Strange readings

[rjblake]

I used your .bin file to update and now I get very strange readings compared to Aaron's. xx:xx:xx is your firmware and yy:yy:yy is Aaron's. Both are configured to used Custom and not MiLike and details on screen are showing correctly. Any thoughts?

2021-01-02 14:13:44 - Device: A4:C1:38:xx:xx:xx Temp: 1280.8c Humidity: 253% Batt: 16% 2021-01-02 14:13:49 - Device: A4:C1:38:yy:yy:yy Temp: 20.6c Humidity: 48% Batt: 92%

[pvvx]

Firmware Ver 0.8, Advertising data:

struct _packed_ {
    uint8_t     size;   // = 18
    uint8_t     uid;    // = 16, 16-bit UUID
    uint16_t    UUID;   // = 0x181A, GATT Service 0x181A Environmental Sensing
    uint8_t     mac[6];
    int16_t     temperature; // x 0.01 degree
    int16_t     humidity; // x 0.01 %
    uint16_t    battery_mv; // mV
    uint8_t     battery_level; // %
    uint8_t     counter; // measurement number
    uint8_t     flags; 
} adv_custom_t;

[rjblake]

Thanks - didn't realise that you'd changed the structure of the advertising. Will need to modify my code accordingly

[ralf-e]

does that mean the FW is no longer compatible with Tasmota BLE or similar? I ask before I update ...

[pvvx]

Advertising Mi

// 16-bit UUID for Members 0xFE95 Xiaomi Inc. https://btprodspecificationrefs.blob.core.windows.net/assigned-values/16-bit%20UUID%20Numbers%20Document.pdf
typedef struct __attribute__((packed)) _adv_mi_t {
    uint8_t     size;   // = 21
    uint8_t     uid;    // = 0x16, 16-bit UUID https://www.bluetooth.com/specifications/assigned-numbers/generic-access-profile/
    uint16_t    UUID;   // = 0xFE95, 16-bit UUID for Members 0xFE95 Xiaomi Inc.
    uint16_t    flg;    //
    uint16_t    dev_id; //
    uint8_t     counter; // measurement number
    uint8_t     MAC[6];
    // +15: 0x0A, 0x10, 0x01, t_lv, 0x02, b_lo, b_hi
    // +15: 0x0D, 0x10, 0x04, t_lo, t_hi, h_lo, h_hi
    uint8_t     data_id;    // = 0x0A or 0x0D
    uint8_t     nx10;       // ? = 0x10
    union {
        struct {
            uint8_t     id1; // = 0x01
            uint8_t     battery_level; // 0..100 %
            uint8_t     id2; // = 0x02
            uint16_t        battery_mv;
        }t0a;
        struct {
            uint8_t     id; // = 0x04
            int16_t     temperature; // x 0.1 degree ?
            uint16_t    humidity; // x 0.01 % ?
        }t0d;
    };
} adv_mi_t, * padv_mi_t;

[pvvx]

does that mean the FW is no longer compatible with Tasmota BLE or similar?

Advertising format 'Tasmota BLE' ? Data format: little-endian / big-endian ? Xiaomi used little-endian. Firmware atc1441 - big-endian. My - little-endian.

[ralf-e]

"tasmota BLE MI ESP32" currently receives either ATC-like or MI-like (original FW, encrypted with KEY or unencrypted for LYWSD02), MI-Like with ATC-FW does not work, no data (4. Sensor in the screenshot). I therefore assume that your current version no longer works with Tasmota either

[pvvx]

The firmware revision is not over yet. The Advertising format will change in new versions ... Current versions have a status - alpha (test only).

[pvvx]

"tasmota BLE MI ESP32" currently receives ...

Tasmota is Proprietary Software. No spec and open-source. This is not supported. Likewise, there will be no normal support ESP32 - is just another proprietary product.