Mavic 2 Pro motors shut down mid-flight

[tbrim]

I experienced a crash of my Mavic 2 Pro yesterday evening; the motors just shut down mid-flight with no warning. I had launched the drone from the base of a hill, ascended to 400 ft, and then flew towards the hill to capture some photos of a structure. I had first lined up the structure on the screen in the Go4 app and then flew the drone towards it. As I reached the point where I wanted to capture photos from I stopped the drone. I confirmed that I still had visual line of sight on the drone, looked back down to my phone screen, and then started descending the drone to better frame the shot. As the drone began descending I heard the motor noise quit, the display in the Go4 app went black, and then a "Weak Signal." message was displayed within the app.

I switched to the Find My Drone section of Go4 and used that to guide me for a hike up the hill to the last known location of the drone. After reaching the point where the blue dot and red triangle were overlapping I started searching for the drone. It took me about 10 minutes to locate it, and another 5 minutes to locate the battery as it was no longer seated in the drone. The battery case had separated and the top of the case had a big dent and several scratches. The electronics in the battery still worked as the green lights around the power button still worked when I pressed it. The individual cells of the battery had separated but appeared undamaged. The drone itself had suffered a break to the foot on the left front arm and the prop on that arm was cracked and chipped. There were several scratches around the body of the drone.

When I got back home I downloaded the flight logs from the DGI Go4 app to my computer. I decrypted the flight log using the log viewer on PhantomHelp as well as AirdataUAV. The log looked normal in every regard. I created a KMZ file of the flight and loaded it in Google Earth. Moving the mouse to the launch point indicated ground elevation of 850 feet. Moving the mouse pointer to the last datapoint from the flight log indicated ground elevation of 951 feet. The last altitude that showed on display and recorded in the flight log was 348 ft - so I was greater than 240 above ground level at the time, well above any obstruction.

This morning I put a fresh battery in the drone and powered it up. The gimbal didn't go through its normal initialization and failed to move at all - so something to do with gimbal control is damaged. I connected the drone to my computer and started up DJI Assistant. It recognized the drone but wasn't able to locate any flight record data on it.

I have DJI Refresh coverage on the drone so filed a ticket with them and will be sending it to them for repair. I mentioned in the details that the battery was severely damaged. I'm surprised that they offered no ability to attach photos of damage to the battery or drone or to upload the flight record. Perhaps they'll be able to extract some data from the drone that I was not able to.

I'm really at a loss why the drone just shut down. Everything in the flight log appeared normal up until the time that things just quit. I'm attaching the flight log in case anyone would like to take a look.

 

[4wd]

It probably simply lost power - bad connection to battery, a few similar cases on here.
When the battery wasn't in after the crash, there's a suspicion it wasn't clipped in 100% - but obviously it could easily eject in the impact.
We've all nearly done it, always double check it's fully seated and clipped in.
Clips and contacts are subject to wear and tear anyway. Be gentle.

 

[tbrim]

It probably simply lost power - bad connection to battery, a few similar cases on here.
When the battery wasn't in after the crash, there's a suspicion it wasn't clipped in 100% - but obviously it could easily eject in the impact.
We've all nearly done it, always double check it's fully seated and clipped in.
Clips and contacts are subject to wear and tear anyway. Be gentle.

That's as good an explanation as any regarding what happened. I always push the battery in place until it clicks and then continue pressing to make sure it's firmly seated before powering it up. Who knows what could happen upon impact after falling from 250 feet. I did find the Mavic in the dirt but there were lots of rocks nearby. It could have it a rock and bounced and then landed where I found it.

 

[runeOfRune]

Post the .DAT log from the app, it has event log, there may be something in it, otherwise it's just battery disconnecting.

 

[tbrim]

Post the .DAT log from the app, it has event log, there may be something in it, otherwise it's just battery disconnecting.

DAT file added to original post.

 

[supernicko]

the mid air power cut could be battery popping out from not 100% seating - sometimes it seems clicked - it can look completely seated in - it sure does power up and run but the descent causes battery to lift up and comeout.... i wouldnt write it off as an answer - i think the m3 will come with a better battery clip system and one that wont let it power on unless seated correctly...

IF your log shows a descent or sharpish stop after a vertical climb then complete power loss no more data - i would look at battery popping out!

 

[sar104]

The DAT event stream log has a number of entries prior to failure that indicate a problem with the battery - specifically that the battery was not properly connected:

94.113 : {task_e}INFO:i2c3 core recovery||
94.113 : 6625 [L-BATTERY][ERROR] CMD_BLACK_BOX read error||
94.124 : {task_e}INFO:i2c3 core recovery||
94.124 : 6625 [L-BATTERY][ERROR] smbus_read CMD_PRESENT_STATE error||
95.101 : {task_e}INFO:i2c3 core recovery||
95.101 : 6674 [L-BATTERY][ERROR] smbus_read CMD_POWER_STATE error||
95.229 : 6681 [L-FMU/LED]action changed. Normal Flash:near vert limit(1)
95.229 : 6681 [L-FMU/LED]type:0, normal flash action:
95.229 : 6681 [L-FMU/LED]c0:0,15;c1:0,3;c2:0,13;c3:2,3;c4:0,10;c5:0,3;c6:0,10;c7:0,3
96.099 : {task_e}INFO:i2c3 core recovery||
96.099 : 6724 [L-BATTERY][ERROR] smbus_read CMD_POWER_STATE error||

The situation was triggered by the sudden pitch excursion around 93 seconds when the aircraft pitched forwards again after heavy braking:




Most likely a case of an improperly latched battery.

 

[Paul2660]

As much as I hate to do it, I now am taping all my M2 batteries before flight. Should not be necessary, but here again is case where someone heard the click, checked the battery and it still came out. Might have been due to a hard flight movement but still it just should not happen. I have 2 M2's and the 2nd one has a slight gap where battery meets fuselage of drone. I always check to make sure also, but adding gaffers tape to each side just makes sure. Just no excuse for this. There is slight side to side play also, unless I add the tape.

The P4 Vr 2.0 solution is really no better IMO. As I figured over time the sensor that has to be engaged will not work, i.e. depress enough and you will get a battery error, even though battery is in there solid. Same solution for that is a small amount of tape to keep the sensor down. I have 5 total P4 batteries, and now none will depress the sensor without the extra small wade of tape. You cannot launch until that sensor is depressed.

To me a more positive latching solution needs to be developed, should not be that hard for any future DJI drone.

Went back to the MPP I have and for sure the battery locks down harder and more solidly than the M2's. No play, no gaps.

Paul C

 

[sar104]

As much as I hate to do it, I now am taping all my M2 batteries before flight. Should not be necessary, but here again is case where someone heard the click, checked the battery and it still came out. Might have been due to a hard flight movement but still it just should not happen. I have 2 M2's and the 2nd one has a slight gap where battery meets fuselage of drone. I always check to make sure also, but adding gaffers tape to each side just makes sure. Just no excuse for this. There is slight side to side play also, unless I add the tape.

The P4 Vr 2.0 solution is really no better IMO. As I figured over time the sensor that has to be engaged will not work, i.e. depress enough and you will get a battery error, even though battery is in there solid. Same solution for that is a small amount of tape to keep the sensor down. I have 5 total P4 batteries, and now none will depress the sensor without the extra small wade of tape. You cannot launch until that sensor is depressed.

To me a more positive latching solution needs to be developed, should not be that hard for any future DJI drone.

Went back to the MPP I have and for sure the battery locks down harder and more solidly than the M2's. No play, no gaps.

Paul C

I think on pretty much every one of these cases I've examined the pilot asserted without doubt that the battery was properly installed, both latches clicked, etc, etc.. Memory overwritten by hindsight.

 

[AMann]

I velcro cinch strap mine in.

Photo later- bought the strap at True Value.

 

[Thomas B]

It is worrisome, but I have not had the issue. I take time to hear a “double” click when I install a battery.

 

[tbrim]

The DAT event stream log has a number of entries prior to failure that indicate a problem with the battery - specifically that the battery was not properly connected:

94.113 : {task_e}INFO:i2c3 core recovery||
94.113 : 6625 [L-BATTERY][ERROR] CMD_BLACK_BOX read error||
94.124 : {task_e}INFO:i2c3 core recovery||
94.124 : 6625 [L-BATTERY][ERROR] smbus_read CMD_PRESENT_STATE error||
95.101 : {task_e}INFO:i2c3 core recovery||
95.101 : 6674 [L-BATTERY][ERROR] smbus_read CMD_POWER_STATE error||
95.229 : 6681 [L-FMU/LED]action changed. Normal Flash:near vert limit(1)
95.229 : 6681 [L-FMU/LED]type:0, normal flash action:
95.229 : 6681 [L-FMU/LED]c0:0,15;c1:0,3;c2:0,13;c3:2,3;c4:0,10;c5:0,3;c6:0,10;c7:0,3
96.099 : {task_e}INFO:i2c3 core recovery||
96.099 : 6724 [L-BATTERY][ERROR] smbus_read CMD_POWER_STATE error||

The situation was triggered by the sudden pitch excursion around 93 seconds when the aircraft pitched forwards again after heavy braking:

View attachment 75072


Most likely a case of an improperly latched battery.

Thank you very much for the analysis. That makes me feel better to know what the issue was - but also makes me feel bad that it was my fault. Between the original Mavic and the new Mavic 2 I have over 650 flights without a single incident. What surprises me is that this was the second flight on that battery. I had completed capturing imagery from one location and then packed up and drove to a second launch site to capture images from a different point of view. Perhaps when either placing the Mavic into, or removing it from, the carry bag I inadvertently loosened the battery.

 

[sar104]

Thank you very much for the analysis. That makes me feel better to know what the issue was - but also makes me feel bad that it was my fault. Between the original Mavic and the new Mavic 2 I have over 650 flights without a single incident. What surprises me is that this was the second flight on that battery. I had completed capturing imagery from one location and then packed up and drove to a second launch site to capture images from a different point of view. Perhaps when either placing the Mavic into, or removing it from, the carry bag I inadvertently loosened the battery.

It's hard to say what might have caused it to be loose. I've seen enough of these now that I'm ridiculously careful in checking the battery before each flight. Now that I've said that I'll probably have the same thing happen.

 

[runeOfRune]

See, I don't believe it was a battery issue. The presence of this:

{task_e}INFO:i2c3 core recovery

...is a common denominator in threads where people had their drone disconnect and become lost. I don't believe people suddenly have problems installing batteries. It's not THAT hard. This may be a whole different issue.

 

[paulatkin73]

It's hard to say what might have caused it to be loose. I've seen enough of these now that I'm ridiculously careful in checking the battery before each flight. Now that I've said that I'll probably have the same thing happen.

somebody posted a video here showing how battery moves a bit in the bay after latching, stating it should not do that.
all mine - all 5 batteries i have - all do that same thing - they click on both sides, but there is definitely some amount of 'play' after they click in place. i did not have any issues so far, but, it is what it is. makes you think twice before flaying above water of any other place where you would not be able to reach it back. and, there is not much space to place a velcro strap there, it does not fit well between downward looking sensors - may cause more issues than help, if it accidentally shifts and blocks any of those.

 

[paulatkin73]

See, I don't believe it was a battery issue. The presence of this:

...is a common denominator in threads where people had their drone disconnect and become lost. I don't believe people suddenly have problems installing batteries. It's not THAT hard. This may be a whole different issue.

i2c is the communication bus to the battery, log message states the link is producing errors. why it producing errors - typically it would be a hardware connectivity issue, but, can be an uart chip overheat, but, well, not very likely. i dunno - does this .dat file have any metric showing cpu core temp, core voltage and load level? i did not see it, but, i did not dig too deep neither.

 

[sar104]

See, I don't believe it was a battery issue. The presence of this:

...is a common denominator in threads where people had their drone disconnect and become lost. I don't believe people suddenly have problems installing batteries. It's not THAT hard. This may be a whole different issue.

I first posted about that error a few weeks ago, but it turns out to be directly linked to the battery errors. @msinger and @BudWalker made the connection. They always occur together and if the condition lasts more than a few seconds it is reflected in the SMART_BATTERY_status field. I'm pretty sure now that it is indicative of a bad battery connection.

Obviously the vast majority of flights do not have battery installation issues or these aircraft would be raining from the sky, but it only takes a moment of inattention to get it wrong.

 

[sar104]

somebody posted a video here showing how battery moves a bit in the bay after latching, stating it should not do that.
all mine - all 5 batteries i have - all do that same thing - they click on both sides, but there is definitely some amount of 'play' after they click in place. i did not have any issues so far, but, it is what it is. makes you think twice before flaying above water of any other place where you would not be able to reach it back. and, there is not much space to place a velcro strap there, it does not fit well between downward looking sensors - may cause more issues than help, if it accidentally shifts and blocks any of those.

I push it in firmly, both sides, and then give it a few tugs to see if it shifts. I don't think a little play is a problem as long as both latches are engaged.

 

[kilomikebravo]

I'm sure this "situation" is of interest to many M2 owners and like SAR, I do the same push and pull routine but my OCD is telling me to start using a Velcro strap. However, as Paul indicated, the real estate for such is quite limited. More pondering required.

 

[Paul2660]

I found that the gaffers tape works as well, it handles heat OK, leaves no residue when pulled off as long as you pull after each flight, which you have to if you are going to fly again. Unfortunate that needs to be done, but there is no recovery from a loss of battery connection. Based on the review of records from OP, it seems as if the battery had several near misses of loss of connection, which makes me wonder if the drone airframe may get hot in flight and flex with flight movements. My M2's do get hot after a couple of flights at least in the summer.

From what I can understand from various posts on this forum and DJI's site, the M2's seem to have two possible issues:

1. Crashes due to a FC failure or reboot, which I guess is fatal and no recovery from unless you are up high enough for the reboot to complete.
2. What appear to be battery disconnections which also have no recovery, and I guess also cause a FC failure. Since the battery connections drop down into individual sockets on the Mavics I can't see flex being an issue for connection interrupt but failure of the connection on
one side or the other sure could.


The battery on the P4 Pro when seat correctly, literally has no flex, you can't move it at all on my P4 Vr 2.0, either the Obsidian black or white batteries, not the case with the M2, especially my 2nd drone which I noticed from day 1 had play, and a bit of flex when battery was inserted.

Paul C