Crashed doing Waypoint in Litchi, battery drained but never got warning

[FantomFlier]

I would have thought the Lit Ap would have given several warnings but that's a huge assumption because I have yet to give it a try.

The Litchi App would indeed have given several warnings, verbally as well. Make sure the volume on your device is not muted or turned down when flying. The pretty lady inside the RC talks to you all the time!

 

[Digdat0]

Lemme amend my request for the .DAT. Sounds like you've pretty much figured out what happened. The .DAT probably won't add much and it's a hassle to retrieve and post. But, if you're curious I'll be glad to look at it.

Here's the DAT file: Dropbox - FLY083.DAT
And Litchi Log: Dropbox - P3PCrashLog_2016-04-21_09-08-21_v2.csv

 

[Digdat0]

The Litchi App would indeed have given several warnings, verbally as well. Make sure the volume on your device is not muted or turned down when flying. The pretty lady inside the RC talks to you all the time!

Litch did give me updates, it kept telling me my battery was at 86%. It got stuck on that, and kept reporting 86% until the battery juice was gone and it ended up landing/crashed. Because it never read below 86%, it thought everything was fine. Here's the Litchi voice settings:

When I look in the DAT file logs, it shows 4.2+ volts per cell, up to the point where the power shut off. Maybe this wasn't Litchi, but DJI itself? I dunno. Thanks for the ideas man

It becomes difficult to analyze in absence of logs.

just posted

 

[flyNfrank]

Didn't quite read everyone's response so I'm not sure if it was brought up or not but, early last fall pilots were starting to report similar battery crash issues to your here. After a lot of heck was raised on the majority of the websites including dji.com a revision took place and firmware upgrade was issued.

The problem took place once the battery reached the lower volt limit. I believe it's around 13.43 volts, or close to that number. I have no idea what firmware you had installed at the time of the crash, but if the update was never made to the A/C and each Battery......what happened to you will take place to anyone else in the same situation.

I'm sorry you crashed and hope you can get back in the air soon.

 

[BudWalker]

Here's the DAT file: Dropbox - FLY083.DAT
And Litchi Log: Dropbox - P3PCrashLog_2016-04-21_09-08-21_v2.csv

That's not the right .DAT, there's no motor start .This flight lasted 165 seconds. The FLYXXX.DAT files have a timeDate if that helps

 

[SanCap]

Digidat0 said, "I'm not sure if I did something wrong, or if the app did, or both. Litchi stopped properly reading the battery properly at 2m 47.9s into the flight. It kept telling me that i had 86% left and was at 15.276V. I should have caught this as a red flag and landed and swapped batteries, but i didn't."

To answer your question both the app failed as well as the pilot. The app stopped reporting proper voltage early in the flight, it is software and software always has glitches. The pilot continued to fly the aircraft still hearing the same voltage from the 2.48 minute mark. At some time long before the aircraft ran out of battery the pilot should have landed to check and try to repair the improper voltage reporting. Obviously the pilot must have known that the battery could not still be at 86% after 20 minutes into the flight. I don't mean to be cruel but this crash was easily avoidable. Maybe others will learn from this. If you see something abnormal, LAND. Sorry for your loss.

 

[shockwave199]

You should visit flylitchi.com and join their forum and report this issue so they can at least have an opportunity to check this out further in case it's a bug in Android litchi.

 

[Digdat0]

That's not the right .DAT, there's no motor start .This flight lasted 165 seconds. The FLYXXX.DAT files have a timeDate if that helps

<facepalm> .. okay, so i got the right one, you were correct the file size is huge. my internet keeps timing out on the 300+mb file, so i used the DataCon tool to convert to a 100mb CSV, that i was able to upload to dropbox. Dropbox - CrashLog.csv

 

[BudWalker]

<facepalm> .. okay, so i got the right one, you were correct the file size is huge. my internet keeps timing out on the 300+mb file, so i used the DataCon tool to convert to a 100mb CSV, that i was able to upload to dropbox. Dropbox - CrashLog.csv

I'm still looking at the end of the flight to see what can be determined. But, the most interesting thing is that it looks like there was some kind of battery failure. Anything to do with the battery stopped changing at about time 180. Here's an example

Cell voltages, current, useful time, etc all stopped changing at the same time. If I had the .DAT I could see if the values were frozen or they stopped being reported. Clearly a hardware issue. IMHO you have a good warranty claim.

I'll look at the rest of the flight later. It appears that the P3 motors were going right up to the end.

 

[Digdat0]

Cell voltages, current, useful time, etc all stopped changing at the same time. If I had the .DAT I could see if the values were frozen or they stopped being reported. Clearly a hardware issue. IMHO you have a good warranty claim

Got the dat file to upload: Dropbox - FLY082.DAT

appreciate the review, i'm curious what ya see

 

[Noodle]

That 21 minute flight time really stand out. It's just too long to risk flying a P3 (even for P3 Standard). Try to keep the mission flight time to 15 minutes.

 

[FantomFlier]

IMHO you have a good warranty claim.
.

Here is a quote from the original poster's first post :

I should have caught this as a red flag and landed and swapped batteries, but i didn't. As a result, the app wasn't properly telling me how much battery or time left, and i wasn't paying attention and it ran out. About the 21 min.....

I do not understand why DJI should entertain a warranty claim? (I doubt that any 3rd party software suppliers give any warranties).

 

[ACC]

I believe that third party's, in this case Litchi, unfortunately get blamed for failures that may not be their issue at all. Third party software does not replace DJIs internal software but rather 'hooks' into it through DJI predefined points. In other words, Litchi commanded and interpreted what DJI provides through their SDK hooks. Because of Litchi's long history, I am not likely to blame it for the problem, though it is not impossible.

It will be interesting to see what other nuggets are garnered from the telemetry data. It almost looks like the battery computer stopped communicating with the drone computer and let it know it was reaching the end of its charge life... Possibly a condition that DJI has not programmed for.


Sent from my iPad using PhantomPilots mobile app

 

[BudWalker]

As I mentioned in my previous post the battery data became frozen or unavailable at about time 180 secs. This is a hardware problem on the P3 itself. The Litchi app was not at fault as it was just reporting the telemetry data given to it.

I looked at the end of this flight and it looks normal in many respects. I'm speculating that the P3 didn't know it was running out of battery right up to the end when it quit.

The P3 went from GPS_ATTI to GO_HOME at time 1045 secs and then went to AUTO_LANDING at time 1285 secs.

It's unclear why it switched to AUTO_LANDING since the battery data was saying the battery was good. There must be other triggers besides battery data that will cause the P3 to switch to AUTO_LANDING.

At the end of the flight motor speeds and their response to control stick inputs seems normal. Starting at time 1294 secs the throttle went to the max and stayed there for the remainder of the log. The throttle had been at max and went through a couple of iterations of neutral-max-neutral.

The response of the motors was normal. E.g. the right front motor looked like

The actual speed and commanded speed are both normal. The drop in speed is coincident with the throttle being reduced to neutral.

There were no unusual acceleration events, i.e.the P3 didn't hit anything (like a fence or the ground)

A stationary P3 will have an acceleration of 1.0 (due to gravity). The excursions in the plot are due the changes in vertical velocity which, in turn, due to the stick inputs.

The recording just stopped at 1306.5 secs when the P3 had an altitude of 233 meters. Seems like it ran out of juice and never saw it coming.

 

[BudWalker]

Here is a quote from the original poster's first post :




I do not understand why DJI should entertain a warranty claim? (I doubt that any 3rd party software suppliers give any warranties).

I should've explained a little better. I'm not saying what DJI should or shouldn't do. I'm saying what I think they will do. When I analyze a flight I often get to know DJI's response to warranty claims. Usually, but not quite always, DJI will repair a P3 under warranty if the damage had it's root cause in some hardware malfunction even if the pilot could have taken steps to mitigate the damage.

 

[FantomFlier]

I should've explained a little better. I'm not saying what DJI should or shouldn't do. I'm saying what I think they will do. When I analyze a flight I often get to know DJI's response to warranty claims. Usually, but not quite always, DJI will repair a P3 under warranty if the damage had it's root cause in some hardware malfunction even if the pilot could have taken steps to mitigate the damage.

Fair enough.

 

[Matt T]

In regards to litchi's capabilities, which the OP questioned to some degree, I can confirm that what happened is far from normal behavior, and likely is entirely a result of litchi receiving faulty battery data. Usually litchi will monitor not just the battery level, but also how much battery it would take to return to home. If it hits that threshold where the amount of battery required to get home is how much is left, it will cancel the mission and come home. I had this happen to me once. I was shooting some agricultural properties, and the one I was currently on only required about a 7-8 minute flight (as I remember it). Headwind on flight out, tailwind back, so ideal conditions. Flat land, no obstacles, LOS control for the full flight. I use litchi for running smooth autopilot missions for video, and already knew the flight time, etc. I was running 3 batteries - 1 in the craft, 1 on the charger, and 1 cooling down. As I had a long day of shooting, I would maintain this cycle, charging in my truck as I traveled between properties. When I was ready to start this one, the battery in my craft was in the 60's percent wise. The next one was almost done charging. I considered the flight time and battery, and decided to go for it, since it wouldn't be a problem for me to set it down in the field and go retrieve it if it didn't make it. It got about 3/4 of the way through the mission, when it hit that threshold and litchi automatically canceled the mission, turned around, and headed straight home. Just as it made it back overhead, it hit 25% and started beeping the normal warning, and then I brought it down to me, switched out with the now fully charged battery, launched again and completed the mission. Based on it arriving back at almost exactly 25%, it seems it is designed to bring the craft back before it gets below that point, accounting for both the trip there and the necessary flight back. It could possibly be that is true for 15% (what I have critical battery warning at) instead of 25%, and just got back earlier because it had a healthy tailwind on the return trip. Either way, I do know litchi has this functionality, and that it works (when all is working correctly between the craft, FW, etc).

 

[MindFever]

Gosh, you guys are brave. Longest flight I've had is 12 minutes and I'm watching the battery indicators and fly time remaining like a hawk.

Sent from my SHIELD Tablet K1 using Tapatalk

Sissy

Seriously though jokes aside ,why would you be affraid ? The battery just doesn't fail not even if it goes under 10% (from the posts around this forum people think an occasional dip below 10% will wreck the battery which I find not true.Long term doing this will affect overall battery life.And I mean doing it a lot.)
You have all the telemetry right there...there's nothing to worry about.
Seriously 12 minutes is being overly conservative.I fly it for 21 minutes if i don't push the motors too much.But mostly it's around 19 or 18 minutes.
Just relax and stretch it to 15 minutes at first.Then go longer...
Ok that sounded like some medicine advice for longer sex lol

 

[MindFever]

Digidat0 said, "I'm not sure if I did something wrong, or if the app did, or both. Litchi stopped properly reading the battery properly at 2m 47.9s into the flight. It kept telling me that i had 86% left and was at 15.276V. I should have caught this as a red flag and landed and swapped batteries, but i didn't."

To answer your question both the app failed as well as the pilot. The app stopped reporting proper voltage early in the flight, it is software and software always has glitches. The pilot continued to fly the aircraft still hearing the same voltage from the 2.48 minute mark. At some time long before the aircraft ran out of battery the pilot should have landed to check and try to repair the improper voltage reporting. Obviously the pilot must have known that the battery could not still be at 86% after 20 minutes into the flight. I don't mean to be cruel but this crash was easily avoidable. Maybe others will learn from this. If you see something abnormal, LAND. Sorry for your loss.

I don't think the poster expected the battery to be at 89 % after 20 minutes. That's just a silly statement.He expected the battery warning to go off which it didn't.
What do you mean software has glitches always...that's a pretty grimm assessment.If a software glitch causes a crash I can't blame the pilot for that.The only blame was he didn't turn back after seeing that battery level got stuck.

 

[movius]

Didn't quite read everyone's response so I'm not sure if it was brought up or not but, early last fall pilots were starting to report similar battery crash issues to your here. After a lot of heck was raised on the majority of the websites including dji.com a revision took place and firmware upgrade was issued.

The problem took place once the battery reached the lower volt limit. I believe it's around 13.43 volts, or close to that number. I have no idea what firmware you had installed at the time of the crash, but if the update was never made to the A/C and each Battery......what happened to you will take place to anyone else in the same situation.

I'm sorry you crashed and hope you can get back in the air soon.

A good reason to update firmware when available.

Sent from my SHIELD Tablet K1 using Tapatalk