I was wondering if maybe the motor had the UAV equivalent of a bird strike...maybe something got sucked into the motor? But then I realized that if this was one of the bottom motors in a X8 config, that’d be all but impossible. Without any data logging, it’s going to be very difficult to reconstruct what actually happened. Are you going to send the motor to KDE for them to examine?
Watching these videos makes your heart race and anxiety spike.. Dude I feel for you! Had my copter freak out a few weeks back.. Definitely not fun.. Does your copter rector flight data?
You raise a good question, James. Bearing in mind that neodymium magnets, at least the ones I've trashed, are (a) very brittle, and (b) actually made of what looks like a compressed powder, I'm not quite sure how abrasion resistant they are. Do you have more experience on this that you can share? My only source is Wikipedia which says: "The ingots are pulverized and milled to tiny particles, which then undergo a process of liquid-phase sintering in which the powder is magnetically aligned into dense blocks. The blocks are then heat-treated, cut to shape, surface treated and magnetized." http://en.wikipedia.org/wiki/Neodymium_magnet I've had situations where even fairly gentle abrasion causes a magnet (not in a Brushless motor, mind you), to wear away like sandstone.... As you say, assuming for a moment, that the motor was rubbing before impact I would have expected it to either (a) sound like a rather upset banshee, or (b) grind almost or actually to a halt.... I don't hear much of that in the audio -- it sounds more like decommutation where one motor is slowing down, but it's a more constant noise until impact. Andy.
I visited Bryan last night to assist him with a small project and I picked up his Gryphon to assess the damage and to get it repaired. It is my opinion as well that the motor shaft broke in flight. There is some pretty clear evidence leaning this way. The motor shaft is broken at the corners of the flat spots where the set screws tighten against the shaft. The machining of the flat spots and the set screws do create stress points and weaken the shaft at that location. In the video you can audibly hear a change in tone that occurs around 2:33. You can also see jello begin in the video at that time indicating additional vibration. Neither of the props on the arm with the broken motor were damaged indicating that they didn’t hit anything solid enough to break the shaft. The wear inside the motor is indicative that it the motor ran for some time without the bell being aligned to the stator. The polished appearance around the bottom of the windings would not have happened from the crash impact. The motor would have stopped spinning too quickly. As you can see from the pictures, the motor shaft is now loose and has slid out the back of the motor. Since this was a bottom motor pushing upward, the bell stayed on and was pushing against the bearing. There is some wear on bell housing where it would have been pushing against the bearing. Why would the shaft fail? Striking something abruptly – This does not appear to be the case as the problem started in the air and there was no propeller damage on that motor. In addition, large forces from an impact tend to bend things. The shaft slid back into the stator stack with a clean break and is not bent. A defect in the metal, metal fatigue, and/or a defect after machining the flat spots are just a few ideas that seem more likely to me. The clean break on the shaft looks more like metal fatigue to me. This aircraft has very little run time on it though. What is next? The motor will go back to KDE Direct for them to examine. Unfortunately, the Gryphon went into salt water and many of the ESCs are crispy. It looks like some have survived and we may be able to pull data from them. We will survey the rest of the electronics and frame and see what needs to be replaced. The MOVI will be sent back for repair/replacement and hopefully some mercy will be given. Do what we can to learn from this.
Thanks for the posting Tim. Excellent images, too. I think we need Patrick's input on this. One question that I don't know the answer to is raised by your comment "Neither of the props on the arm with the broken motor were damaged indicating that they didn’t hit anything solid enough to break the shaft." What I don't know is whether a prop would be strong enough to communicate a shaft-shearing force to the shaft without the prop itself breaking? If the shaft's breaking absorbed the impact, could a prop survive? Do you have any data on this that you could share? One thing that does emerge from this, given that one motor stops on an X8 is that it seems that an X8 responds less well to a single motor failure than a flat 8? (Note: this is a question, not a statement!) Steve: What's your take? Andy.
For what it’s worth, I am not sure I want to be the first guy to post that a motor and/or ESC went out on my X8 and it handled just fine. Some of us (myself included) have experienced the loss of a motor (it was a BL in my case) in flight on flat 8 configs, and at least in my case, it was barely noticeable. No yaw disturbance, and because I wasn’t near the power redline, the loss of 12.5% of my thrust also was not a catastrophe. I don’t think the loss of a motor on an X8 would be any different, but that’s more a hope than a statement of fact. It’s clear from Bryan’s video that the copter was yawing significantly (look for the copter’s shadow) before it crashed, which leads one to believe that maybe this theory isn’t true. But this brings up the question of how the copter’s instruments sense the loss (or impairment) of a motor other than simply detecting a yaw and correcting it. Bryan, which A2 mode were you flying in? Does the A2 have a fully-manual mode? Is it possible that in this mode the A2 is not able to detect and/or compensate for the loss of a motor? From what I know of MK’s flight control algorithms, there isn’t really a “fully manual” mode. The flight control is always doing some manner of tracking of the copter’s movements, and therefore will send power as needed to straighten out the copter, and if a motor’s not being responsive, will try other motors. But maybe I’m wrong...happens all the time.
WTH!? Are those motors not equipped with a hardened steel shaft?? That shaft looks like some sort of cheap brittle metal. Axi motors use hardened steel shaft...unbreakable. Also, KDE MR motors utilize 4mm shafts, similar to all other heavy lift series motors...i've always thought 5mm should be the minimum, Axi motors (*2800-5000 series) use 5mm & *6mm hardened steel shaft. Patrick Koegler, is this really the case? These motors are not equipped with a hardened steel shaft?
Regarding the data on props being strong enough to transmit forces to the shaft. All I have is my thirty years of flying model aircraft of all types and witnessing many crashes. In general when a propeller strike is involved to the point where a motor or engine gets damaged, the propeller does not survive. I have never seen a motor shaft sheer off while the prop is unscathed. Let's remember that these are Xoar wood props and that they are pretty fragile. The propellers on the other side of the Gryphon broke, but the motor shafts did not. Let's take a look at the brushless motor design. The only real loads on the shaft that broke are the forces needed to hold the propeller and and bell in alignment to the stator and motor mount. It doesn't transmit any torque, so even if the propeller and bell were to stop, the only mass that the shaft needs to stop is the shaft itself. On the other hand fatigue caused by cyclical loading does not need to involve large forces. On the flight controller question, I believe Bryan was in attitude mode which is self leveling. The DJI Naza, WK-M and A2 controllers have algorithms to compensate for a motor loss. As long as there is sufficient headroom on power, the aircraft will fly normally. If the power is marginal to maintain altitude and to keep the craft level, the first thing the controller gives up compensating for is yaw in order to keep the aircraft level. I have purposely tested my X8 with one motor unplugged and it flies very well as long as the payload is not too great for the available power. The A2 does not have a "full manual" mode. It is more like an angle lock/heading hold mode. It has a manual mode, but it is still stabilizing.
I would say there are a lot variables that will never give us the true answer to this accident. I have to assume that the copter had a blade strike prior to this flight that caused a hairline crack on the shaft. I have never seen a shaft fail except when their was a previous blade strike. I have had several shafts break on fixed wing planes. Blade motors and Himax motors shafts failed but only after a blade strike or a crash. I even had one himax shaft break in mid flight, sent the entire bell and prop sailing off into the bushes. Again, it takes alot of force to shear a shaft like that. I have seen a motor fail in flight on a multi rotor due to small magnetic stone kicked up from friend taking off directly on dirt ground. The stone must have been between magnets for awhile because the copter was flying along fine and then all of a sudden the motor came to an abrupt dead stop in flight on a F550. when we took the motor apart, the culprit was wedged in the bell housing. Water is very hard like Patrick said and last week I was flying fixed wing and thought I would skim lake tahoe, this did not work, the wheels grabbed the water and pitched the plane forward and the prop instantly (APC) shattered into multiple pieces, But the shaft on the Himax motor was fine, not bent at all. dried the electronics for a coulpe days and it flys fine. I have seen wear on the dji stock motors for the f550 due to large dust getting kicked into the motor and it kind of churning around in the bell housing causing wear marks. Can you please take the shaft out and test its hardness by setting it in a vice and tapping it with a punch, you can tell right away if its hardened or not.
I for one am grounding my X8 4014 until we have more data about the motor evaluation. Too much investment flying around as I have to have 200% trust in my stuff. My X8 is my spare copter because a like the better flight times of my flat 8 so I can be afforded some downtime with the X8. I too would like to know if the shafts are hardened steel and would like to see the failed one tested. Bryan since you are running 17's and this motor is designed for 18's could it be that there's extra RPM-torque stress on the shaft if they are not hardened? I also noticed that your 17's have a 5.0 pitch which would make the motors run even that much harder IMO.
Wow, this is not very encouraging, you are basically flying around my upgrade list. I was looking at those KDE motors and would like to see what they are made of before committing. I also would like to find out if you don't mind how much it will cost for the MOVI rebuild. I have been flying over water lately and that question is always in the back of my mind.
Even very high RPM should not lead to broken shaft if the props are fairly balanced. I would never think my motors' shafts would broke until seeing this. Is there a possibility that specific motor's shaft was manufactured with errors and couldn't be realized in inspection by KDE? I also would like to hear from KDE about this shaft problem since we transformed all of our heavy lift copters to KDE motors. An inspection from Patrick would be appreciated a lot.
Bryan, I'm also very sorry to hear about your crash. The most important point here is that no one was injured. I'm also running an X8 with these motors, so you have my complete attention. I'm wondering if there is any way to bench check these to identify any issues. Anyone have any ideas? Also, has anyone specifically tested running an X8 with the current MK stack with one motor out? -m
Bryan, I for one am deeply appreciative of you sharing your experience with the forum here. I hope what you have learned so far is adequate compensation for the effort you've put in to this process. I hope we can continue to contribute to your efforts.
HI Bryan I don't think anyone is bashing KDE and I didn't take your post as bashing anyone. I would have done no different had it been myself. I for one love to have KDE around compared to Tiger because of their attention to customer service. I really like the improvemments of the KDE over the Tiger motors. If there's a weaklink in these motors its not any blemish on KDE as I feel they will do what ever they need to do to make it right. I think KDE has already improved upon many of Tigers design flaws and if they need to make a change to improve it that much more I'm sure they will do it.
Bryan: I am very sorry to hear about your crash. Thanks for all the information being shared. I, too, fly these motors in an X8 configuration and will be taking a good look at my set up and am curious to know how the copter should handle in a typical motor loss scenario.
It should be said here that we must all remember that these copters have to be able to work when circumstances are less than perfect. The must upsetting thing about this shouldn't ncessesarily be the possibility that a motor went bad in flight. That should be expected on some level...they're mechanical devices and they are subject to pretty astounding amounts of force, and even a slight defect can be a disaster. It's like a Formula One car versus your family minivan. Any slight imperfection is dramatically multiplied at 10,000 RPM, 800 horsepower and 200 MPH. A similar defect could lurk for months in your family car. And we're not flying minivans...these things are closer to race cars. The hope is (was) that the redundancy of 8 motors would save the day, and that may be a false assumption. I wonder if somebody can reach out to the Kopterworx guys and ask if they have any experience with the loss (or impairment) of a motor on one of their many X8 copters. Or anybody else with experience with X8 configurations? EDIT: And I must reiterate that my experience with redundancy is with the MK flight controller, and Bryan was flying a DJI.
I've started digging into the electronics on the Gryphon to see if the CC Phoenix Edge Lite 75's had data on them. Both of the ESCs on the arm with the broken shaft appear to have survived the crash and I was able to pull data from them. What I didn't realize was that once the data log is full it stops recording and the data has to be cleared for it to log again. So the data on the log file goes back to the initial testing of the machine and has no recent data. I expected it to be more like a black box and have data for the most recent flights. On top of that the data clearing process is not practical for multi-rotors. It involves powering up the ESC at full throttle setting for a number of seconds. With the flight controller in the mix this can't be done.