What happens when a rotor fails...

[Dependabletv]

Hi All
I'm relatively new to multi rotor flying and plan, in time, to add aerial filming to the services offered by our video production company in the UK. We're currently training very successfully on a DJI 550 which is a fantastic machine and makes me very excited about the possibilities a larger machine will offer. We have plans to put a high end camera onto a larger machine, possibly with the new Movi gimbal. Time will tell on that front.

As part of our training, and ultimately risk assessment, I'm keen to understand what happens if we get a rotor failure. Apart from the obvious answers (you crash!). I presume no element of control remains or does attitude hold at least attempt to keep you level. Do the remaining rotors allow controlled descent speed albeit without the ability to control direction? Does more rotors or coaxial rotors make a difference?

I'm sorely tempted to take our 550 to a big open space, take off a rotor and hit full power but I'd rather find a more scientific and less potentially expensive way to learn. Do any simulators offer the ability to simulate failures. And finally, does anybody have any experience of any emergency parachutes for such situations.

We are hoping to fly a seriously expensive camera and want to explore all possibilities. All experience and advice very gratefully received.

Regards
Richard

 

[modles]

If you're lifting anything expensive, use an octo. From experience.

 

[Dependabletv]

Hi Modles
thanks for the reply. We'll have to use an octo for both payload capability and stability irrespective of any safety concerns. Can you elaborate on your experience? Have you had a rotor failure and if so, what happened? I want to be able to add a section on this into our operations manual as well as being as prepared as possible.
Thanks
Richard

 

[Dewster]

DJI implemented a kind of fail safe if you lose a motor, rotor, or ESC on a HEXA. It will cause your craft to spin in an effort to counter the loss of one control arm. You'll have to switch to Intelligent Orientation Control so that you can guide your spinning craft to a safe area to land. It may be just enough to reduce the amount of damage to your craft and property. Before the firmware update a loss of a motor, rotor or ESC was an assured crash. I've seen videos where operators were able to save their craft before DJI's update and many where they have not. You'll need to make sure that you keep your craft balanced and light as possible with your camera gear.

I experienced a bad crash which almost totaled my flat Hexa when a prop broke. I decided to follow a trend and thought I was upgrading my props when I switched from APC to carbon fiber props. In less than five minutes of flight, the carbon fiber prop broke. I heard a snap, all the motors on my craft went crazy as my flight system tried to stable the craft. DJI WKM has/ had a feature that would kill the motors if the craft rotated over ?? degrees during Attitude, GPS mode. It was a safety feature. My craft dropped straight from the sky after a brief flutter of activity.

As a result of my crash I wanted to have one that would survive loss of a rotor. I went with a Y6 design (this was before DJI's spinning recovery update). Some say the only true redundant design is the Octocopter. Your chances of recovery on a Hexa will depend on the weight of your craft, what you were doing when you lost your rotor ( fast forward flight, altitude etc), pilot experience, and lots of luck. I run my batteries in parallel, I try to keep my craft light and balanced, and most importantly: purchase good props. Most crashes are from flying with cheap (in design, not cost) props. Don't just look at the material that they are made of, look at their design. Read through the forums. Many will try, crash, burn and report back on the quality of defective props.

My prop induced crash of my flat Hexa destroyed my FPV gear, the entire frame, two out of six motors and one ESC. The only thing I needed was a partridge in a pear tree. Lol. My Lipo was also destroyed. It was a 6000 mah that was punctured. A ton of heat came from the ruptured battery, but it still provided power to some of the twitching motors.

Ive seen tests where a prop would be removed from a craft to see how it will fly. Most of the tests involved just hovering in place. A true test would be for it to be in motion, with weight to simulate camera gear etc.

Best advise is to use good props (and price is not a determining factor of quality).

 

[modles]

Yes some good advice there. Here's mine...

If you are running a hexa and you loose lift from one prop (for whatever reason). DON'T RELY ON THE DJI FUNCTIONALITY!

Switch from gps mode to attimode, but DONT enable IOC, doing so will tempt you to try to steer the craft, best thing to minimise damage is to bring it down spinning in a straight descent. If you are coming down over water or obstacle, then control very slowly and gently, once you are about 10m above. Also, dont try and bring it down to fast.

If you try and control it laterally unless it's a deadly calm day and you are not heavily loaded it will oscillate out of control. I found that the manic corrections of the zenmuse is what caused it to flip.

After you've done all that, go out and get an octo, but even then follow the same idea.

Also a good point on the intelligent motor control, i would agree its best not to use that.

 

[modles]

out of interest does anyone know what firmware version their 'motor failure saving' feature was introduced?

 

[Dependabletv]

Hi Dewster,
This is really helpful stuff and good advice. Thank you. Clearly, weight will be a big issue for us. Sadly, there is no such thing as a light broadcast quality camera so we will be on the heavier side.

Am I right in thinking the "fail safe" feature of an octocopter is that it can effectively kill an opposing motor and still work as an, albeit, dramatically out of balance hex? If so, what are the implications for a co-axial octo? I'm guessing (and that's the problem here isn't it) that it could probably cope but not as well?

can you point me to any of the videos you refer to. I've really struggled to find anything about this issue
thanks again
Richard

 

[modles]

It does not kill the opposing motor, it doesnt know which motor is failed as they are essentially dumb. My best guess is that it looks for a build up in the Integral accumulator in the yaw PID loop and when that goes over a threshold for a period of time it knows something is wrong and there is and induced yaw due to motor failure. Well that's how I did it on my home brew controller anyway. You then just increase the yaw in the opposite direction.

A coaxial octo should be just as good as a normal octo in this arrangement, but remember that the efficiency of coaxial octo is not as good as a V or + setup

 

[FerdinandK]

On a quad you have almost no chance of saving the craft in case of a motorfailure (on one motor/prop).

On a hexa you can save the hexa if the craft is able to fly safely with 4 motors. So your hexa should be well overpowered, and the best chances you have in "manual" mode, where you have full throttle range, and you only need to compensate with help of the gyros. So a hexa needs to be well overpowered (should hover at 1/3 of max trust available). There is no difference between an Y6 and a hexa, both have six motors, and both behave the same. Typically Y6 are more likely overpowerd (since the might have been tris in the past, and/or the frame is much smaller and lightweight compared to a flat-hex).

The more motors you have, the better your chances. You can save an octo with a motor/prop (one) if you can fly safely with 6 motors. Here it can happen, that you do not even recognize a motor-failure (e.g. when you fly fpv), but you can hear it, it is a very different sound compared to normal. Typically all props "sound" on the same tone, if you loose a prop/motor, you hear an "interval", since one part of the motors are running at different rpms to compensate.

In the beginning I experienced two motor-failures on a hexa wich lead to a chrash (not dropping from sky, but incoming fast). Both hexas were flown on the very limit of there capabilities, so no reserves in thrust. The next hexa with a motorfailure I could save, it was well overpowered and the motorfailure resulted in a strong drift to one side (but it was controlable in manual mode). I could land it with only small damage, one leg of the landing gear broke). The next motor-failures I had on an X12 (2 bells went off) which I could land without any problems. My last experience was loosing a prop on an X8, were during flight (FPV) I only noticed that it was a bit slow in full foreward (and sounded "different"), which was explained to me after landing, since one rear prop was missing.

The more props the better your chances, but your best chance you have with training to fly in manual mode (on whatever FC you use). Do not rely on "special features". If you "feel" a drift, do not fight against the drift, just use the "drift" to bring the copter home (by yawing the copter). If it does not like to yaw in one direction, use the other. Do not panic, look - think - act, you can panic if the copter is down. You (your pilot) have to exercise this, by flying close (and above) the limits with smaller copter. Starting with one prop less is completely different, than loosing it during flight, since the most unstable moment is when the FC has to correct itself after loosing one motor (the best with your help in manual mode).

Here again my video of my X12 using 2 motor in flight:
https://vimeo.com/47266306
and just to state, that a prop/motor failure is not the only unexpected thing that can happen
X12 with propeller icing:
https://vimeo.com/56590314

best regards
Ferdinand

 

[Dependabletv]

Hi All,
This is such useful stuff thank you so much. Modles, I hadn't fully considered the implications of the gimbal in a fail state, let alone thinking it might cause a flip! That's really not great! But I can see how its attempts to counteract the movement would clearly add to an already bad situation.

So I presume the more rotors you have the smaller the problem become in relative terms - crudely, on an octo you lose an eighth where as you lose a sixth on a hex? As I say we plan to work with an octo so that's a good thing! The reason I ask about coaxial is that we've been advised that they can be more stable in poorer conditions. That's probably a subject for a separate thread but good to know that it's not a more dangerous option.

Ferdinand - FPV in fog! I think I'm relieved to say that's somewhere we won't be going as we'll be restricted to 400ft/500m or line of sight. But it's great to see you manage to recover each situation. Do you think that losing two motors actually helped you? Were they an opposing pair? It may be that this put you in a better position than losing just one? You've also made me realise that I really do need to be able to control the machine in manual mode. That's somewhere I haven't gone yet as the attitude and GPS modes are so good, switching them off seems counter intuitive. I was also advised against it by the supplier we are working with.

That's also a good point to make here - we are working with a great BNUC-S licensed supplier to build the machines for us rather than attempting to build them ourselves (although I have had to rebuild most of the f550 to add landing gear which was a really useful exercise)

Keep the comments coming. This is brilliant!
R

 

[FerdinandK]

On a hexa you do not loose 1/6 of the trust, you loose 1/3, since there are only two turning in one direction, compensating 3 turning in the other direction. So if you loose (let´s assume) a CW Motor, the CCW Motors have also to reduce thrust to compensate the yaw momentum.

On an octo you loose 1/4 of the thrust ...

A pilot on any aircraft will not manually fly the aircraft if not needed (or just for fun) only if the situation requires more attention (takeoff and landing) or in case something goes wrong, he has to take over (and is not starting to reprogram the autopilot). In a Simulator, they do not train how to relax while using the autopilot, they train, what to do if one engine stops, what to do if they loose one aileron, they just train, what to do in case (one or )some of the systems are failing.

On the copter, flying with GPS / Atti is a starting point, but if you want to be able to handle unexpected situation, you have to be able to go back at minimum FC-suppurt (cause all the rest might have gone crazy).

best regards

Ferdinand

P.S.: Of course if you are not able to control the craft in manual-mode it is not wise to switch into this mode (as suggested by the supplier). Typically you should also not start to learn how to fly manually with a okto with a cam underneath. That is also the reason, why all the people who are a bit longer in the MC-scene tell, that you should start with something small and cheap to learn, and then start to do fly a heavy lifter with a cam attached.

 

[RCJardin]

This an interesting thread as I am in the same position of enhancing our broadcast camera services. Perhaps the just announced Black Magic Pocket camera will become the new 5D, in which case a bit of weight will be saved and we will have broadcast 220mbts Pro Res HQ pictures shot on a camera designed for video.

 

[modles]

Yeah make sure you have a disposable toy to keep your eye in with. Fly in Atti over GPS whenever you can and be sure to practice in manual when possible. These light evenings provide lots of time to get yourself into trouble (and get yourself out of it).

I find to simulate a motor failure in flight, just losen one of the prop nuts (wont work with notched dji props), then you are pretty much guaranteed that at some point it will come loose

If you come from a background of flying RC heli's you shouldn't find manual mode a challenge.

When are you booked in for your BNUC?

 

[Dewster]

Hi Dewster,
This is really helpful stuff and good advice. Thank you. Clearly, weight will be a big issue for us. Sadly, there is no such thing as a light broadcast quality camera so we will be on the heavier side.

Am I right in thinking the "fail safe" feature of an octocopter is that it can effectively kill an opposing motor and still work as an, albeit, dramatically out of balance hex? If so, what are the implications for a co-axial octo? I'm guessing (and that's the problem here isn't it) that it could probably cope but not as well?

can you point me to any of the videos you refer to. I've really struggled to find anything about this issue
thanks again
Richard

Here are a few vids

DJI prop break on Hexa (long video crashes in ocean):

http://youtu.be/Yr_aGFdRys4

Hexa lost prop in flight (best recovery video):

http://youtu.be/qrMuROJIJLY

another prop failure video (the owner's opening comment had me laughing):

http://youtu.be/VVv_xRudghU

DJI Y6 motor failure test:

http://youtu.be/vvnbme5_R0Q

DJI flat Hexa motor failure test:

http://youtu.be/amwizUBkCPA

DJI Octo loses prop in flight (you'll see the prop come offin flight)


http://youtu.be/GKd_IJSCLyU

 

[Dependabletv]

Thanks Dewster - I'm not sure whether to be encouraged by the saves or to give up when I see the fails! I take some comfort in the fact that by considering this at this stage I hope I'm taking the issue seriously enough, and treating it with enough respect to stand the best chance of coping with it.

I have flown smaller "indoor" heli's for a while so the transition to the F550 was actually a joy. Much more controllable but, as I've said to many people, I'm glad I started with a basic knowledge as the consequences are so much greater with something bigger. Saying that I am very much a TV Producer/Director first and foremost and we have both the cameras and filming experience to cover that side really well - and to know what broadcasters and film makers will want from the footage - both visually and technically. This is the most fantastic way of potentially combining work with a hobby as well as getting stunning images. I used to be heavily involved in aerial filming for the BBC so know both the capabilities and limitations of a conventional helicopter. I don't believe for one minute that these will ever replace full size helicopters but they do provide exciting new possibilities that can't be delivered in any other way.

I had been put off manual mode as the guys that built the F550 told me to avoid it - initially at least. I then accidentally had it switched on before take off recently and it took me 30 seconds of wildly spinning props to work out what was wrong. I presume this will turn off everything that makes it so easy to fly? So it will invert if I tell it to etc etc. Not sure that's a roller coaster I want to ride yet but I suppose I should - and if I switch it on with some height to play with I can always switch it off again.

BNUC booked for early May. Can't wait to get stuck in.

This is a great resource. Thanks again for all the advice and info,
Richard

 

[Dependabletv]

This an interesting thread as I am in the same position of enhancing our broadcast camera services. Perhaps the just announced Black Magic Pocket camera will become the new 5D, in which case a bit of weight will be saved and we will have broadcast 220mbts Pro Res HQ pictures shot on a camera designed for video.

This is probably one for a separate thread but the new Blackmagic camera is interesting although I think it's smaller size comes with a small sensor. The best analogy I've heard is that it's like super 16mm. I think it may also be a CMOS sensor with the inherent rolling shutter issues which aren't great on moving cameras. But then it's a really great price and will almost certainly deliver stunning images. I think it just enlarged the minefield again!

 

[modles]

Sounds like your approaching this with the correct level of planning and respect

Best of luck and keep in touch!

 

[3DJIM]

Great thread, Subbed . The bottom line is make sure your payload & multi are insured cos crap is gonna happen!!. An interesting note is that our excess doubles once we are airborne !!


Jim.

 

[Dependabletv]

Hi All,
Just to report back an evening of manual mode flying without incident. Just like going back to the mini heli I have. Much less relaxing than attitude or GPS clearly but not as big a deal as I'd been lead to believe. (I had images of the thing either turning upside down or heading for the horizon) It does ascend rather quickly though!
R

 

[Dewster]

This an interesting thread as I am in the same position of enhancing our broadcast camera services. Perhaps the just announced Black Magic Pocket camera will become the new 5D, in which case a bit of weight will be saved and we will have broadcast 220mbts Pro Res HQ pictures shot on a camera designed for video.

I think there is a real market for light weight aerial camera systems. You'll definitely see professional cameras designed for multi-rotors when ever the aviation authorities come out with a plan for commercial UAS.