 km apart
- Two climbs: 126 meters and 112 meters
- 13.06 volts left with 10 volt recommended minimum
- Ground + 10 meters
Time: > 1hr and 20 minutes (81.43 minutes)
- Octa 2XQuad 6Up+2Dn (wanted to fly something novel that would show off the flexibility of APM 2.5)
- 3.18 kg AUW
- Li-Ion battery
Attached are the flight logs. Later I'll provide:
o Earthview of flight
o Altitude gains
o Details of the copter (design and weights)
You didnt mention the actual battery capacities.
The ship is a Octa 2XQuad w/ 6Up+2Dn. While the octa shares common weight (the controller, receiver, etc.), the added structure required for an octa usually out weighs the advantage. Hence, the most efficient ship for duration flight is probably a quad. But for photography the ship needs an open front and redundancy.
I also wanted to demonstrate the completely flexible manner in which 3DR APM was designed and implemented. Hence a complex ship that would support a camera (open front and rear ends for field of view); props at three different levels, along with some props up and some down (why not).
Not the best photo, but it is the photo taken from my 2.5 story man-cave (camera was on auto) during the record flight on a return leg to the man-cave window.
A more detailed look (both halves are the same).
Each half of the ship is basically a quad with one inverted inboard motor.
ESCs are stripped of all insulation, painted with silicone, and put in the prop intake to cool.
The battery, 40 combined into a 4S10P, had to be wrapped in PVDF foam (made prop clearance a bit fun; PVDF is a nontoxic fire certified foam) since it was -2C out (significant weight gain). You can see the little green batteries looking under the main motor spar. Even with the foam, the battery only produced about 80% of what it would in the summer.
The construction process followed my blog on carbon tube construction using braided carbon composite tubes, aerospace adhesive, and zip ties.
I'll work on the video and flight record next.
Very impressive! I am looking forward to see the record video.
Nice!. Did you use "standard" octa config?
I use crossed sticks underneath style quads, and different prop levels don't matter at all ;-)
Next solar power addition (summer)
What's nice about APM is that you can easily customize the code for any layout. I have a draft wiki if you are interested. The code simply replaced the code for a V ("variant") octa. The motors were numbered 1 to 8 going from left to right and then fore to aft.
| add_motor_raw(AP_MOTORS_MOT_1, 0.691, 0.522, 1.000, 1);
add_motor_raw(AP_MOTORS_MOT_2, -0.691, 0.522, -1.000, 2);
add_motor_raw(AP_MOTORS_MOT_3, 1.000, 0.000, -1.000, 3);
add_motor_raw(AP_MOTORS_MOT_4, 0.382, 0.000, -1.000, 4);
add_motor_raw(AP_MOTORS_MOT_5, -0.382, 0.000, 1.000, 5);
add_motor_raw(AP_MOTORS_MOT_6, -1.000, 0.000, 1.000, 6);
add_motor_raw(AP_MOTORS_MOT_7, 0.691,-0.522, 1.000, 7);
add_motor_raw(AP_MOTORS_MOT_8, -0.691,-0.522, -1.000, 8);
Is this really a world record?
This is the elevation chart from the flight log showing the two +100 meter elevation climbs.
and converted to excel
As you can see, I'm not a very good pilot. A lot of waste and lack of elevation control. Later I'll experiment with altitude hold (this flight was, except for a few seconds, flown entirely in stabilize mode).
This is the earth view of the flight. This is the top view. It's an important perspective for understanding the video. The man-cave is the right building where I brought the copter up to the 2.5 floor window where the camera was located. The camera is pointed north from the man-cave to the turn-around point of the second way point (upper left of the photo). The photo is rotated so north is from the man-cave to the second way point (going lower right to upper left in the photo).
This is a side profile of the flight.
I'm working on the video now. I've tried two compiles and both ran out of memory. If need be, I'll publish the 2754 photos (less than a gig versus over a gig for a video).
So far, I'm very dissatisfied with the video process. To film something this long, I used a canon shooting at 0.5 Hz with a super good lens. But it wasn't adequate. The objects are too small. The field of view trying to cover half a kilometer by 130 meters high is too wide. I need to rethink how these events are documented. Maybe the only way is a device that can go for two hours or more and be carried around with you on a helmet (or have a friend film it) without running out of juice or memory. Suggestions would be greatly appreciated. What camera can go for 2 hours in the cold? What camera can film for 2 hours and not run out of memory? How can you process a file that large? How do you share a file this large? If the flight was stationary around the window of the man-cave, it would have been fine, but that's not what flight is about. So suggestions please.
Finally broke the video into three parts. The videos are compiled from 2754 time-stamped Canon stills taken every 2 seconds. At a video rate of 20Hz, the videos are 40x actual speed. Sometimes the copter goes outside the FOV of the camera (had to trade between showing detail and scene area). The white box that is the most visible part of the copter is the thermal blanket on the battery. If you lose track of the copter, watch where I am and the copter is usually 5 to 120 meters up.
Stage 0: Shows the flight prep and takeoff location about 1/3rd the way between the mancave and way point 2.
Stage 1: The octa
- is launched
- brought to the man-cave viewing window
- sent north on it's first slow 120 meter climb reaching the apex about .3km away
- reaches the second way point (near the lone fir tree at the end of the field)
- turn SE over the orchard fan towards way point three
- turns SW back towards the mancave
Stage 2: The octa
- shows at the man-cave
- pauses near the launch site
- returns to way point 2
- returns to way point 3
- heads back to the man-cave
Stage 3: The octa
- shows back at the man-cave
- begins second 100+ meter climb
- returns to launch
- return copter to mancave
Astounding! Huge congrats. Can you compile all the stuff in this thread and turn it into a single blog post, which will appear on the front page of the site? It deserves the attention!
Great accomplishment, novel design and very careful choice of components really produced a serious endurance copter.
I would actually think any piloting inadequacies only add to your accomplishment, hard enough to maintain adequate concentration for that long a flight.
Clearly the Panasonic's (and low enough power requirements for hover and flight to accommodate them) are still the way to go for records.
But your unusual frame layout proved it had genuine flight capabilities.
And I am sure it would turn in impressive flight times even with a conventional LiPoly and a camera.
Why did you go with 15" DJI props instead of T-motor, actual test or couldn't get the T motor props?
An impressive flight but I'm curious about exactly what you mean with the words "World Record". Could you please tell us: Was this adjudicated by the FAI? What is the source for the "Flight Requirements"? How does this relate to the current FAI (pending certification) endurance record for electric helicopters in class F5-Open of 2 hours and 31 minutes? You refer to "records [that] also exist for hover," by whom and under what certifications might those be? I ask all of this because there are people working on or considering long-duration flight projects who need to know what the current numbers to beat, as it were, might be. Thanks!