Ardupilot vs motor failure testing
Вставка
- Опубліковано 9 лют 2025
- Although the results didn't show much conclusively, I think it's safe to say losing one motor doesn't faze Ardupilot in any significant way provided there is still enough thrust, and the prop direction doesn't matter either.
I also tested cutting both motors in auto mode, but forgot to include that in the video. The plane pitched up a bit in an attempt to maintain altitude, which resulted in parachuting downward at 7m/s at which point I got scared and recovered it in manual mode. You can see a clip of that here: ua-cam.com/video/4N046PUA5Ts/v-deo.html
Erm if Andrew was auto compensating and automatically adjusting for the switched of motor leading to your video is it not logical to assume that having an autopilot doing micro adjustments alongside you flying leads to the exact same video Andrew produced? Andrew is making minor adjustments in his head leading to you saying in effect that his data would not translate to lets say my video or stave's down the road, but what would be the outcome if you removed the autopilot (clue in the name there) all together and just flew the plane on servos, esc's, motors and you? then actually tried those exact same tests with no computer assistance, do you think there would be any noticeable difference then?
saturday with a new iforce2d video is a good saturday
Any day with iforced 2d is a relaxing day for me i like 5o chill while watching his videos very calming videos
Nice to see you have nice spring weather and are up flying again.
Really appreciate the inclusion of the "RC input" OSD. This a great visual aid to helping understand what is happening. Suggestion: for different flight modes (10:45), would suggest changing stick-end, or stick-box colour based on flight mode (@10:45 Manual, FBW_, etc).
With the motor-failure simulation (12:50), the right wing is visually tilted (~5-10º) relative to the horizon, and suspect the fuselage has yawed left to maintain a straight track over ground. To better visualize the yaw, I'd suggest tapping a ~7-10cm length of thread, or yarn on, or just in front of the main wing. (alternatively could attach it to video antenna, so just above the wing to eliminate surface friction). If you've ever seen video of a sailplane, often there is such a 'yaw-string' tapped to the canopy to help human pilots coordinate rudder and aileron inputs to help maximize efficiency (drag reduction).
With just the right motor running roll-rate to left is high-rate, and roll-rate to right is low-rate. Logs should be able to tell difference in º/second. (or vice-versa depending on direction of rotation of the powered motor)
Great stuff, specially the bit about me being an experienced pilot. Chopping a motor had a much bigger effect on my AlbaBird
Nice to see you using a flysky . Same one i got.
Great view of the New Zealand countryside.
That was fun and interesting !!! and I love your plane design ..... I've run out of switches on my flysky so about to try a turn knob for the flaps...... Looking forward to your next video , good job Man 😉😎
Hi iforce2d, I'm not surprised you saw no difference between the props inwards vs outwards spin direction. The benefit of having inwards spinning props is for when you are in the climb and have a higher angle of attack than straight and level flight. In that case the centre of thrust is moved towards the downward moving blade of the prop - and for the inwards spinning props that means towards the middle of the plane. Therefore you get a lower yaw moment, and a lower Vmca (velocity of minimum control airplane). This is likely much more pronounced in full sized aircraft which will have a MUCH lower thrust/weight ratio, and therefore a much greater angle between the incoming airflow and the aerofoil in a max performance climb. With the amount of thrust of an RC plane the effect might be quite limited. If you want to see the difference between the inward vs outward spin direction I'd suggest you repeat this test in a steep climb. I'm not convinced even then for an RC plane you'll see much of a difference, but perhaps in the telemetry?
I have never thought of this, thanks very much. Then again, i only fly single motor aircraft :D
Another amazing video! I could watch hours of this channel
To answer your question in some other data that might be interesting.
-looking at the pitch, roll, throttle curves is interesting, bit I wonder if looking at the pitch, roll, yaw ERROR might highlight any sensitivity.
-Looked like there was a slight difference in turning radius at each waypoint. Wonder if this is significant.
-Current being about the same is interesting. Theoretically makes sense that for either case, it would take the same energy to keep the bird up, but you might expect the single prop configuration to fly a bit more inefficiently. Maybe your vibey motor is adding inefficiency of its own.
-A control lap with both props left untouched to compare to the prop off runs would be nice.
This one seems to have a nice enough thrust-line that it barely notices a motor-out. But does that also make adding differential thrust in the control mix to boost yaw authority a moot point?
Man, you fly in amazing place!
Hi. Do you know website or youtube that has a clear step by step to build ArduPilot RC Plane?
Thank you
Haha... finally, a use for those knobs! 🙂
Hi, thank you for show beatiful plane and excellent tests. I am struggling with a KillerWhale plane to make auto landings. Do you have some recommendations or could share some landing parameters. Please continue to share your results, thank you. Regards from Bo in Sweden
Hi Bo, I have not done any auto-landing, but if I put landing gear on this plane someday, maybe I will :)
You should cut some scoops out of the fuselage and move the propellers closer...maybe even a little bit _into_ it.....then have those scoops shaped so they smoothly direct the airflow under the wings at the root.
and maybe even slightly re-design the wing root to make a bit of downward thrust with that airflow (or adjust the flaps to take better advantage of it for just making that thrust during takeoff and on final). This may be outside the intended flight characteristics here, but it sounds fun lol.
Add this data point to your calculations:
Ardupilot powered VTOL Ranger T1
Q-Loiter mode, taking off. Get to about 10M, and starboard ESC decides to identify as roadkill and lets the magic smoke out.
Interestingly, AP caught the failure and kept the vehicle shiny-side-up during the decent, gliding it into a gentle landing
Pretty cool. Not an RC guy yet, but seems like you could use a little rudder input programmed in for asymmetric thrust.
Yes, if you wanted to do this intentionally for some reason that would be helpful. The question I was trying to answer was how well Ardupilot manages when a motor fails unintentionally, without any specific 'programming' to take care of it.
I see. I find it fascinating the stuff you're doing. In airliners if we don't apply rudder or rudder trim the autopilot will have a hard time compensating. And it looks like Ardupilot is basically an autopilot from what I can tell. Waiting for the retracts you talked about! :) @@iforce2d
but when motors are properly matched to their esc and load, they never fail. Props do fail a lot though on RC. Am I right?
Great aircraft :) I have been building my designs for the last few years with literally same technology, orange foam ("penopleks"), fiberglassed surface and multiple sections on a wing cut from profile shapes. the only thing is, I have no idea how to paint them or make them look nicer. How would you go about painting one of your models to make them look all nice and shiny? :)
I actually made a video all about my attempts to paint these. It didn't go very well, mainly because of the fiberglass fabric pattern on the surface. I can't really sand the surface much because I usually only have one layer of glass and it would be worn away too quickly before becoming smooth, plus the epoxy is very hard. If you want a truly shiny surface I think you need to make molded parts, instead of this method.
ua-cam.com/video/7Fjj5aLm2Qw/v-deo.html
Hi mate, where do you usually get your batteries from in NZ? I love in South Auckland, and I used to get them from Hobbyking HK warehouse, but they don't ship from there any more.
I actually haven't bought any lipos for a long time, maybe two years. But I used to get them from HobbyKing Australia warehouse. In recent years I have only bought li-ion cells on TradeMe and built my own.
which one is the lift motor?
How about a moving average filter block for the motor speed.
What do you need in order to run Ardupilot? A pixhawk? I wonder, why running Ardupilot is not possible with a regular embedded linux device, such as a Raspberry pi?
ardupilot.org/rover/docs/common-autopilots.html
@@iforce2d Thanks!
What a beautiful plane! 🎉
HI, I BUILT nrf24 transmitter and receiver which output ppm using your sketch. What i want is to get my quad battery voltage to display at transmitter on an oled screen.
May you please help how to configure this ❤
ua-cam.com/video/lf7bO5KO6rc/v-deo.html
ua-cam.com/video/OC5KFWCgRdM/v-deo.html
Really interesting, I see ardupilot in my future
have you used ardupilot yet? did you succeed?
Where did you flew it? The landscape is amazing!
With such large propellers I would expect some gyroscopic effect that could make some difference between clock- and counterclockwise rotation in the case of one engine off
The torque is the problem not really gyro effect imo.
Thanks!
How did you put your "rc input" into video? How can I do this for my videos?
Not easily. I wrote a program to read in the iomixer log and generate 38377 png files from it. Then I used ffmpeg to join the png files into a video. You can use things like Dashware and the regular Ardupilot log for a more typical approach.
Hi, is it possible to get some files for the plane ur using?
www.iforce2d.net/sketches/stackedSlices/
Why are you using such a cheap radio? There are better radios with longer range and more options for what you are testing. Your plane looks awesome BTW.
I have never felt this radio to be inadequate.
You really should get a Tx with Open Tx Firmware on it, you can do all that mixing on the Tx, without all your extra bits of mixers etc.
Or he can flash opentx on his Flysky I6X. :)
If you angle each motor outwards towards the tip of each wing at approx. 5 degrees, it will assist flight stability greatly with the loss of one motor.
Respect!
The twin motor and props are ok but plastic prop loosing efficacy through vibration and flex, change them and see
This is like GNU Radio Software program!
Nice plane but that FPV runway..... *drools*
I might have to make some landing gear for this thing and do it properly sometime :)
Why'd you laugh? It is well designed. :)