New Mavic Vibrating Violently - SOLVED!

[JDPEagle]

I finally received my Mavic Pro on Monday, December 12th (after a two-and-a-half month wait). The drone flew flawlessly up until this afternoon.

When the Mavic took-off, it began to shake violently and even had trouble holding its hover position. Here is what we know:

1. The Mavic has never been crashed,
2. The propellors have been carefully inspected (and even replaced with a new set),
3. All motors and parts feel solid and tight,
4. The IMU and compass have been recalibrated multiple times, and
5. The Mavic flew perfectly yesterday evening (and a dozen times prior to that).

Anyone seen this before? Any thoughts short of returning the drone to DJI?

SOLVED

When I first received the Mavic, I was unaware of the new DJI Go app. My initial unboxing and firmware upgrade was using the DJI GO 3.1.1 app. Of course, there were no apparent issues with the operation or behavior of the aircraft using the older app (with more than 15 flights over 5 days). When I learned of and downloaded the DJI GO 4 app, the aircraft began the shaking as seen in the above video. Initially, thinking that it could have been a result of the app, I disconnected my mobile device, restarted the UAS, and the issue persisted without the mobile device attached. I also attempted the flight with the older DJI GO 3.1.1 app, and there was no resolution. I immediately discounted the mobile app as the culprit.

After MANY failed attempts and tests to resolve the issue, I downgraded the firmware (to 01.02.9), and re-updated the firmware using the DJI GO 4 app. PROBLEM SOLVED!

My take away/conclusion: Do NOT upgrade the Mavic firmware with anything other than the DJI GO 4 app. In addition, the 01.03 firmware update from 3.1.1 is somehow different, or incomplete, compared to the one initiated through 4.0. The 3.1.1 updated aircraft became unstable once the 4.0 app was connected (and disconnecting the device made no resolve the issue, nor did the use of the 3.1.1 app).

YMMV. Thank you for all of the insights and troubleshooting efforts, folks!

 

[halley]

It's definitely an imbalance on one rotor somewhere. My vote from the video is the right rear. You've said you have swapped blades, so check the motor for more or less turning resistance or small debris.

 

[BD0G]

Looks like front right to me. I think there is a way to upload your log file and examine the amount of current that each motor was pulling or the RPM's of each motor. Then you could determine if one motor had the props turning way faster or way slower than the others. You could also potentially determine if one motor is drawing excessive current while trying to compensate for internal resistance.

 

[Frits]

Are you sure you don't have mismatched blades (black on black and white on white)?

 

[vashon100]

If you look at 0:37 on the vid, the left rear blade looks to have a lot of play in it. Is it mounted well (maybe missing a tab) or rivet slop?

 

[erkme73]

You must have a very sensitive microphone on the camera - it's picking up the ultrasonic clicking from under the MP. The front right looks to be spinning quite a bit slower than the others on startup. That may just be an illusion, but it's the first place I'd look.

The extreme test video on YouTube shows the affect of cutting one prop nearly in half, and it doesn't shake as bad as yours. It almost sounds like it's hitting a rev-limit and constantly compensating. Do any of the motors feel tighter than the others?

 

[JDPEagle]

I have thoroughly checked all of the props (and even replaced them with brand new ones). All rivets and characteristics are the same amongst all (and no damage). As erkme73 said, I have also seen many videos with severely damaged props not having this kind of effect on the aircraft.

All motors spin with the same resistance, and there is no free play.

I have the DAT logs off of the Mavic, and I don't know if they indicate individual motor amperage or RPMs. If so, could someone point me as to how to view the log files?

 

[erkme73]

Calling my man-crush @BudWalker... If anyone can figure out the mysteries of your DAT, he's your guy.

 

[vashon100]

Have you tried running it w/o props on while supported by the body only (securely only 1" off the ground) to see if the spin of the motors alone give that vibration?

 

[JDPEagle]

Have you tried running it w/o props on while supported by the body only (securely only 1" off the ground) to see if the spin of the motors alone give that vibration?

I have. There is no perceivable vibration.

At this point, if I had to place a bet, I'd say that there is a connection issue with one of the motors (bad soldering joint or wire itself). This would cause a twitchiness in a motor, and could cause this type of vibration. This vibration would not be perceptible without the props attached, and this would, obviously mean a return to DJI.

 

[BudWalker]

I finally received my Mavic Pro on Monday, December 12th (after a two-and-a-half month wait). The drone flew flawlessly up until this afternoon.

When the Mavic took-off, it began to shake violently and even had trouble holding its hover position. Here is what we know:

1. The Mavic has never been crashed,
2. The propellors have been carefully inspected (and even replaced with a new set),
3. All motors and parts feel solid and tight,
4. The IMU and compass have been recalibrated multiple times, and
5. The Mavic flew perfectly yesterday evening (and a dozen times prior to that).

Anyone seen this before? Any thoughts short of returning the drone to DJI?

I have the DAT logs off of the Mavic, and I don't know if they indicate individual motor amperage or RPMs. If so, could someone point me as to how to view the log files?

Could you Dropbox the .DATs so that I could take a look? You can also look yourself by using CsvView. But, I'm particularly interested in looking. Currently.the motor speed data is available in CsvView. The accelerometer and gyro data is also available and may yield something useful. I haven't quite figured out the motor amperage data yet.

EDIT: Just now figured out motor load/current. It'll be in the next version. Please provide a Dropbox link to the .DAT so I can take a look.

 

[JDPEagle]

Could you Dropbox the .DATs so that I could take a look? You can also look yourself by using CsvView. But, I'm particularly interested in looking. Currently.the motor speed data is available in CsvView. The accelerometer and gyro data is also available and may yield something useful. I haven't quite figured out the motor amperage data yet.

EDIT: Just now figured out motor load/current. It'll be in the next version. Please provide a Dropbox link to the .DAT so I can take a look.

Thanks for taking a look, Bud! Here's the link: Dropbox - DJI_ASSISTANT_EXPORT_FILE_2016-12-16_20-09-38.DAT

The DAT file includes three short flights (<2 mins each) I anxiously look forward to what you uncover!

 

[Aeroworks]

Looks like a bent propeller shaft or something. Also, could just be an illusion but the front right prop looks like a drastically different rpm.

Does the vibration intensify with increased throttle/rpm?

If it were my mavic, I would put 1 prop on at a time and while holding it firmly (or in a vice) power it up and see which motor gives the vibration.

Sent from my SM-T713 using MavicPilots mobile app

 

[erkme73]

Thanks for taking a look, Bud! Here's the link: Dropbox - DJI_ASSISTANT_EXPORT_FILE_2016-12-16_20-09-38.DAT

The DAT file includes three short flights (<2 mins each) I anxiously look forward to what you uncover!

This is how BudWalker is developing a cult following... Next man-crush to commence shortly...

 

[BudWalker]

Thanks for taking a look, Bud! Here's the link: Dropbox - DJI_ASSISTANT_EXPORT_FILE_2016-12-16_20-09-38.DAT

The DAT file includes three short flights (<2 mins each) I anxiously look forward to what you uncover!

I looked at this .DAT. Although interesting, I can't give you a specific cause, e.g. this motor is defective. Maybe other members will have some ideas.

The 2nd flight which actually had 5 separate "flights" where the Mavic's motors were started. There was one "flight" with props followed by another "flight" where the props were removed. Here are the motorSpeeds

Here are what I've termed motorLoads. I think it is a measure of current but I'm not sure how the data is obtained.

Without props the motorSpeeds increase but the motorLoads don't. The interesting part is the motorLoads with the props on.

For comparison I did a similar flight with my Mavic



The motorLoad variance with props is much large on your Mavic. I used Excel to look at the variance for each motorLoad hoping that would one would stand out indicating a specific motor. But, nothing could be seen there.

Just for fun take a look at the gyro data. It can be seen that it wobbles (that's a technical term) more about the X and Y axes than the Z axis just like in the video.

 

[JDPEagle]

I looked at this .DAT. Although interesting, I can't give you a specific cause, e.g. this motor is defective. Maybe other members will have some ideas.

The 2nd flight which actually had 5 separate "flights" where the Mavic's motors were started. There was one "flight" with props followed by another "flight" where the props were removed. Here are the motorSpeeds
View attachment 2610
Here are what I've termed motorLoads. I think it is a measure of current but I'm not sure how the data is obtained.
View attachment 2609
Without props the motorSpeeds increase but the motorLoads don't. The interesting part is the motorLoads with the props on.

For comparison I did a similar flight with my Mavic
View attachment 2611
View attachment 2612

The motorLoad variance with props is much large on your Mavic. I used Excel to look at the variance for each motorLoad hoping that would one would stand out indicating a specific motor. But, nothing could be seen there.

Just for fun take a look at the gyro data. It can be seen that it wobbles (that's a technical term) more about the X and Y axes than the Z axis just like in the video.
View attachment 2613

That is extremely interesting data (and quantitative visualization) of what we see in the video! Like you, I'm rather surprised to not see a more consistently oscillating value about one of the four motors. I am also surprised to see that all four motors seem to be operating at their full RPM range (and rather smoothly at that). However, there could still be a motor issue that is not being accurately reported back to the flight controller (perhaps indicating a stabilized RPM when in fact a dead spot in the motor is creating an oscillation).

An earlier post Aeroworks mentioned doing "run-ups" with each motor with a prop attached. I did this, and did not notice any vibration at all amongst all four motors.

The mystery continues, and Mavic #2 is on the way from Best Buy. And a HUGE "thank you" to Bud for this analysis@

 

[halley]

Is the indicated flight in the first strip chart from the same hop that we see in the video? I don't see a lot of maneuvering in the video. The variation in speed here is confusing to me. I would expect such a variation to occur on neighboring motors, not opposite motors, during a maneuver. For example, to pitch forward, the R Front and L Front would droop momentarily while the R Back and L Back would power up. The labels here show that both "white blades" on opposite corners dipped significantly and were compensated by the "black blades." The result would be level flight (I guess) but strange anyway.

 

[JDPEagle]

Is the indicated flight in the first strip chart from the same hop that we see in the video? I don't see a lot of maneuvering in the video. The variation in speed here is confusing to me. I would expect such a variation to occur on neighboring motors, not opposite motors, during a maneuver. For example, to pitch forward, the R Front and L Front would droop momentarily while the R Back and L Back would power up. The labels here show that both "white blades" on opposite corners dipped significantly and were compensated by the "black blades." The result would be level flight (I guess) but strange anyway.

The result of the increase/decrease RPMS of corresponding props would actually result in yaw. And that would have been from a different maneuvering flight from the one depicted in the video.

 

[halley]

The result of the increase/decrease RPMS of corresponding props would actually result in yaw.

Yeah I realized that about the same time you were typing a reply. The differential torque is employed for yaw/rudder action.

 

[BudWalker]

Is the indicated flight in the first strip chart from the same hop that we see in the video? I don't see a lot of maneuvering in the video. The variation in speed here is confusing to me. I would expect such a variation to occur on neighboring motors, not opposite motors, during a maneuver. For example, to pitch forward, the R Front and L Front would droop momentarily while the R Back and L Back would power up. The labels here show that both "white blades" on opposite corners dipped significantly and were compensated by the "black blades." The result would be level flight (I guess) but strange anyway.

This is actually two "flights". The first with props. The second (where you circled) is without props. So there wasn't any maneuvering. Since there the AC isn't moving and reacting to the motor speed there is no feedback to adjust motor speeds. If this had been a real flight with props the slower LBack and RFront would have both caused the AC to spin CCW.