Do not trust at a glance but read the numbers

[Mauri59]

It's my habit to be cautious, but I didn't pay enough attention to all the numbers.
Before takeoff I always do a checklist, which until a few days ago I thought was accurate and sufficient. As luck would have it, it's my habit to have the drone approach when the app indicates the battery is below 40%.
By the time the Mavic Pro was 100m away from me and 50m up, the battery was down to 30% and automatic RTH started. Since it was close and at low altitude, I interrupted RTH to take a couple of pictures. Suddenly, just 14 seconds later, the remaining battery power had plummeted to less than 10% and the drone began the landing procedure, without me being able to do anything.
My checklist also included checking the status of the batteries, but until yesterday, my eye was on the voltages to check that they were correct and look for any imbalances, while for the charge level I trusted the full battery image.
What I hadn't noticed, was the reduced mA capacity (only 2/3) at "full" battery. Battery that had charged without problems up to the 4th LED and then turned off, that has only 47 flights and that if I try to check the remaining life time (holding the button for 6 seconds) says that it is still between 90% and 100%.

 

[Meta4]

Since it was close and at low altitude, I interrupted RTH to take a couple of pictures. Suddenly, just 14 seconds later, the remaining battery power had plummeted to less than 10% and the drone began the landing procedure, without me being able to do anything

It sounds like the battery wasn't properly charged at the start.
Your flight data will have a lot more information to show what was going on.
Go to DJI Flight Log Viewer | Phantom Help
Follow the instructions there to upload your flight record from your phone or tablet.
That will give you a detailed report on the flight data.
Come back and post a link to the report it provides and someone might be able to analyse it and give you an understanding of the cause of the incident.

 

[Mauri59]

Thank You for the answer.

I would gladly look at the log, but Win10 20H2-19042.1110 crashes (with blue screen) when I try to install the Ipad Mini2 driver.
That said, it's pretty obvious that the flight range estimation system takes voltage as a reference, and that when the 2150mA is almost depleted, the voltage suddenly plummets and the estimated remaining range as well, sending the drone into a forced landing.
With other battery packs there are no problems, while recharging the "incriminated" pack the situation is repeated, with a charge that does not go beyond 2150mA (instead of 3400-3800mA), but 4.2V each cell.
What I don't like is the misleading information that the led system of the battery gives me. The pack seems to charge correctly (one led flashing, then two, then three, finally four, then they go off) and even, by holding the button for 6 seconds, the battery is given in excellent condition.

 

[Mauri59]

It sounds like the battery wasn't properly charged at the start.
Your flight data will have a lot more information to show what was going on.
Go to DJI Flight Log Viewer | Phantom Help
Follow the instructions there to upload your flight record from your phone or tablet.
That will give you a detailed report on the flight data.
Come back and post a link to the report it provides and someone might be able to analyse it and give you an understanding of the cause of the incident.

...finally I was able to access the log, which seems to confirm the sudden drop in autonomy with an inconsistency between the autonomy data shown and the actual ones, which led to the assumption of having a higher autonomy, and suddenly, through some other parameter (or protection), led instead the drone to land on the spot.

The strange thing in the log is the indication of a critical value in the battery when it was still at 16% charge (guess apparent), with the concomitant activation of autolanding (which there was no way to stop), while the subsequent indication of Forced Landing occurred at 5% battery.
Two inconsistent data, because the critical threshold was set at 10% and not 16%, and it was at 10% that Forced Landing should have been activated, not 5%.

12m 0s P-GPS 16satellites 316.5ft N/A 0mph 337.1ft 37% 10.737V 3.58V 3.58V 3.58V 0V

12m 49.8s Auto Landing 16satellites 317.5ft N/A 0mph 336.4ft 16% 10.778V 3.59V 3.59V 3.59V 0V Obstacle Avoidance will be disabled in landing.; Critically Low Power. Aircraft Landing.

14m 13.5s Forced Landing 15satellites -4.9ft 6.9ft 0mph 337.5ft 5% 10.683V 3.56V 3.56V 3.56V 0V

 

[Meta4]

...finally I was able to access the log,

How about posting it so those familiar with reading the data can see what's in there?
It's going to have useful information, particularly your cell voltages at different times through the flight.
The % indication means nothing unless the battery was fully charged at the start of the flight

 

[Mauri59]

How about posting it so those familiar with reading the data can see what's in there?
It's going to have useful information, particularly your cell voltages at different times through the flight.
The % indication means nothing unless the battery was fully charged at the start of the flight

0m 0s

Starting Motors

14satellites

0ft

0.3ft

0mph

5.9ft

100%

12.561V

4.19V

4.19V

4.19V

0V

Data Recorder File Index is 359.; Taking Off.

0m 1.2s

Auto Takeoff

14satellites

0ft

0.3ft

0mph

5.9ft

100%

12.345V

4.12V

4.12V

4.12V

0V

Taking Off.

0m 3s

Auto Takeoff

14satellites

1.3ft

1.6ft

2.2mph

0.6ft

100%

11.943V

3.98V

3.98V

3.98V

0V

Home Point Updated.; Home Point Recorded. RTH Altitude: 100m.

There is no other useful information in the log, because the log started from 100% charge and with normal voltage values for a "full" Li-Po cell.
The problem, as you can see from the battery status image I uploaded, is the inconsistency of showing a 100% full charge, when the capacity of a nominal 3830mA battery pack is instead only 2150mA.
The drone should not have initiated an automatic landing at 16% (critical value was set to 10%) nor a forced landing at 5%. The only possible inference is that some other type of protection may have been involved, but this is not indicated in the log.
I guess the "smart" battery is not that smart, if it indicates as completed the charge when one or more cells are actually defective, and I am very surprised that there is no routine that compares the actual capacity (which is indicated, those 2150mA) with the nominal one (3830mA). How can 2150mA and 100% coexist is the question? (probably, because the charge level is extimated only from the voltage).
It was enough that I had checked the capacity value as well, before taking off, to realize that inconsistency, but I acted like the controller and GO4 app, that is, I trusted the 100% charge indication.
(and it won't happen again)

 

[Ralph thompson]

If u’d been on this forum for long enough u’d have seen a number of posts from people very knowledgable in LIPO technology & there advice is don’t go by the percentage, it assumes that you take care of your batteries and maintain them. it’s likely that if u discharge the battery until it automatically switches off, let it cool, then fully recharge it, it will return to normal, assuming no bulges or cracks.

 

[Mauri59]

If u’d been on this forum for long enough u’d have seen a number of posts from people very knowledgable in LIPO technology & there advice is don’t go by the percentage, it assumes that you take care of your batteries and maintain them. it’s likely that if u discharge the battery until it automatically switches off, let it cool, then fully recharge it, it will return to normal, assuming no bulges or cracks.

Hi Ralph,
actually, I've read all about lithium polymer batteries, and all my battery packs have always been treated according to standard, almost maniacally: never left 100% charged beyond three days (discharged 50% manually), charged at least once every three months if not in use, stored at temperatures between 18° and 28° Celsius, never used in cold environments. The pack that gave the problem (one of four) is not even the most used one (only 47 flights).
In any case, I was wrong not to pay attention to the limited capacity of the battery shown by the GO4 app already before take-off, but I am surprised that the software, while detecting such an anomalous data (I did not notice, but the 2150mA was on the display), associates it to 100% charge, without detecting the inconsistency and without displaying a warning, as if the voltage was the only parameter of reference.
All is well, apart from the expense of a new pack: the drone landed without a single scratch and was easily recovered, the faulty battery was set aside and replaced with a new one. (but I really don't think I'll ever, in person, fly an aircraft powered by a LiPo battery).

 

[Interficiam Stercore]

It's my habit to be cautious, but I didn't pay enough attention to all the numbers.
Before takeoff I always do a checklist, which until a few days ago I thought was accurate and sufficient. As luck would have it, it's my habit to have the drone approach when the app indicates the battery is below 40%.
By the time the Mavic Pro was 100m away from me and 50m up, the battery was down to 30% and automatic RTH started. Since it was close and at low altitude, I interrupted RTH to take a couple of pictures. Suddenly, just 14 seconds later, the remaining battery power had plummeted to less than 10% and the drone began the landing procedure, without me being able to do anything.
My checklist also included checking the status of the batteries, but until yesterday, my eye was on the voltages to check that they were correct and look for any imbalances, while for the charge level I trusted the full battery image.
What I hadn't noticed, was the reduced mA capacity (only 2/3) at "full" battery. Battery that had charged without problems up to the 4th LED and then turned off, that has only 55 flights and that if I try to check the remaining life time (holding the button for 6 seconds) says that it is still between 90% and 100%.View attachment 132397View attachment 132398

I'll start heading home at 50%, unless I know I will have a tail wind pushing me home faster than no wind or a headwind. If I have a 3-5 mph tailwind to help me go home, I will risk the 40% low batt level. I always check to see if I have a tail wind in the direction of my home point by looking at the WINDY app and or also doing a short reverse flight toward home. If it goes faster coming home than it does going out, I know there will be a tail wind helping me reach my home point.

 

[AVLJMA]

If u’d been on this forum for long enough u’d have seen a number of posts from people very knowledgable in LIPO technology & there advice is don’t go by the percentage, it assumes that you take care of your batteries and maintain them. it’s likely that if u discharge the battery until it automatically switches off, let it cool, then fully recharge it, it will return to normal, assuming no bulges or cracks.

Maintaining the batteries is key. I have a personal log and use Airdata. I learned from flying manned aircraft that triple checking everything is key to survival. It is not different with drones.

 

[Ralph thompson]

I would raise a ticket with DJI, they having an amazing support service, they’d want u to return it & turn-around might be a couple months but they take quality seriously. At worst, they’re costs are reasonable but I wouldn’t be surprised if they just replaced it at no cost, even way out of the warranty period. Good luck.

 

[Mauri59]

I'll start heading home at 50%, unless I know I will have a tail wind pushing me home faster than no wind or a headwind. If I have a 3-5 mph tailwind to help me go home, I will risk the 40% low batt level. I always check to see if I have a tail wind in the direction of my home point by looking at the WINDY app and or also doing a short reverse flight toward home. If it goes faster coming home than it does going out, I know there will be a tail wind helping me reach my home point.

In this case, the problem was that the full charge of the battery pack was actually only two-thirds of its rated value (i.e. 2150mA instead of 3830mA).
The 2150mA is equivalent to 56% of nominal capacity, so I took off with a battery with a capacity of just over 50%, even though graphically the cells were full, and the remaining range given at 100%.
What amazes me, is that despite the fact that the overall capacity indicated (2150mA) is reduced by 44%, there is no warning from the app, a bit as if DJI had provided for the use of (smaller) batteries other than those for sale.
Basically, when the voltage dropped to 30%, yes 30% remained in the battery, but it was 30% of 2150mA and not 3830mA, which is just 16.8%.

 

[Mauri59]

I would raise a ticket with DJI, they having an amazing support service, they’d want u to return it & turn-around might be a couple months but they take quality seriously. At worst, they’re costs are reasonable but I wouldn’t be surprised if they just replaced it at no cost, even way out of the warranty period. Good luck.

The "failed" battery pack doesn't have many flights, but is too old for warranty work. The solution was to buy another one and learn from experience that the mA capacity matters much more than the estimated percentage remaining. This % estimate is valid only if the battery pack has a capacity close to the nominal one. Considering that a cell might fail in flight, from now on I will tap my finger on the screen every now and then at the battery symbol and check the remaining mA.

 

[Mauri59]

Maintaining the batteries is key. I have a personal log and use Airdata. I learned from flying manned aircraft that triple checking everything is key to survival. It is not different with drones.

What happened to me, looks a bit like some airplane crashes that occurred due to incorrect indication of fuel level compared to the actual amount in the tanks. If the pilots had calculated how much fuel had been loaded and used, they would have noticed that the onboard gauge was indicating the wrong value.

Fuel System Maintenance Error: Tuniter ATR72 TS-LBB Ditching 6 August 2005 - Aerossurance

An error in a stores database entry was the first of a series of errors that collectively resulted in both engines flaming out and a ditching in the Med.

aerossurance.com

 

[MAvic_South_Oz]

the battery was down to 30% and automatic RTH started

To be clear, the Low Battery Warning shown set on yours at 30% is not when the RTH kicks in.
That is only a warning the pilot can set in Go4 for his own use / benefit.
So too the Critical Low Battery Warning, you can set this for 10%, 15%, whatever you like.

RTH will kick in when the aircraft on board computer figures it's time, based on distance from home and estimated RTH battery needs.
This is usually pretty conservative, I've usually noticed landings using RTH (rare for me) of around 15% to 20% battery remaining, though RTH doesn't usually take into account wind, so a pilot really needs to be aware of this and make sure they use discretion if conditions require it.

Post your Phantom link to the flight record, it really is the only way the finer detail of the data can be seen by those wanting to help you understand the forced landing incident.

 

[Mauri59]

To be clear, the Low Battery Warning shown set on yours at 30% is not when the RTH kicks in.
That is only a warning the pilot can set in Go4 for his own use / benefit.
So too the Critical Low Battery Warning, you can set this for 10%, 15%, whatever you like.

RTH will kick in when the aircraft on board computer figures it's time, based on distance from home and estimated RTH battery needs.
This is usually pretty conservative, I've usually noticed landings using RTH (rare for me) of around 15% to 20% battery remaining, though RTH doesn't usually take into account wind, so a pilot really needs to be aware of this and make sure they use discretion if conditions require it.

Post your Phantom link to the flight record, it really is the only way the finer detail of the data can be seen by those wanting to help you understand the forced landing incident.

In the flight log, the posted lines are the only ones with meaningful indications.
As far as I'm concerned, I think I know how my drone flies and what it does (after 5 years of use), however I discovered (I'm an electronic engineer) that the GO4 app's indication of remaining range only makes an estimate from the voltage, WITHOUT considering the actual battery capacity, although it does indicate it.
If the initial capacity is substantially reduced (and my "fault" was not paying attention to this fundamental data) there is something wrong with the battery pack, even if the cells are represented full and their voltages are normal and leveled.
In a case like this, the remaining range is wrong (we start from a 56% represented as 100%) and the behavior, i.e. the discharge curve, will be faster and especially unpredictable, accelerated.
The drone was set as in the picture, yet it started the descent from a different value than the preset one, 16% instead of 10% and after 14 seconds from the 30% warning, it was already impossible to stop it.
The purpose of my post was to point out the importance of looking at the total capacity (mA) of the battery pack, before and during flight, because the residual percentage figure, when a pack is faulty (but not reported), is a bogus and misleading indication.