This afternoon I took my drone for a flight/range test, I took it just over 1000 meters and started the flight home at 60% making sure I had more than enough battery to make it home.

About 1/3 of the way home the battery suddenly dropped from 49% to 0% instantly, several warnings appeared on the screen including:

"Critically low power. Aircraft is landing."
"Propulsion output has been limited to ensure battery health."

At this point I tried to cancel the landing but that failed.

The height dropped from 56 meters to 17 meters in 25 seconds, I managed to recover the altitude to 29 meters before I lost connection completely and the Mavic Air self-landed in a field over 500 meters from my location. I got into my car and managed to locate it using my DJI Go app.

This of course is a major worry, I purchased the Mavic Air at an airport and the extra batteries from DJI via Amazon. I live near the coast and I have been flying over water quite a bit, if this happened over water I would no longer have my Mavic Air.

I cannot post the YouTube link in here, but if you copy and paste " 8Ml1enHPMsI " (without the quotes) into YouTube search you will find the video.

Obviously after this I have lost confidence in the Mavic Air, I would like to know why this happened and if it is safe for me to fly the Mavic Air?

Should I make a claim or send it back?

Any suggestions appreciated.

Here is your link.

It’s a real strange one, you can’t tell much from that short replay, except that there was some problem/fault with battery reading, I believe only dji can throw some light on this.
You never said if your aircraft was damaged or how bad.

You next move should be to contact dji file a case and take it from there.

hallmark007 Posted at 2018-3-28 15:17
It’s a real strange one, you can’t tell much from that short replay, except that there was some problem/fault with battery reading, I believe only dji can throw some light on this.
You never said if your aircraft was damaged or how bad.

There was no damage to the aircraft, it landed safely and was recovered. My concern is it happening again, especially over water!

Featherlake Posted at 2018-3-28 15:40
There was no damage to the aircraft, it landed safely and was recovered. My concern is it happening again, especially over water!

Well at least that’s a good result, when you look at your problem, first thought is faulty battery, but if for some reason Aircraft has something to do with sudden drain on battery you need to get both checked out and an answer from dji as to how/why this happened would be useful. Send it back..

Man that is crazy at least no crash and you were able to recover the craft.  Hope you are able to get the issue resolved and am thinking most likely a bad battery.

hallmark007 Posted at 2018-3-28 15:52
Well at least that’s a good result, when you look at your problem, first thought is faulty battery, but if for some reason Aircraft has something to do with sudden drain on battery you need to get both checked out and an answer from dji as to how/why this happened would be useful. Send it back..

Glad you got'er back! I would find a large field and fly it again with another battery but keeping in in VLOS and try to replicate the fault. Fly any batteries that are fully charged (excluding the one you had an issue with) if there is no issue I would fully charge the battery that incurred the error message and fly it close VLOS for the entire flight down to 20% or so to see if it happens again. Just a process of elimination before you send it in. It may just be a battery error or could be a drain of the battery by the AC.

Someone else posted the exact same thing just a day or two ago.  I think their battery % was also on the 40's when it suddenly dropped to 0.

Any point in extracting Mavic Air's internal logs to see if there is more information there that might help?

QuadKid Posted at 2018-3-28 16:01
Glad you got'er back! I would find a large field and fly it again with another battery but keeping in in VLOS and try to replicate the fault. Fly any batteries that are fully charged (excluding the one you had an issue with) if there is no issue I would fully charge the battery that incurred the error message and fly it close VLOS for the entire flight down to 20% or so to see if it happens again. Just a process of elimination before you send it in. It may just be a battery error or could be a drain of the battery by the AC.

"I would fully charge the battery that incurred the error message and fly it close VLOS for the entire flight down to 20% or so to see if it happens again."

Anybody make a battery tester for DJI batteries?  Would be safer than test flights.  
Next best would be to strap drone down and spin it up as a first test.  No problems, move on to actual test flight.

In case of Mavic Pro batteries, the four middle connetors are 12.92v, two pairs to either side are ground, outter most connector on one edge is ?3.27v and ?9.2v, based on analysis post on youtube.  Supposedly four ground connectors are wired together, and four 12.92v connectors are wired together.  
Which should make it easy for those familir with basic electronics to make a battery load tester for Mavic Pro batteries.

I think you should send your drone to DJI for inspection. It look like there was not the problem about your battery capacity and something from drone drained your battery to 0%. If that was the case, your drone would directly crash rather than went into landing mode. I guess it is voltage sensor error (could be hardware or software) that made the software thought your battery was empty and went into landing mode. Without professional inspection from DJI, you never know when it will happen again.

Hi Featherlake, I'm sorry to hear about the accident. Please contact our support to get a case number and send it in for full diagnosis: https://www.dji.com/support, thanks!

Hi Featherlake, would be nice to see your flightlog, if you like use this link http://www.phantomhelp.com/LogViewer/Upload/ and post the uploaded link on this forum.
Love to see and analyze this one.
Cheers JJB

JJBspark Posted at 2018-3-28 21:50
Hi Featherlake, would be nice to see your flightlog, if you like use this link http://www.phantomhelp.com/LogViewer/Upload/ and post the uploaded link on this forum.
Love to see and analyze this one.
Cheers JJB

Here you go:

http://www.phantomhelp.com/LogViewer/3N7D0HFYYZ02Z6ZPKWOH/

Featherlake Posted at 2018-3-29 00:23
Here you go:

http://www.phantomhelp.com/LogViewer/3N7D0HFYYZ02Z6ZPKWOH/

Had a look at your file. #1 and 2 batt cell values shows values between 1 and 65 VDC, #3 cell doesn`t show values. The Total of 3 cells looks normal and the percentage shows 100% at start and drops to 49%. Usefull batt power value (in seconds) is much lower than it should be, my best guess is that your start at 100% batt is actually lower.
I suggest you check batt cell values in the GO app, mayby u see some differences in values between batterys.

Hope that DJI can explain what went wrong with your drone/batt/app.
Cheers JJB

JJBspark Posted at 2018-3-29 01:49
Had a look at your file. #1 and 2 batt cell values shows values between 1 and 65 VDC, #3 cell doesn`t show values. The Total of 3 cells looks normal and the percentage shows 100% at start and drops to 49%. Usefull batt power value (in seconds) is much lower than it should be, my best guess is that your start at 100% batt is actually lower.
I suggest you check batt cell values in the GO app, mayby u see some differences in values between batterys.

Where does the log show these values. I used the link provided and #3 cell values are there and nowhere do I see the 65VDC values for the other cells....Are we looking at the same thing?

Yeah, I dont see these values either.

Edit:I do know what went wrong I think, I just need to clarify something. Stay tuned!!!!!

Edit:Ahh! Lookking elsewhere, I do see some stupid spikes of Cell voltage. Hmmm!

Your battery looks healthy and ok. Why your Mavic Air suddenly calculates your remaining battery capacity to 0% in 7m 12.6s I can not say exactly. A "gifted" guess will be there is a software/hardware malfunction in your drone that makes it suddenly incorrectly calculating 0% of power on the battery.

I do not want advice on refreshing the firmware but leaves it to an DJI employe.  

Would personally send battery and drone return to seller with the message to get the stuff repaid on the warranty.

Regards Leif.

Simmo1 Posted at 2018-3-29 03:57
[view_image]

I do not think that CsvView is reporting the data correctly....A voltage spike of 1300+% (5V-65V) seems highly unlikely.

Featherlake Posted at 2018-3-28 15:40
There was no damage to the aircraft, it landed safely and was recovered. My concern is it happening again, especially over water!

After drone landed, did you look at battery level after turning drone off and back on?

What about powering down drone, removing battery, replacing battery, power drone backup, then looking at battery level?

ggeorgis Posted at 2018-3-29 22:57
I do not think that CsvView is reporting the data correctly....A voltage spike of 1300+% (5V-65V) seems highly unlikely.

The log data which is used by CsvView presents some strange values at random intervals.
Is it the aircrafts logging system which produces this? Or another app in the workflow?

Here a sample:

If is indeed the onboard stuff that is making the error, other onboard flight systems use this data to make decisions too?? Hopefully it is CSView reading the .txt wrong?

I've asked Mr Walker to perhaps look in and comment...

An electrical short of the magnitude required to drop that battery from 50% to zero in a few seconds would of fried the aircraft. You would also smell it. This looks like a battery problem.  Remember the integrated systems on this aircraft trigger various functions from readings from a number of sensors. One of these is the battery information being passed to the CPU. "When condition x =  A then do y type logic". My guess, and it's purely a guess is that the battery was never at 100% despite it showing this when you started. The battery read value could have either corrected itself at around 48% and shown the true value close to 0% or the battery itself is faulty and fails at a certain point under load. These type of batteries are not supposed to have or develop a "load failure" but I am am sure you are going to see more and more battery issues as they near the end of their recharge cycle. Even with cell phones today as the battery recharge cycles climb you will notice that even at full charge the phones does not quite last as long.

For those that enjoy conspiracy theories. On exactly the same date as this incident my Samsung Galaxy S7 did the same thing. I climbed into my Jeep at 7:30 with my phone on 67%, I put the phone in my phone holder and noticed it has just switched off. When I restarted the battery was at 0%. It has never happened before and is fine again now. Go figure. - solar flare, radiation from a secret test or Batman. Was just strange.  

Kevjones1959 Posted at 2018-3-30 02:11
An electrical short of the magnitude required to drop that battery from 50% to zero in a few seconds would of fried the aircraft. You would also smell it. This looks like a battery problem.  Remember the integrated systems on this aircraft trigger various functions from readings from a number of sensors. One of these is the battery information being passed to the CPU. "When condition x =  A then do y type logic". My guess, and it's purely a guess is that the battery was never at 100% despite it showing this when you started. The battery read value could have either corrected itself at around 48% and shown the true value close to 0% or the battery itself is faulty and fails at a certain point under load. These type of batteries are not supposed to have or develop a "load failure" but I am am sure you are going to see more and more battery issues as they near the end of their recharge cycle. Even with cell phones today as the battery recharge cycles climb you will notice that even at full charge the phones does not quite last as long.

For those that enjoy conspiracy theories. On exactly the same date as this incident my Samsung Galaxy S7 did the same thing. I climbed into my Jeep at 7:30 with my phone on 67%, I put the phone in my phone holder and noticed it has just switched off. When I restarted the battery was at 0%. It has never happened before and is fine again now. Go figure. - solar flare, radiation from a secret test or Batman. Was just strange.

I think this is the most plausible explanation.

I recharged the same battery overnight and flew again, it functioned perfectly and the aircraft has functioned normally ever since the incident.

It's still concerning that a partially change battery can display 100% at take off.

Featherlake Posted at 2018-3-30 02:38
I think this is the most plausible explanation.

I recharged the same battery overnight and flew again, it functioned perfectly and the aircraft has functioned normally ever since the incident.

Yes I agree. We are currently relying on the green LED indicators on the battery and the app. I dont know what more they could do actually.

Keule Posted at 2018-3-30 00:49
The log data which is used by CsvView presents some strange values at random intervals.
Is it the aircrafts logging system which produces this? Or another app in the workflow?

The phone I use for flying (Galaxy S7) is used for flying only, I have a few apps and had recently installed Facebook because I wanted to look at sharing live video on Facebook at some point.

During the flight I turned on my WiFi because I wanted to update the satellite map to see exactly where I was, when I did this Facebook started bombarding me with dozens of notifications at once (remember this isn't my main day to day phone). My phone was buzzing away for at least a minute before all the notifications had stopped, I'm not sure if this had any affect on things?

Featherlake Posted at 2018-3-30 02:45
The phone I use for flying (Galaxy S7) is used for flying only, I have a few apps and had recently installed Facebook because I wanted to look at sharing live video on Facebook at some point.

During the flight I turned on my WiFi because I wanted to update the satellite map to see exactly where I was, when I did this Facebook started bombarding me with dozens of notifications at once (remember this isn't my main day to day phone). My phone was buzzing away for at least a minute before all the notifications had stopped, I'm not sure if this had any affect on things?

Do me a favor, I am one of these guys who dont believe in co-incidences. Ask on your facebook page if anyone experienced any sudden loss of battery power on a cell phone or battery powered piece of electronic equipment on Wednesday or strange behaviour with their phones. I am going to do the same.

Simmo1 Posted at 2018-3-29 03:57
[view_image]

jesus christ...for the layman and non-technical, what does this mean?

Here's something to ponder and understand.

I had an Inspire 1.......the reason I no longer have it was due to a catastrophic battery failure...it puffed to the point of disconnecting from the mount.

But prior I had read that if you put in a partially charged/used battery that the electronics would assume 100% charge.......so I tested this.   I went to the local RC field, put in a battery that had 60% charge from a flight the previous day.    Took off,  flew out 300ft at about 20ft high cruzed around and then after 2.5 min I got a critical low battery warning!!!  and the AC landed right there....walked out about 200ft in the field to get her........That was a KNOWN issue, and I believe was solved with the following UpDate.......but it happened.

Randy

Acenspades Posted at 2018-3-30 11:44
jesus christ...for the layman and non-technical, what does this mean?

Its the graph of what was going on in the aircraft in respect to the 3 cells, throttle, altitude and the smart battery drop. Easy to see?? But why, that is the question.

Simmo1 Posted at 2018-3-30 16:10
Its the graph of what was going on in the aircraft in respect to the 3 cells, throttle, altitude and the smart battery drop. Easy to see?? But why, that is the question.

Graph (created by CsView?) does not appear to go along with the data in DJIFlightRecord_2018-03-28_[18-00-19].csv

The battery pack appears to be fine.  It looks like battery information got corrupted by Drone's processing.  The voltages are same just before battery indicates 0% and voltages go up afterwards, yet battery is still indicating 0%.

Does anyone know the format (records / fields) or breakdown of DJIFlightRecord_2018-03-28_[18-00-19].txt file?

The reply from Mr BudWalker:  'I think this is a problem with the .txt to .csv converter that CsvView uses. '
He also suggested perhaps taking a look at the .DAT file.

The .txt to csv converter that CsvView uses has been fixed. Thanks to ferraript, the TXTlogToCSVtool developer.

https://phantompilots.com/threads/tool-win-offline-txt-flightrecord-to-csv-converter.70428/page-28#post-1347355

The next version of CsvView will have the updated version of TXTlogToCSVtool. If you want to use the new TXTlogToCSVtool before then
1) download the attached .zip and unzip
2) copy the TXTlogToCSVtool.exe from that .zip to the <CsvView Install Dir>/executables directory. Usually this will be the C:/Program Files (x86)/CsvView/executables directory

An update on my post:

I have been using the MA for a few days following the incident with no issues whatsoever. The only thing I can think of is that the battery that failed at 49% was not fully charged before the flight.

So make sure that your batteries are fully charged before you set off. For some reason the software indicated that it was a full charge when it wasn't, hopefully this is something that DJI will look into and fix.

Featherlake Posted at 2018-4-1 14:09
An update on my post:

I have been using the MA for a few days following the incident with no issues whatsoever. The only thing I can think of is that the battery that failed at 49% was not fully charged before the flight.

I think the battery WAS fully charged actually. I think you are putting a fully charged battery in the MA and it is actually ignoring the battery state for some reason or screwing it up. Once the battery reaches around 50% it is then subtracting the current reading from the current reading. In other words we start at 100% then we use 48% instead of subtracting the 100-48 to give remaining percentage it is subtracting the used from remaining causing the craft to believe the battery has suddenly gone to 0. and initiating all the RTH / etc. Else someone has produced a heap of bad batteries.  I suspect the software update on MA with the battery update. Something does not sound right. IN some cases once the pilot has recovered the craft and checks the battery a "dead" battery is showing 22% available. Also if in fact the power of the battery really went to 0 namely nothing. I dont believe the software could even initiate a RTH and the motors would just switch off and it would fall out of the sky with zero power. In other words it would just switch off and die.

I'm 99% certain that I'm way off base with this theory - but - I get my TV signal from a transmitter close to where you were flying.
I don't suppose there's any way your MA got fried by a burst of microwaves??