Welcome to InspirePilots.com

Join the leading DJI Inspire community for free!

Defective motor or ESC ?

Discussion in 'Inspire 1 Help' started by Sciro16V, Jun 14, 2015.

  1. Sciro16V

    Joined:
    Mar 29, 2015
    Messages:
    12
    Likes Received:
    2
    Hello. Have a question to inspire users about stoping spined motors. When i land and stop motors left stick down, 3 of them are stoping almost immediately and one of them not. Front right one is spinning much longer than rest. Is it normal behavior or i should worry and resend my inspire to dji? All motors are spinning same strenght and not locking or sound diffrient. Sorry for my english. Thanks for answers. If sumeone want to see video of my problem i can make one fast.
     
    #1 Sciro16V, Jun 14, 2015
    Last edited: Jun 14, 2015
  2. The Editor

    The Editor Moderator
    Staff Member

    Joined:
    Aug 7, 2013
    Messages:
    6,723
    Likes Received:
    3,880
    Nothing to worry about.
    As long as all your motors are running at the same temp and no particular one is running hotter than the others. (Check with your fingers immediately after landing and you can tell if any of them are running hotter) you will be fine.
     
  3. Sciro16V

    Joined:
    Mar 29, 2015
    Messages:
    12
    Likes Received:
    2
    OK thanks for tips i will check them asap.
     
  4. fei8820

    Joined:
    May 22, 2015
    Messages:
    297
    Likes Received:
    63
    Not to worries.... Like what the editior said .. I been running hobby shop since 2007 and selling some dji product as well.. I can say I never heard about motor or esc faliure ...
     
  5. Kauaigene

    Joined:
    Apr 8, 2014
    Messages:
    313
    Likes Received:
    96
    Then again...
     
  6. fei8820

    Joined:
    May 22, 2015
    Messages:
    297
    Likes Received:
    63
    Ops sorry for your lost... Did any other damages for this fall? Does the esc n motors faliure cause this ?
     
  7. Lee

    Lee

    Joined:
    Mar 20, 2015
    Messages:
    4
    Likes Received:
    0
    Did you lose a prop ?
     
  8. Kauaigene

    Joined:
    Apr 8, 2014
    Messages:
    313
    Likes Received:
    96
    No prop loss or another other damage. Did the failure cause the crash or was it the other way around? Thats the question and hopefully DJI will have the answer.
     
  9. IIIDaemon

    Joined:
    Jan 24, 2015
    Messages:
    70
    Likes Received:
    6
    Location:
    Texas
    .
    ...Sciro, just for grins.. 'Nudge' your right stick jussssss slightly see if that make a diff (chk all axis), you can try same while idling...
    .
    IIIDaemon
    www.GasRecovery.net
     
  10. IIIDaemon

    Joined:
    Jan 24, 2015
    Messages:
    70
    Likes Received:
    6
    Location:
    Texas
    .
    ...Kauaigene, your incident looks very similair to 'Ferdi's'.. sudden tumble out of the sky for no 'Apparent' reason.., how many hrs
    on the craft? did you happen to insp bearings afterward, any chance of some close up (macro) images of the motors & bearings..?
    .
    IIIDaemon
    www.GasRecovery.net

    (...i suggested it months ago.. we should ALL be recording these incidents WITH the craft's serial numbers...)
     
  11. Kauaigene

    Joined:
    Apr 8, 2014
    Messages:
    313
    Likes Received:
    96
    Less than 10 total....lots of trouble free flights though.
    Once the error came up it was straight into the box to DJI so no inspection...will pay more attention if there is a next time.
     
  12. IIIDaemon

    Joined:
    Jan 24, 2015
    Messages:
    70
    Likes Received:
    6
    Location:
    Texas
    .
    ...Kauaigene, Thanks for your reply, if you don't mind PLEASE let us know what DJI has to say about this or what they do if anything, thx again...
    .
    IIIDaemon
    www.GasRecovery.net
     
  13. Joe R

    Joined:
    Feb 13, 2016
    Messages:
    97
    Likes Received:
    29

    I know this is an old post but I may have the answer to your problem. Hopefully you have already figured it out and had it corrected.

    Most ESCs have the ability to engage an electronic "Brake" (electronically slow the motors down), this includes the Inspire. It sounds like the brake feature in the right front ESC is not working like the other three.
    That electronic braking after landing helps to keep the motor from spinning too long and thereby prevent producing voltage that feeds back into the ESC after shutdown.

    Let's say that you are flying along and want to stop the Inspire from moving farther by putting the stick back to center, have you noticed that the Inspire does not stop as good as it used to?

    TEST: You can try decreasing the brake strength in the app to see if the other motors spin longer at shutdown. Then try increasing the brake strength in the app to see if the other motors stop quicker at shutdown. If the right front motor doesn't change at all, the brake feature in that ESC is for sure not working. I would definitely replace that ESC in this case.

    NOTE: The electronic brake would not affect the performance of the motor during flight except for when releasing the stick back to center for stopping the Inspire from going farther than you want it to. The GPS positioning system helps with holding it in place but the braking strength is what stops it faster or slower depending on the settings in the app. The app setting is merely electronically adjusting or programming the brake strength in the ESC.

    The motor spinning longer than it should could eventually cause more problems with that ESC. It could damage the FETs (Field Effect Transistors)

    NOTE: Recent research has revealed that DJI did not installe Schottky diodes at the output of the ESC to the motor to prevent Back EMF (BEMF = Back Electro Magnetic Field, which is voltage generated by the motor feeding back into the ESC). DJI actually uses this Back EMF (BEMF) to help regulate the speed of the motor and the braking system more accurately.

    I'm writing this in case you haven't figured it out yet and to let others know that if the the motor spins longer than it should, there is a problem with the electronic brake in the ESC. In that case, the ESC does need to be replaced to prevent eventual total ESC failure.

    I hope this helps you and others that may have they same problem.

    I'm just trying to help by sharing my knowledge and experience.
    Just remember, if you have something is not behaving as it should, there "IS" a problem with it somewhere and eventually it will become a bigger problem!

    Regards,

    Joe
     
    #13 Joe R, Apr 14, 2016
    Last edited: Apr 14, 2016
    Mark.B.Barry and slim.slamma like this.
  14. Mark.B.Barry

    Joined:
    May 22, 2016
    Messages:
    2
    Likes Received:
    0
    HI Joe
    Thank you for shearing your knowledge with others...
    one of my inspire motors(actually forth motor) turned less than others and during flight back side of inspire is tilted and I have seen that drone move one-way in the air...
    do you think this issue is caused by ESC or motor?
    and shall I change them or not?
    best regard dear
     
  15. Joe R

    Joined:
    Feb 13, 2016
    Messages:
    97
    Likes Received:
    29
    Hi Mark,

    It's difficult to say really.

    It could be the motor, the ESC or the flight computer or even the IMU.
    A few simple do it yourself tests can be done to eliminate the motor in question as the problem. 1) You can rotate the motors by hand slowly to see if you feel more resistance in that motor. 2) You can start the motors, leave them at idle, then shut them down to see if the motor in question stops before the others. 3) Another thing to try is to fly it for a few minutes, then see if the motor in question is warmer than the others, it could be a fault with the windings getting too hot from bearing or debris restriction (caused by too much current draw).

    The ESC is another story, you'd have to physically measure the output voltage to that motor which would require some disassembly and some special measuring equipment, not just a volt meter.
    It could be that the flight computer isn't sending the correct pulse width to the ESC, the only way to tell that is to do a partial teardown and connect the wire that carries the pulse width to the ESC to an oscilloscope and measure the width, even then, you'd need to know what the correct pulse width is supposed to be.
    It could also be the IMU not sending the correct signals to the flight computer. That would also require an oscilloscope and you would need to know what it's supposed to output to the flight computer.

    I hate to say it but if the motor bearings or other motor related restrictions aren't the problem, the only choice is to send it in to DJI to have them test it.

    I have the test equipment for all of that but I don't know what measurement would be correct. The flight computer could be sending a signal at 1600µs but I wouldn't have any way of knowing if that is the correct pulse width. Frequency plays a role in this as well, in other words, is the pulse width supposed to be cycled at 60 Hz or at 300 Hz (times per second)? Example: Slow analog servos typically operate at 50 to 60 Hz while some digital servos operate at 560 Hz. The same applies to ESCs, I'd have to know what the correct frequency is (sometimes called the refresh rate, meaning how many times it updates in 1 second), Slow servos usually operate at low frequencies while very fast servos like tail servos on helicopters operate at very high frequencies .
    This is the type of information I haven't been able to get my hands on with the Inspires. DJI doesn't publish it and I haven't had any Inspires that I could tear down to find out how they work (at least the basics things like pulse width and frequencies).

    Edit: One thing I forgot to mention is the average flight time. If you're getting lower flight times and have good batteries, a motor with bad bearings can cause higher current draw which causes lower flight times. Usually only the motors can cause this unless the batteries are getting weak.


    I hope this helps!!

    Good luck!!

    All the best,

    Joe
     
    #15 Joe R, May 28, 2016
    Last edited: May 28, 2016
    Mark.B.Barry and EMCSQUAR like this.
  16. BudWalker

    Joined:
    Mar 27, 2016
    Messages:
    28
    Likes Received:
    4
    Another "test" you can do is look at the motor speed, motor commanded, and motor load values. These are found in the .DAT file which is on the bird itself (not the .txt file on the tablet). Go here to see how to retrieve the .DAT file. The .DAT file will also contain all kinds of info that can be used to see if and why your Inspire isn't flying straight. The .DAT file will be large so you'll need to provide a Dropbox link so that others can access it.
     
    Mark.B.Barry likes this.
  17. Joe R

    Joined:
    Feb 13, 2016
    Messages:
    97
    Likes Received:
    29
    Please pardon my comment if I happen to be totally wrong about this but doesn't datcom depend on the flight computer and IMU for it's calculated readings rather than actual measured readings?
    One tech guy told me that they use a system that gives an overall health reading of the flight computer and the IMU but the rest is just info that is projected and not actually measured other than current draw. Meaning there isn't an actual feedback system in each ESC or motor. I know the motors don't use hall effect sensors so the rpms have to be simply calculated by the flight computer.
    If the flight computer or the IMU is out of whack, wouldn't that mean that the flight data shown on datcom would be incorrect or at least some sort of interpreted "Guess"?
    As stated, DJI told me that all data is calculated by the current that goes through a current sensing circuit much like a resistor form of a current transformer (CT). If I have this right, I think it's actually an array of current sensing resistors that are used to measure the current draw for each ESC and subsequently, each motor. Are they feeding me a bunch of crap?
    Can someone tell me if there are sensors on every part of the Inspire other than current sensing circuitry and IMU data that is interpreted by the flight computer (and stored in the flight computer)?
    I was also told by DJI that the reason that most people don't ever find out what actually happened to their Inspire or Phantom for that matter is that they don't really know because of the lack of individual sensors. All they can do is see that something was wrong with the flight computer or the IMU.
    Oh yeah, almost forgot, it does calculate voltage as well if I was informed correctly.
    Bottom line is, I have a close friend who works for DJI which is where I get this info. I can't say his name but he is one of the techs at the US repair center.

    But, If someone can tell me that there are sensors in each component, I would appreciate it very much!!

    Thanks!!

    Joe
     
    #17 Joe R, May 28, 2016
    Last edited: May 28, 2016
  18. BudWalker

    Joined:
    Mar 27, 2016
    Messages:
    28
    Likes Received:
    4
    DatCon was implemented by reverse engineering the contents of the .DAT file. So there is always some question how the P3 or Inspire acquired, processed, and stored the values that appear in the .DAT. How well a value is understood depends on the value. In addition to the values that DatCon extracts from the .DAT file it also generates several values that are useful when analyzing a flight. This is all documented here.
    Maybe you're referring to the roll, pitch and yaw values that occur in one record type in the .DAT. These are almost certainly the FC's predicted values. But, that same record type also has accelerometer(X,Y,Z), gyro(X,Y,Z), and magnetometer(X,Y,Z) data which are almost certainly the actual measurements. Barometric altitude is kind of interesting. When the A/C is powered up one of the first values to appear is the baroAlt. But, baroAlt changes to have a value very close to the gpsAlt when gpsAlt becomes available. I'm supposing that the FC uses baroAlt because it's smoother, but it has to be corrected by the gpsAlt which is much noiser.

    I don't know that much about how ESC works. The motor speed commanded values have to be values generated by the FC. I suspect the motor load value is some kind of current measurement. I don't know how motor speed actual is measured, but the units is RPM. The motor data has been used on several flights to determine if a prop broke in flight instead of breaking when the A/C hit the ground. Here's an example.
    upload_2016-5-29_6-16-59.png
    When the prop broke the speed and commanded went to the max values. The load was reduced due to less load because the prop had broken

    In the next example the pilot reported that the A/C would sometimes attempt to motorStart and then shutdown reporting an ESC failure. We looked at the last flight he had
    [​IMG]
    Although the motorSpeed was being maintained the motorCommanded was going to the max more often. At the same time the motorLoad is going to 0 more often. This suggests that motorLoad is a current measurement and that there is an intermittent open connection in the ESC circuitry for that motor.

    Here's another example that isn't so clear. Maybe you can provide some insight. The pilot reported that the A/C rolled over and flew into the ground. It's a long story but this flight had two incidents where this happened to all four motors. There are acceleration spikes associated with a spikes in the motorCommanded.
    upload_2016-5-29_6-39-8.png
    upload_2016-5-29_6-39-27.png

    I don't really understand why the acceleration spikes are associated with motorCommanded spikes, not load or actual speed. But, notice that the motorLoad goes to 0 during this interval. I concluded that it was some kind of ESC issue since it happened to all 4 motors twice in the same flight.

    Hope this answers some of your questions.
     
    Mark.B.Barry likes this.
  19. Joe R

    Joined:
    Feb 13, 2016
    Messages:
    97
    Likes Received:
    29
    Hi Bud,

    Thanks for the in-depth response! We really need more in depth responses along with associated material to study for the purpose of trying to better understand how it works.
    In my opinion, the better we understand how it works, the better our chances are of knowing when to fly and when not to fly.

    Most of what you have posted makes perfect sense but some of it needs more research on my part, especially the relationship between acceleration and motor command.

    In your first example where the prop broke (or came off) in flight makes sense to me. The corner of the aircraft that the prop was on would drop causing the IMU to command the FC to increase the voltage to that motor and thereby increase the rpms to raise that corner (or at least attempt to).
    There is a direct relationship between voltage and current. Higher voltage results in lower current (and visa versa).
    Just for clarification, a motor rated a 10 amps running at idle will likely only draw 3 amps or so but the voltage will be the same. You can load the motor all the way up to it's rated current and still have the same voltage, start going over that rated current by overloading that motor and the voltage will drop proportionally with the current and begin to overheat.
    The Inspire motors are 3 phase AC motors, the ESC inverts DC to AC which is why there are 3 wires from the ESC to the motor but only 2 wires from the battery to the ESC (not counting the PWM control wire).

    Since they are AC (alternating current) motors, I can give an example with a standard industrial 3 phase motor. Many modern industrial motors are brushless inrunners that can be controlled with a variable frequency drive (VFD) to obtain very low speeds and still retain high levels of torque. In the case of a VFD speed controller, the voltage is the same throughout the entire range of speeds from 1 Hz to 60 Hz, it simply reduces the number of pulses per second. The fact that the voltage is still the same at any speed is what allows it to retain the torque.
    If you run an AC motor on 240 volts that's rated for say 10 amps, you can increase the voltage to 480 volts if the motor is rated for it (dual input voltage) and the current will drop proportionally to 5 amps or exactly half the amps at exactly twice the voltage.
    That explains why the current was lower at the same point where the voltage was higher.
    Another way to look at it is that when the prop broke or came off, the current suddenly dropped and caused the voltage to rise at the same time but only since the command signal increased at the same time.
    Rather than voltage and current leveling off with each other, the voltage went high in an attempt to raise that corner where the prop broke since the command signal increased (I'm just trying to say the same thing in different ways here).
    This also suggests that the motor speed is calculated by the amount of voltage "And" current that the FC reads.
    BTW, I would have to completely agree with your analysis of what happened in this one.
    Note: I'm sure you already know all about the relationship between current and voltage but there may be some out there new to this so I done my best to explain it so others can learn.

    In the second example where the command signal first maxed out followed by a period where the command signal doesn't drop, you can see where the motor speed quickly steps down in that same time period, then when the motor speed increases for a split second, the command signal drops for a split second and is followed by a repeat of that same erratic behaviour. The erratic behaviour in the load or motor current suggests that at least one component in that ESC failed causing the erratic starting and stopping.
    At the end where the current (load) spiked upward is probably where the ESC completely failed but without more data past that point, that's just an half educated guess.
    Overheating could be the cause of what happened here and it could be something as simple as debris in the motor causing it to draw more current, overheat and burn an FET or MOSFET.
    There may not have been enough debris there to completely jam the motor but just enough to slow it down. The reason I say this is because of the ramp up in command signal, the erratic current and the fact that the speed didn't increase proportionally with the command signal (all just prior to the command signal maxing out).
    Believe it or not I had a small hard shell bug get sucked into one of the motors on my S800 and it caused damage to that ESC, it didn't burn it up or cause the motor to quit but they are much stronger motors than what the Inspires have. I did have to take the motor apart to clean it and replace a MOSFET on that ESC. The MOSFET got pretty hot but it was still working, I darn sure wanted to replace it so it wouldn't completely fail later. Under a 20X boom type stereo microscope it was easy to see that it got too hot by seeing the bubbled appearance on the case.
    This is the microscope I use which is great for a lot of things, not just electronics: Microscope
    I would have to completely agree with your analysis of what happened in this one as well.

    The third example is another story altogether to me.
    I feel there was a major fault but not a complete failure in the FC or the IMU or both in this case.
    I need to do some research to make an educated guess on this one.
    The theory of an ESC failure doesn't make sense to me since all four motors did it at the same time. The reason I say this is because there is an ESC for each motor so that would mean that all four ESCs would have to fail at the same exact time and do it twice at the same exact time so I respectfully disagree with an ESC be at fault in this one.
    I also feel that this may be one of those rare cases where the data may be corrupt and thereby giving false readings in the software made for interpreting what happened in this event.
    Keep in mind that I do have a theory about what happened but I'd like to research it some before writing my theory. I'd rather it be somewhat of an educated guess instead of just a wild guess. ;)

    I'll get back to you with my theory on what happened to the third one as soon as I can. I think and I repeat "I Think" I can shed some light on this one. Lol
    I'm noticing one thing right away that explains one of the dilemmas but I could be wrong so I'd rather do some research first to be more certain.

    My first post about this was my concern for good data versus corrupt data and the affects it could have on interpreting it correctly but this has brought up some great material for discussion and learning not only for me, but maybe you as well as others.
    In the case of what brought this subject up in the first place, I will have to say that if the software shows an increase in the command signal, a decrease in speed (voltage?) and an increase in current but with fairly smooth lines, that could indicate a motor that's restricted. But with the same levels with erratic (jumpy) lines, that could indicate that there is something wrong with the ESC and possibly the motor too. Usually the motor is what causes the ESC to get too hot because of an increase in current draw (high current equals high heat)
    I have to admit that I haven't used this software to interpret events because I haven't had any events to analyze..........YET. Lol

    Rest assure, when an event does occur, I will use the software. :)

    Actually now that I think about it, I did have one event but it turned out to be that the IMU calibration was way out. A full calibration cured the problem and it was from using the dynamic method to balance the props which I no longer recommend doing since the Inspires can't be switched to a completely manual mode where the IMU doesn't have any influence over the speed of the props or the attitude of the frame.

    Thanks, Bud for your time and efforts in helping out with this!! Your examples provide valuable information :)

    Nice work, BTW!!! :)

    Joe
     
  20. BudWalker

    Joined:
    Mar 27, 2016
    Messages:
    28
    Likes Received:
    4
    Joe,
    I'm reading your response. I realized that I forgot to include this in my earlier post. You might find it interesting. It's a re-enactment of the experiment that was done to find the motorLoad and motorSpeed values in the .DAT.

    Brought to you courtesy of @Luap
     
    #20 BudWalker, May 29, 2016
    Last edited: May 29, 2016