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,

Are you aboe to rre-connect manually? If so, the whitelist is not the problem, as if using whitelist and the peer is not whitelisted, there would be no way to connect.

However, I notice anothe rimportant point looking at the screenshot, and that is that there is no bond. You need to establish a bond for the reconnection to happen. With Andrioid, an app can explicitly initiate bonding, but with iOS the app cannot do it. One solution can be to make sure that one of the characteristic in a service you use requier security, and then iOS will atomatically initiate bonding when attempting to use that service. (I see that you requier SEC_JUST_WORKS for HID which should be OK though, as proivded that HID is used).

[Chaoyue Ying]

Hi,

At present, it is necessary to manually connect the iPhone to NRF52832, but I hope that the iPhone can automatically connect after bonding.

The screenshot is to show the services defined in 52832, so there is no bonding.

I did an experiment, after the iPhone scanned and bond 52832 in the Settings -> Bluetooth interface, turned off the Bluetooth switch and restarted, 52832 always showed no connection, just like the video below:

As you can see, I had to manually tap after the Bluetooth restarted to trigger the phone's connection to the 52832(Bluetooth name is Anny_ETC_1961).

[Einar Thorsrud]

Hi,

There is unfortunately not much we can do about iOS behaviour as that is controlled by Apple, but I am a bit suprised about this. What happens if you insteead of toggling Bluetooth off and on again on the phone resets the nRF devices? Does the phone re-connect automatically then?

[Chaoyue Ying]

Hi,

When I instead of toggling Bluetooth off and on again on the phone resets the nRF devices,the result is same as above.

I'm not sure if you noticed that I uploaded the project files earlier. If yes, I hope you can give me some suggestions for improvement according to the code.

而不是切换蓝牙开关手机重置nRF设备

[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.