How can I modify my board so that my iPhone automatically connects to it

[Chaoyue Ying]

I'm testing my code based on nrf5SDK v17.0.0 with nrf52832 board.

By nrf connect application,I can see the broadcast packet from my board.It looks something like this.

Some of my hypothetical users are iPhone users.Then I  test my board with iphone.

It is OK when I use iphone to connect,pair and bond with my board.

But when I restart the bluetooth on iphone or reenter Bluetooth broadcast range of my board, iphone will not connect to it automatically.But Android phone can connect to it automatically.

Besides,I tried the example that hid_mouse,The example can achieve this effect,but it seem based on whitelist.But I want to support multiple phones to connect and support them reconnect automatically at the same time.

[Einar Thorsrud]

Hi

The iOS Bluetooth stack will only automatically connect to a devices which has services that are natively supported, such as HID. For devices without those servics, you will need have an app running on the phone that handles the (re-)connection. This is normally requierd anyway for devices with such services in order to use/interact with the services on the device (given that they are not natively supported in iOS).

[Chaoyue Ying]

Hi, 你好,

I think something not mentioned by omission before.In the raw data,you can see the uuid in advertising packet is set as 我想有些东西之前没有被遗漏。在原始数据中，可以看到advertising packet中的uuid被设置为我想有些东西之前没有被遗漏。在原始数据中，可以看到advertising packet中的uuid被设置为我想有些东西之前没有被遗漏。在原始数据中，可以看到advertising packet中的uuid被设置为

BLE_UUID_HUMAN_INTERFACE_DEVICE_SERVICE.BLE_UUID_HUMAN_INTERFACE_DEVICE_SERVICE。BLE_UUID_HUMAN_INTERFACE_DEVICE_SERVICE。BLE_UUID_HUMAN_INTERFACE_DEVICE_SERVICE。

The services on my board as follow:我的板子上的服务如下：我的板子上的服务如下：我的板子上的服务如下：

Just add a nus service to the hid_mouse sample.That is using the same HID service with the hid_mouse sample.They have the same uuid,service,character and HID map.只需向hid_mouse示例添加一个nus服务。这是在hid_mouse示例中使用相同的HID服务。他们有相同的uuid，服务，字符和HID地图。只需向hid_mouse示例添加一个nus服务。这是在hid_mouse示例中使用相同的HID服务。他们有相同的uuid，服务，字符和HID地图。只需向hid_mouse示例添加一个nus服务。这是在hid_mouse示例中使用相同的HID服务。他们有相同的uuid，服务，字符和HID地图。

I don't know why they are different.So I guess maybe due to whitelist.我不知道它们为什么不一样。所以我想可能是白名单的缘故。我不知道它们为什么不一样。所以我想可能是白名单的缘故。我不知道它们为什么不一样。所以我想可能是白名单的缘故。

Below is my code,I hope to get further guidance from you.下面是我的代码，希望能得到你的进一步指导。下面是我的代码，希望能得到你的进一步指导。下面是我的代码，希望能得到你的进一步指导。

ETC（test）.rar

我觉得有些东西reload-alertreload-alertreload-alert

[Einar Thorsrud]

Hi,

I have looked at your project, and I don't see anything that sticks out. It includes a HID mouse service with a standard UUID as far as I can see, and just has an additional NUS service. I am not able to build your Keil project, but can you test on your end to see if for instance removing the NUS service makes a difference? Or if not, backtrack and see how much and what you need to remove to get the same behaviour as with the HID mouse sample (that iOS automatically reconnects)?

This is a bit of trial and error, as we do not know what happend internally in iOS and why it decides to not reconnect.

[Chaoyue Ying]

Hi,

Thanks for your suggestion.I found the cause of the problem.

As you can see from the file, the Bluetooth name is written this way:

char * namecat(void)
{
	ble_gap_addr_t addr;
	sd_ble_gap_addr_get(&addr);
	NRF_LOG_HEXDUMP_INFO(addr.addr,BLE_GAP_ADDR_LEN);
	static char name[13] = DEVICE_NAME;
	char addr_name[4];
	for(uint8_t i = 0; i < 2;i++)
	{
		addr_name[2*i] = num_to_char((addr.addr[i])/16);
		addr_name[2*i+1] = num_to_char((addr.addr[i])%16);
	}
	strcat(name,addr_name);
	name[13] = '\0';
	return name;
}

After the modification, the Bluetooth name is written like this:

static char name[14] = DEVICE_NAME;

char * namecat(void)
{
	ble_gap_addr_t addr;
	sd_ble_gap_addr_get(&addr);
	NRF_LOG_HEXDUMP_INFO(addr.addr,BLE_GAP_ADDR_LEN);
	char addr_name[5];
	for(uint8_t i = 0; i < 2;i++)
	{
		addr_name[2*i] = num_to_char((addr.addr[i])/16);
		addr_name[2*i+1] = num_to_char((addr.addr[i])%16);
	}
    addr_name[4] = '\0';
	//strcat(name,addr_name);
    memcpy(&name[9],addr_name,5);
	return name;
}

/**@brief Function for the GAP initialization.
 *
 * @details This function sets up all the necessary GAP (Generic Access Profile) parameters of the
 *          device including the device name, appearance, and the preferred connection parameters.
 */
static void gap_params_init(void)
{
    ret_code_t              err_code;
    ble_gap_conn_params_t   gap_conn_params;
    ble_gap_conn_sec_mode_t sec_mode;

    BLE_GAP_CONN_SEC_MODE_SET_OPEN(&sec_mode);

    //char *name = namecat();
    namecat();

    err_code = sd_ble_gap_device_name_set(&sec_mode,
                                          (const uint8_t *)name/*DEVICE_NAME*/,
                                          strlen(name/*DEVICE_NAME*/));
    APP_ERROR_CHECK(err_code);

Although the problem has been solved, what puzzles me is why these two approaches lead to different results.

[Einar Thorsrud]

Hi,

That was an interessting find. I did not test this on my end, but looking at the two namecat() implementations they look equivalent to me. Are there any differences if you inspect the name array with a debugger or print it in a log?

[Chaoyue Ying]

Hi,

There doesn't seem to be a difference between the two methods from the logs.

Now I have a new problem.

If there's a way to speed up the connection back to my board.

Adjusting the TX power is one way, but it's not enough.

I also tried to shorten the broadcast interval.

According to the manual, the minimum broadcast interval can be adjusted to 20ms, but its performance is still poor.

I even use while (NRF_LOG_PROCESS()) in the main loop; Replace idle_state_handle(); Because I guess the low power consumption also has an impact on the back connection.

So what else can I do to speed up the connection?

[Einar Thorsrud]

Hi,

As a periphearl, the only thing you can do to speed up re-connection is to increase the advertising interval. The advertising module used in the nRF5 SDK and in use in the hids_mouse sample use slow advertising after a long time (longest interval), fast advertising and aditionally high duty cycle directed advertising immediately after a disconnect to ensure fast reconnection. You can adjust when to use these modes, and the intervals used. Note however that it is also very important how frequently and for how long time the central (phone) scans, as this needs to overlap in time in order to establish a connection, and you do not have much control on how this is handeld on the phone side.