Mavic 2 wants a compass calibration 30 days after last one

[BudWalker]

Your are right when I drag the file into the dropbox folder is shows the size as 836 bytes yet the flight recorders show the file size as 11.9 mb. I must be doing something wrong. Any ideas.

Try directing iTunes to do the file xfer.

 

[bcoles]

Try directing iTunes to do the file xfer.

Here is the new link copied as you suggested using iTunes to save it in Dropbox. Hope this works. Dropbox - 2018-10-05_13-27-25_FLY037.DAT

 

[bcoles]

Here is the new link copied as you suggested using iTunes to save it in Dropbox. Hope this works. Dropbox - 2018-10-05_13-27-25_FLY037.DAT

I have a tee time at 12:30 so not be around for 5 hours. I will be back. Thanks for everyones help. Rejoin later, Bryan.

 

[Fplvert]

As @sar104 observed (below) the calibration was requested because "magnetic interference", not the 30+ day rule.

Has the M2P requested another calibration after the one you did for this flight?

It would be interesting to see the .DAT file immediately before you started getting the calibration request. Since the .DAT files are relatively small you could zip up several.

Did you add a tracker or some other external device? The calibration made a significant change. Here is the magMod before and after the calibration. It changes from about 1230 before the calibration to about 1500 after.
View attachment 49048

Thanks for looking, I was quite surprised when the compass asked to be calibrated. I’ve started motors in the same spot many times.

I haven’t had any more compass calibration requests.

My prior 5 or 6 flights have been Waypoint Litchi missions in various locations. I might try to retrieve those just for curiosity sake.

No tracker on this one, it has been rock solid.

 

[gnirtS]

Ive jusr moved mine again and got prompted to calibrate again. Bits of log:-

27.655 : 1620 [L-COMPASS][mag_cali_pt] distance_from_last 965291.6 m|
28.526 : 1671 [L-COMPASS][mag_cali_pt] distance_from_last 965291.6 m|
28.612 : 1676 [L-FMU/LED]action changed. compass out of range:need_cali(1)
29.396 : 1722 [L-COMPASS][mag_cali_pt] distance_from_last 965291.7 m|
30.266 : 1773 [L-COMPASS][mag_cali_pt] distance_from_last 965291.6 m|
31.137 : 1824 [L-COMPASS][mag_cali_pt] distance_from_last 965291.6 m|
32.007 : 1875 [L-COMPASS][mag_cali_pt] distance_from_last 965291.6 m|
..
64.212 : 3762 [L-COMPASS][mag_cali_pt] distance_from_last 965293.3 m|
65.082 : 3813 [L-COMPASS][mag_cali_pt] distance_from_last 965292.6 m|
65.953 : 3864 [L-COMPASS][mag_cali_pt] distance_from_last 965292.2 m|
66.226 : 3880 [L-COMPASS][scale cali(0)] fill num:[194]
66.226 : 3880 [L-COMPASS][scale cali(0)] estimation error:[15.7]
66.226 : 3880 [L-COMPASS][scale cali(0)] succeed! bias:80.3 -360.0 -1024.7 scal:3.590 3.572 3.835|
66.226 : 3880 [L-COMPASS][save data] app cali all success
66.226 : 3880 [L-COMPASS][save data] in user index mode
66.226 : 3880 [L-COMPASS]mag cali pos and time saved success!
66.226 : 3880 [L-COMPASS][mag_cali_pt]lat:0.030011, lon:1.927658
66.226 : 3880 [L-COMPASS][mag_cali_pt]height:5.1, date:20320216
66.226 : 3880 [L-COMPASS]req gimbal recover when cali end

So move it 600 miles, demands calibration. Further flights in this area gave no calibration prompts.

Incidentally, i dont know what those lat/lon fields are - its nowhere near a real location in any units you can think of.

 

[gnirtS]

What if the 30 day rule had something to do with the eventLog stream date of 2032 on @gnirtS but @sar104 and @BudWalker updated to 2018 correct date and time? Where would the eventLog stream be getting its data from? Is it possible that @gnirtS might need calibration less often since his has been calibrated for the next 14 years according to the eventLog stream file? His date of 2032 is 14 years into the future, so any calibrations before that date are not being recorded. Maybe the day after 20320216, 20320217, @gnirtS will show the updated calibration.

If DJI does not come up with a cure, maybe there is a way to fast forward the date on whatever chip is keeping track of the date, so an owner wont be prompted for continual calibrations.

A lot of hardware devices do not have an offline, backed-up real time clock (they don't need one). It could be the case here the drone hasn't got one so relies on a pre-programmed arbitrary time (2032 etc) until it gets a GPS lock and can accurately set its GPS time.

I'd expect routines like distance and 30 day since calibration to be run after GPS time has been set so i cant see it changing the frequency required.

 

[gnirtS]

That is QUITE a holiday, amigo.

To illustrate:-


Local wildlife is fun too:-

 

[BudWalker]

Ive jusr moved mine again and got prompted to calibrate again. Bits of log:-

27.655 : 1620 [L-COMPASS][mag_cali_pt] distance_from_last 965291.6 m|
28.526 : 1671 [L-COMPASS][mag_cali_pt] distance_from_last 965291.6 m|
28.612 : 1676 [L-FMU/LED]action changed. compass out of range:need_cali(1)
29.396 : 1722 [L-COMPASS][mag_cali_pt] distance_from_last 965291.7 m|
30.266 : 1773 [L-COMPASS][mag_cali_pt] distance_from_last 965291.6 m|
31.137 : 1824 [L-COMPASS][mag_cali_pt] distance_from_last 965291.6 m|
32.007 : 1875 [L-COMPASS][mag_cali_pt] distance_from_last 965291.6 m|
..
64.212 : 3762 [L-COMPASS][mag_cali_pt] distance_from_last 965293.3 m|
65.082 : 3813 [L-COMPASS][mag_cali_pt] distance_from_last 965292.6 m|
65.953 : 3864 [L-COMPASS][mag_cali_pt] distance_from_last 965292.2 m|
66.226 : 3880 [L-COMPASS][scale cali(0)] fill num:[194]
66.226 : 3880 [L-COMPASS][scale cali(0)] estimation error:[15.7]
66.226 : 3880 [L-COMPASS][scale cali(0)] succeed! bias:80.3 -360.0 -1024.7 scal:3.590 3.572 3.835|
66.226 : 3880 [L-COMPASS][save data] app cali all success
66.226 : 3880 [L-COMPASS][save data] in user index mode
66.226 : 3880 [L-COMPASS]mag cali pos and time saved success!
66.226 : 3880 [L-COMPASS][mag_cali_pt]lat:0.030011, lon:1.927658
66.226 : 3880 [L-COMPASS][mag_cali_pt]height:5.1, date:20320216
66.226 : 3880 [L-COMPASS]req gimbal recover when cali end

So move it 600 miles, demands calibration. Further flights in this area gave no calibration prompts.

Incidentally, i dont know what those lat/lon fields are - its nowhere near a real location in any units you can think of.

The lat, long values are expressed in radians.

 

[Former Member]

A lot of hardware devices do not have an offline, backed-up real time clock (they don't need one). It could be the case here the drone hasn't got one so relies on a pre-programmed arbitrary time (2032 etc) until it gets a GPS lock and can accurately set its GPS time.

I'd expect routines like distance and 30 day since calibration to be run after GPS time has been set so i cant see it changing the frequency required.

Just curious, can you name some modern electronic devices that depend on an internal clock that dont have an offline, backed up, real time clock? Even my $1.99 Walmart digital desk clock has a backup AA battery in it. Every computer I have owned since 1995 has a backup battery for its BIOS.

I would expect distance and position to be run after the GPS time has been set, but how much can you rely on GPS if your internal clock has a bug? In your case, your GPS time on your drone was not set, or it was set by a satellite somewhere in orbit that has a date set to the year 2032.

By the way, the GPS unit and compass are 2 separate mechanisms and processes. You can fly without GPS lock and your compass will still work and your DAT time will be correct. Atti mode does not rely on GPS lock, but the flight will be recorded in your DAT file.

 

[gnirtS]

Just curious, can you name some modern electronic devices that depend on an internal clock that dont have an offline, backed up, real time clock? Even my $1.99 Walmart digital desk clock has a backup AA battery in it. Every computer I have owned since 1995 has a backup battery for its BIOS.

The Raspberry Pi machines and derivatives to pluck one out of my head immediately. And a lot of real time I/O based boards.

 

[Former Member]

The Raspberry Pi machines and derivatives to pluck one out of my head immediately. And a lot of real time I/O based boards.

I think it's safe to say that either your Mavic 2's onboard hardware backup time calculator is malfunctioning or your firmware has a bug in it. Have you tried calibrating again since the last calibration where it gave you the year 2032? Updated your Mavic recently to the latest firmware?

 

[gnirtS]

Assuming you mean v1 firmware yes its running. And 2032 shows after every single calibration.

As does 2 other friends that own M2Ps.

 

[Former Member]

Assuming you mean v1 firmware yes its running. And 2032 shows after every single calibration.

As does 2 other friends that own M2Ps.

How often are your friends drones requesting calibration

 

[gnirtS]

One has been asked once since purchasing (he moved a few hundred miles on a holiday). The other asked on arrival before first flight, not since (hasnt had it 30 days yet and hasnt moved it very far to see if that triggers again).

 

[Former Member]

I'm going to check all of the Mavic 2s that I have to see what their internal date is

 

[DanMan32]

The Mavic 2, and the Mavic Air .DAT files from the AC are encrypted. However, the tablet .DAT files are not encrypted. Look here to see how to retrieve the tablet .DAT.

Encrypted or just an uncommon compression algorithm? When I selected all available entries to download, it ended with 'Compressing Files' and only created one file rather than one for each of the selected items. 7-Zip didn't recognize the compression though. I suppose compression would effectively be encryption if you didn't know the algorithm.
Did you know that Microsoft Visio files are multiple files compressed? Not sure it is true Zip but most compression programs can read it and display all the files it contains.

 

[kilomikebravo]

I'm going to check all of the Mavic 2s that I have to see what their internal date is

I bet that is going to be interesting, TD.

I’ll check mine too if you tell me where to look. Presume it’s the DAT file on the device? If so what’s the line look like? I’ll let you know.

KB

 

[msinger]

If so what’s the line look like?

See the posts above.

 

[BudWalker]

Encrypted or just an uncommon compression algorithm? When I selected all available entries to download, it ended with 'Compressing Files' and only created one file rather than one for each of the selected items. 7-Zip didn't recognize the compression though. I suppose compression would effectively be encryption if you didn't know the algorithm.
Did you know that Microsoft Visio files are multiple files compressed? Not sure it is true Zip but most compression programs can read it and display all the files it contains.

The DJI Assistant 2 will produce a file that looks something like DJI_ASSISTANT_EXPORT_FILE_2018-09-01_18-04-59.DAT. It is compressed but it's a zlib compression scheme - not the same as a .zip file. You can use ExtractDJI which comes bundled with DatCon or CsvView to decompress this file. If there are any "real" .DATs inside the compressed file then ExtractDJI extracts them. Both DatCon and CsvView will use ExtractDJI to decompress one of these files.

With the Mavic Air and Mavic 2 the "real" .DAT is encrypted before it gets compressed. So ExtractDJI will be producing the encrypted .DAT. Also, the Mavic 2 requires a different version of DJI Assistant 2 called DJI Assistant 2 for Mavic 2". There are a few extra steps to retrieving the "real" .DAT files - but is hardly matters since they are encrypted.

 

[BudWalker]

I bet that is going to be interesting, TD.

I’ll check mine too if you tell me where to look. Presume it’s the DAT file on the device? If so what’s the line look like? I’ll let you know.

KB

Look at post #1. You'll need to use DatCon and look at the eventLog stream, set the time axis to include the stuff before the flight begins, and probably set Invalid DatHeader OK.