Bug in AP_Hott_Telem.cpp (Copter 4.0.4dev)

ArduCopter Copter 4.0
1

There is a minor bug in AP_Hott_Telem.cpp (Copter 4.0.4dev), probably because of a faulty Graupner documentation:
the gps_fix_char has to be the 3rd last byte in the GPS message, so the correct code is this:

void AP_Hott_Telem::send_GPS(void)
{
    // GPS message
    struct PACKED {
        uint8_t start_byte = 0x7c;    //#01 constant value 0x7c
        uint8_t gps_sensor_id = 0x8a; //#02 constant value 0x8a
        uint8_t warning_beeps;        //#03
        uint8_t sensor_id = 0xA0;     //#04 constant (?) value 0xa0
        uint16_t alarm;               //#05
        uint8_t flight_direction;     //#07 flight direction in 2 degreees/step (1 = 2degrees);
        uint16_t gps_speed_kmh;       //#08 km/h
        uint8_t pos_NS;               //#10 north = 0, south = 1
        uint16_t pos_NS_dm;           //#11 degree minutes
        uint16_t pos_NS_sec;          //#13 position seconds
        uint8_t pos_EW;               //#15 east = 0, west = 1
        uint16_t pos_EW_dm;           //#16 degree minutes
        uint16_t pos_EW_sec;          //#18 position seconds
        uint16_t home_distance;       //#20 meters
        uint16_t altitude;            //#22 meters. Value of 500 = 0m
        uint16_t climbrate;           //#24 m/s 0.01m/s resolution. Value of 30000 = 0.00 m/s
        uint8_t climbrate3s;          //#26 climbrate in m/3s resolution, value of 120 = 0 m/3s
        uint8_t gps_satelites;        //#27 sat count
        uint8_t free_char3;           //#28 ???
        uint8_t home_direction;       //#29 direction from starting point to Model position (2 degree steps)
        int16_t vel_north;            //#30 velocity north mm/s
        uint8_t speed_acc;            //#32 speed accuracy cm/s
        uint8_t gps_time_h;           //#33 UTC time hours
        uint8_t gps_time_m;           //#34 UTC time minutes
        uint8_t gps_time_s;           //#35 UTC time seconds
        uint8_t gps_time_hs;          //#36 UTC time 0.01s units
        int16_t vel_east;             //#37 velocity north mm/s
        uint8_t horiz_acc;            //#39 horizontal accuracy
        uint8_t free_char1;           //#40 displayed to right of home
        uint8_t free_char2;           //#41
        uint8_t gps_fix_char;         //#42 GPS fix character. display, 'D' = DGPS, '2' = 2D, '3' = 3D, '-' = no fix
        uint8_t version = 1;          //#43 0: GPS Graupner #33600, 1: ArduPilot
        uint8_t stop_byte = 0x7d;     //#44
    } msg {};
2

Thanks

Can you submit a PR to master, please

Also, if you could help adding the other sensors would be great.

Next would be to make the Graupner sensors provide data to ArduPilot :wink:

3

Hi @LuisVale ,
First: Thanks a lot for implementing Hott in ardupilot. The code works well and is well written, good job.

Honestly, i don’t know how to do that …

I don’t think, the other sensors are neccessary. The microcopter display works, and that’s the main thing.
On top of my wishlist would be: Making Arducopter parameters editable via textmode !

4

I’ve created a PR with @fs007’s changes in it here: https://github.com/ArduPilot/ardupilot/pull/13873

Is someone able to test that PR?

5

can you should a photo of the result before and after this change?
also, what receiver and transmitter are you testing on?

6

hott
hott2395×1294 390 KB
I’m using a Graupner mx-20 transmitter and a gr-16 receiver.
Here is a photo that shows the character AFTER this change. Before the change, the character isn’t shown at all.

7

Thanks! We’ve merged the fix into master. Please test and confirm it is good

8

Yes, please do this change. The documentation was correct, but for any reason we did implement it wrong.
On the mz-32 which I use, I can’t see the mistake, because it does display the 3D fix itself.

9

There’s another thing that could be improved in this lib:
A “D” should be shown when dgps mode is reached
(uint8_t gps_fix_char; //#42 GPS fix character display, ‘D’ = DGPS, ‘2’ =2D, ‘3’ = 3D, ‘-’ = no fix)

Sometimes it’s important to wait for dgps before arming.

In fact, the existing code doesn’t do that and RTK mode isn’t handled well also.

I would suggest something like this code (beginning line 279):

switch (gps.status()) {                 
    case AP_GPS::NO_GPS:
        msg.gps_fix_char = ' ';
        break;
    case AP_GPS::NO_FIX:
        msg.gps_fix_char = '-';
        break;
    case AP_GPS::GPS_OK_FIX_2D:
        msg.gps_fix_char = '2';
        break;
    case AP_GPS::GPS_OK_FIX_3D:
        msg.gps_fix_char = '3';
        break;
    case AP_GPS::GPS_OK_FIX_3D_DGPS:
        msg.gps_fix_char = 'D';
        break;
    default:
        msg.gps_fix_char = 'R';
        break;
    }
10

Again, I’m unable to test that patch myself, apart from, “does it
compile?”.