Jeti EX Bus GPS not working

[andypnz]

Jeti EX Bus GPS not working

[andypnz]

Ok, managed to sort that - the code was not handling southern hemisphere -ve latitude (or -ve long either)......

If you like I can share with you the changes I made to fix this, and also improve the displayed gps resolution....

[aepub]

I also use JETI. When using the RP2040, the GPS has no signal, and the GPS TX is connected to the 6 pins of the RP2040. Replace with ATmega328PB, and the GPS signal is normal.
The author has not replied to any questions recently.

[aepub]

After I modified led. c and led. h, the RGB indicator light flickered. At least that way, I knew that rp2040 was working. I tested MS5611, ACS758-150B, and the test voltage, and everything was basically normal. The only problem currently is that GPS data is not displayed on the remote control.

[dgatf]

Ok, managed to sort that - the code was not handling southern hemisphere -ve latitude (or -ve long either)......

If you like I can share with you the changes I made to fix this, and also improve the displayed gps resolution....

Yes, please share. Thank you

Is it with the RP2040?

[andypnz]

Yes, RP2040

Its been a while since I looked at this :-)

project 230218_1554hrs_WORKING.zip

I have attached the project directory - but from memory it was just changes to these two files. But if these don't cover it I am afraid you will need to do some file comparisons with my attached zip.....

jeti.c, around line 280 change to:

               int32_t value = (int32_t)*sensor->value;
               if (value < 0)
               {
                    format |= 1 << 6;
                    value *= -1;
                }
                *(buffer + *buffer_index) = sensor_index << 4 | sensor->type;
                uint8_t degrees = value / 60000;
                uint32_t degreesDecimals = ((value % 60000)); 
                *(buffer + *buffer_index + 1) = (uint16_t)degreesDecimals;      // degrees (dec, l)
                *(buffer + *buffer_index + 2) = (uint16_t)degreesDecimals >> 8; // degrees (dec, h)

nmea.c, around lines 23-24:

    *parameter.lat = 692761.0;               //-692.761166667;  // 11º32'45.67" +N, -S
    *parameter.lon = 10320.0;               //-2500; //-1251.964833333; // 20º51'57.89" +E, -W

around line 131:

            *parameter->lat =(((atof(degrees) * 60.0 + minutes)) *1000.0) * lat_dir ;

around line 139:

            *parameter->lon = ((atof(degrees) * 60 + minutes)) * 1000 * lon_dir; 

[dgatf]

Ok, I see. Negative coordinates where not correct. I'll add value *= -1; to fix it. Commit 2b9397b
I guess multiplier 1000 is to increase accuracy. I may add it later

But based on the debug, the issue was related to gps module not being read by rp2040 and that's why all were zeros in the transmitter

[andypnz]

I think you really need to add the multiplier - without this, the GPS accuracy is very poor.......

[dgatf]

The issue was that the formatting for coordinates for jetiex was wrong. No need to add multipliers. Fixed with 8bf664f

As a side note:

Jeti uses 3 decimals on minutes. This is 2m accuracy. Float has a mantissa of 7 digits, which is ok in most cases for 3 decimals on minutes, except for longitudes > 160°, where accuracy drops to 20m

NMEA protocol sends 5 decimals on minutes, which is 0.2m accuracy (usual gps modules don't have this accuracy though). This requires 10 digits mantissa. A double (16 digits mantissa) instead a float is required to hold all digits. This implies to create sensors for all rx protocols that holds a double instead a float. I may do this as an enhancement