Battery Discharge Results Using Modified Andriod App

[Liger 1956]

Using the FF6 Android app that @AnafiPilot1 has created I did a simple hover test using a fully charged battery and recorded the FF6 screen on my mobile phone. I then watched the video and noted down the % remaining battery and the battery voltage that FF6 reported every 30 seconds.

I used FF6 version 6.6.2.
The Anafi is on Firmware Version 1.6.3
The battery is on Firmware Version 1.0.9.0
The Maximum Battery Capacity is reported as 99%
The battery was fully charged about 1 hour before the flight.
The outside temperature was 8 Deg C. The battery was just warm to the touch at the end of the flight.

A graph of the battery voltage and % capacity remaining is below



Points of Interest (at least to me)
- It was notable that whilst the reported battery voltage would fluctuate up and down slightly, probably not helped by the wind, the reported % remaining battery always went down.
- The minimum voltage that the battery went down to was 7099mV which is about 3.5v/cell.
- After the Anafi landed the battery voltage recovered to 7222mV. I recharged the battery to about 7.4volts about 15 minutes after the flight.
- At a flight time of 20:14, remaining battery of 10% and a voltage of 7115mV FF6 stated "Very Low Battery - 3 minutes before auto land". I then landed the Anafi myself.

If other forum members could do similar tests with the most up to date firmware and with batteries that are showing a less than perfect maximum capacity then we would have more data to look at.

 

[Gipsz Jakab]

Using the FF6 Android app that @AnafiPilot1 has created I did a simple hover test using a fully charged battery and recorded the FF6 screen on my mobile phone. I then watched the video and noted down the % remaining battery and the battery voltage that FF6 reported every 30 seconds.

I used FF6 version 6.6.2.
The Anafi is on Firmware Version 1.6.3
The battery is on Firmware Version 1.0.9.0
The Maximum Battery Capacity is reported as 99%
The battery was fully charged about 1 hour before the flight.
The outside temperature was 8 Deg C. The battery was just warm to the touch at the end of the flight.

A graph of the battery voltage and % capacity remaining is below

View attachment 5114

Points of Interest (at least to me)
- It was notable that whilst the reported battery voltage would fluctuate up and down slightly, probably not helped by the wind, the reported % remaining battery always went down.
- The minimum voltage that the battery went down to was 7099mV which is about 3.5v/cell.
- After the Anafi landed the battery voltage recovered to 7222mV. I recharged the battery to about 7.4volts about 15 minutes after the flight.
- At a flight time of 20:14, remaining battery of 10% and a voltage of 7115mV FF6 stated "Very Low Battery - 3 minutes before auto land". I then landed the Anafi myself.

If other forum members could do similar tests with the most up to date firmware and with batteries that are showing a less than perfect maximum capacity then we would have more data to look at.

I also made a spreadsheet from my previous flight. I flew from 100% to the storage voltage.
I used FF6 version 6.6.2.
The Anafi is on Firmware Version 1.6.3
The battery is on Firmware Version 1.0.9.0
The Maximum Battery Capacity is reported as 97%
The battery was fully charged about 1/2 hour before the flight.
The outside temperature was 6 Deg C.

The lesson of the table: the% display of the battery charge does not seem to reflect the change in voltage, it is too linear. I think it has more to do with the elapsed time than the battery voltage.
This may be due to the fact that in the case of a weak, worn-out battery, it looks the same according to the elapsed time and then the battery will run out sooner and fall out of the sky by more than 20-30%.

 

[Liger 1956]

@Gipsz Jakab Your graph appears to indicate that after 12 minutes of flight you still have a battery voltage of about 7.6v but I only have just over 7.2v in my battery even though the health of both batteries is very similar. My flight was simply pressing the take off button and then letting it hover without touching the sticks. Could you please weigh your Anafi, with battery? Parrot state a weight of 320g however as my Anafi is wearing a vinyl skin and some small 3D printed joint protectors on the legs it weighs in at 331g (3.4% heavier than Parrot state) on my kitchen scales. This could be a cause for the difference in voltages but we would need more data to be sure although my battery appears to have a larger initial drop in voltage during the first 2 minutes of the flight than yours.

 

[AnafiPilot1]

During my testing - I do understand why Parrot does not display voltage.

I have a known weak battery - and it still works great, but I noticed while flying in race mode, the voltage does sag quite a bit - at 60% battery level pushing full speed moving forward and up, voltage was dropping from 7.5v to 7.1v, then when I slow down the voltage recovers to 7.5v. Battery level did not change at all (which is as I would expect).

When I get a chance I will try and log this battery from 100% to landing while flying.

 

[Gipsz Jakab]

Using the FF6 Android app that @AnafiPilot1 has created I did a simple hover test using a fully charged battery and recorded the FF6 screen on my mobile phone. I then watched the video and noted down the % remaining battery and the battery voltage that FF6 reported every 30 seconds.

I used FF6 version 6.6.2.
The Anafi is on Firmware Version 1.6.3
The battery is on Firmware Version 1.0.9.0
The Maximum Battery Capacity is reported as 99%
The battery was fully charged about 1 hour before the flight.
The outside temperature was 8 Deg C. The battery was just warm to the touch at the end of the flight.

A graph of the battery voltage and % capacity remaining is below

View attachment 5114

Points of Interest (at least to me)
- It was notable that whilst the reported battery voltage would fluctuate up and down slightly, probably not helped by the wind, the reported % remaining battery always went down.
- The minimum voltage that the battery went down to was 7099mV which is about 3.5v/cell.
- After the Anafi landed the battery voltage recovered to 7222mV. I recharged the battery to about 7.4volts about 15 minutes after the flight.
- At a flight time of 20:14, remaining battery of 10% and a voltage of 7115mV FF6 stated "Very Low Battery - 3 minutes before auto land". I then landed the Anafi myself.

If other forum members could do similar tests with the most up to date firmware and with batteries that are showing a less than perfect maximum capacity then we would have more data to look at.

Analyzing the video made by Solaris8x86, I made a graph of the bad battery.
One row of data came every 3 seconds, so I used data from every 10 rows as a basis. I added the battery voltage of the battery by adding the two cells to 3 decimal places (I did not use the rounded value on the left):

 

[AnafiPilot1]

People we have got some real good data - managed to retrieve the black box telemetry from an autolanding event ( Thanks to @ROBANAFI for providing the files ).

It is a fascinating and helps explain a lot - I will let the graph do the talking. note: the battery level line should have continued at 71% level till flight end, I missed a data point!

 

[Liger 1956]

That is very interesting data. I remember @ROBANAFI posting about his auto land in a wood below his take off point.

Have you managed to get the battery voltage from the .bin file in the FDR directory or another source? From the data once the auto land has been initiated then it appears that even if the battery recovers somehow it does not get cancelled.

Do you have any idea why the battery voltage recovered to 7500mV during the auto land? The slope of the altitude graph appears to indicate that the motors were still running and that it was a "controlled " descent. Also I would have thought that a RTF would have been triggered before the auto land unless the RTF trigger us based on the remaining battery % and not voltage.

 

[Bud FPV]

Great review ! The sag is normal ,there the same thing in FPV racer ,and I guess the lipo HV from parrot is not make for that ,may be less sag with normal Lipo of FPV racers . The HV lipo has a price (less C ,discharge courant) ,they are no miracle .

 

[Gipsz Jakab]

Great review ! The sag is normal ,there the same thing in FPV racer ,and I guess the lipo HV from parrot is not make for that ,may be less sag with normal Lipo of FPV racers . The HV lipo has a price (less C ,discharge courant) ,they are no miracle .

Although different from LiPo, this is abnormal.

 

[Liger 1956]

@AnafiPilot1 Do you, or @ROBANAFI know the Battery Health % as reported by FF6 for the flight?

 

[AnafiPilot1]

If we go back to the Anafi in the Ohio River thread, as I said and ultimately proved correct (per statement by Parrot to solaris8x86 ). Parrot have introduced some sort of voltage threshold which triggers an critical battery alarm, causing the Autoland.

During landing as the load on the battery has reduced the voltage has been able to recover, but Parrot have chosen to make the Critical Auto Land a one way thing - once triggered it will execute the landing regardless of changing situation (again as I suspected in the Ohio River thread).

Clearly @ROBANAFI 's battery in this case is faulty or worn, maybe one cell has a very high internal resistance, and under load the voltage is sagging ( Current Drawn x Internal Resistance = Voltage Drop ). Judging by the wild voltage fluctuations with varying load - it does point to a cell with a very high IR.

His data shows that Parrot have the Critical Battery Alarm set around ~6600mV. and clearly Parrot did this to address Anafi falling out of the sky problems when the battery still has plenty of 'charge' left!

Notice how the battery level is totally disconnected from the actual battery voltage - during Autoland it stayed at 71% at which time the voltage went from 6600mV to 7500mV !!

 

[Bud FPV]

Sorry ,I may be didn't undestand , the charts to 5.01Pm come from rivers incident ? Sorry for the question !

 

[Bud FPV]

If we go back to the Anafi in the Ohio River thread, as I said and ultimately proved correct (per statement by Parrot to solaris8x86 ). Parrot have introduced some sort of voltage threshold which triggers an critical battery alarm, causing the Autoland.

During landing as the load on the battery has reduced the voltage has been able to recover, but Parrot have chosen to make the Critical Auto Land a one way thing - once triggered it will execute the landing regardless of changing situation (again as I suspected in the Ohio River thread).

Clearly @ROBANAFI 's battery in this case is faulty or worn, maybe one cell has a very high internal resistance, and under load the voltage is sagging ( Current Drawn x Internal Resistance = Voltage Drop ). Judging by the wild voltage fluctuations with varying load - it does point to a cell with a very high IR.

His data shows that Parrot have the Critical Battery Alarm set around ~6600mV. and clearly Parrot did this to address Anafi falling out of the sky problems when the battery still has plenty of 'charge' left!

Notice how the battery level is totally disconnected from the actual battery voltage - during Autoland it stayed at 71% at which time the voltage went from 6600mV to 7500mV !!

Problem of tempo on trig ?

 

[Gipsz Jakab]

Sorry ,I may be don't undestand , the charts to 5.01Pm come from rivers incident ? Sorry for the question !

from this video, first case:

The voltage values start from a low point and then suddenly drop from 32% to a state of exhaustion. The shape of the curve is nearly normal, the only problem is why BMS assigns out-of-place percentage values to these values?

 

[ROBANAFI]

@AnafiPilot1 Do you, or @ROBANAFI know the Battery Health % as reported by FF6 for the flight?

@AnafiPilot1 Thanks for doing this

@Liger 1956 i've just checked and the faulty battery only has 65% capacity! I've not been using it since the auto landing but luckily my other battery is a fairly healthy 96%. It was a really windy day and with the condition of the battery it might have pushed it to far?

 

[AnafiPilot1]

@AnafiPilot1 Thanks for doing this

@Liger 1956 i've just checked and the faulty battery only has 65% capacity! I've not been using it since the auto landing but luckily my other battery is a fairly healthy 96%. It was a really windy day and with the condition of the battery it might have pushed it to far?

Yes most likely the power to maintain flight in wind is quite high, would explain the voltage drop.

The Blackbox telemetry is quite a complex json format, 3 levels of data streams and on change logging with timestamp (very different from the standard flight log json file).

I just extracted the info I needed, but it would be nice if we could parse the whole file, to extract location / pitch / yaw / azimuth / xyz speed / etc - so we could get a better picture of the load the drone was on at a particular voltage.

Maybe @h-elsner could add another parser for the Blackbox telemetry json to his excellent Anafi Log project!!

 

[ROBANAFI]

Yes most likely the power to maintain flight in wind is quite high, would explain the voltage drop.

The Blackbox telemetry is quite a complex json format, 3 levels of data streams and on change logging with timestamp (very different from the standard flight log json file).

I just extracted the info I needed, but it would be nice if we could parse the whole file, to extract location / pitch / yaw / azimuth / xyz speed / etc - so we could get a better picture of the load the drone was on at a particular voltage.

Maybe @h-elsner could add another parser for the Blackbox telemetry json to his excellent Anafi Log project!!

Also its worth noting that the log file you asked for was already in FF6, it wasn't from the SD card. So if anybody does happen to lose the Anafi and can't get it back there will be a log file in the app.

 

[Liger 1956]

I did not know that there was a Blackbox json format file recorded as well as the standard one that we know about and use in the ShowAnafiLog software. Can you let me know the name of the file, and it's location, so that I can look for one on my phone. It may be easier to parse it after a test flight than watch a screen recording and manually note down the figures.

 

[ROBANAFI]

I did not know that there was a Blackbox json format file recorded as well as the standard one that we know about and use in the ShowAnafiLog software. Can you let me know the name of the file, and it's location, so that I can look for one on my phone. It may be easier to parse it after a test flight than watch a screen recording and manually note down the figures.

/ Android / data / com.parrot.freeflight6 / files / logs

I have no idea how it was decoded though.

 

[Gipsz Jakab]

Analyzing the video made by Solaris8x86, I made a graph of the bad battery.
One row of data came every 3 seconds, so I used data from every 10 rows as a basis. I added the battery voltage of the battery by adding the two cells to 3 decimal places (I did not use the rounded value on the left):
View attachment 5120

I plotted the battery level according to the voltages on the graph !!! This is what FF6 should have shown: