In this exercise, you will learn how clock selection affects the total current consumption of a peripheral device. You will begin by exploring the most important topics related to clock selection and configuration, starting with comparing Online Power Profiler estimations for both clock options. You will then use the PPK2 to measure average current consumption over various intervals and identify connection events.

Additionally, you will observe how the accuracy of the low-frequency clock source affects window widening and average current consumption.

Optionally, you will monitor radio transceiver activity using a logic analyzer or oscilloscope to gain deeper insight into the radio transceiver behavior.

A handful of steps in this exercise are marked as optional as they require additional hardware:

• Logic analyzer or oscilloscope
• nRF52840 DK or dongle, or any BT sniffer

Exercise steps

In the GitHub repository for this course, go to the base code for this exercise, found in l5/l5_e1.

This base code is just a modified version of the peripheral_lbs sample.

1. Disable serial logging.

Minimize the current consumption to include only the Bluetooth activity by disabling the unnecessary components and processes.

Add the following line to the prj.conf file to disable serial logging:

CONFIG_SERIAL=n
CONFIG_CONSOLE=n
CONFIG_UART_CONSOLE=n

CONFIG_SERIAL=n
CONFIG_CONSOLE=n
CONFIG_UART_CONSOLE=n

To observe and verify the current consumption, use the PPK2 using source meter mode.

Part 1 – Bluetooth LE connection

2. Change the Bluetooth LE connection interval.

The clock selection matters the most for the connection phase. When the peripheral and the central are connected, they exchange data periodically. For this phase, not only does the low-frequency clock current matter, but also the calibration procedure and peripheral reception time increase due to clock accuracy.

The default connection interval for the peripheral is 30 ms (24 * 1.25 ms). use Kconfig search to check the default value of the CONFIG_BT_PERIPHERAL_PREF_MIN_INT Kconfig option

The device with this configuration consumes around 80 µA.

This exercise demonstrates how the clock selection affects the current consumption. The difference should be around a few µA, so configure the application to have a much lower base current consumption. For applications that do not need to exchange data so often, the easiest way to do this is just to increase the connection interval.

2.1 Add the following lines in the prj.conf file:

CONFIG_BT_PERIPHERAL_PREF_MIN_INT=800       
CONFIG_BT_PERIPHERAL_PREF_MAX_INT=800
CONFIG_BT_PERIPHERAL_PREF_TIMEOUT=400
CONFIG_BT_GAP_AUTO_UPDATE_CONN_PARAMS=y

CONFIG_BT_PERIPHERAL_PREF_MIN_INT=800       
CONFIG_BT_PERIPHERAL_PREF_MAX_INT=800
CONFIG_BT_PERIPHERAL_PREF_TIMEOUT=400
CONFIG_BT_GAP_AUTO_UPDATE_CONN_PARAMS=y

This gives a connection interval of one second. The connection parameters were covered in Lesson 3 of this course.

3. Using an external crystal oscillator.

3.1 Select a low-frequency clock source. When using the nRF Connect SDK, use Kconfig search to check the default value of the CONFIG_CLOCK_CONTROL_NRF_SOURCE Kconfig option.

Observe that the external oscillator is the default source for the low-frequency clock. You can verify this in the .config file containing the Zephyr configuration parameters for the project, found in the {build}/l4_e1/zephyr/ folder.

CONFIG_CLOCK_CONTROL_NRF_K32SRC_XTAL=y
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_SYNTH is not set
# CONFIG_CLOCK_CONTROL_NRF_HFINT_CALIBRATION is not set
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_500PPM is not set
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_250PPM is not set
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_150PPM is not set
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_100PPM is not set
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_75PPM is not set
CONFIG_CLOCK_CONTROL_NRF_K32SRC_50PPM=y

CONFIG_CLOCK_CONTROL_NRF_K32SRC_XTAL=y
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_SYNTH is not set
# CONFIG_CLOCK_CONTROL_NRF_HFINT_CALIBRATION is not set
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_500PPM is not set
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_250PPM is not set
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_150PPM is not set
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_100PPM is not set
# CONFIG_CLOCK_CONTROL_NRF_K32SRC_75PPM is not set
CONFIG_CLOCK_CONTROL_NRF_K32SRC_50PPM=y

The application uses the external crystal. Notice that the system is configured to use 50 ppm accuracy of the clock by default. This fact will be helpful in the next steps of this exercise.

For the central, the clock accuracy should be 50 ppm or better.

3.2 The crystal is mounted on the nRF54L Series DK boards. It requires capacitors for the operation. The nRF54L Series SoCs are equipped with programmable internal capacitors.

You can check their configuration by browsing the devicetree (for the nrf54L15 SoC it will be: <path_to_sdk>/zephyr/boards/nordic/nrf54l15dk/nrf54l_05_10_15_cpuapp_common.dtsi).

&lfxo {
	load-capacitors = "internal";
	load-capacitance-femtofarad = <17000>;
};
&hfxo {
	load-capacitors = "internal";
	load-capacitance-femtofarad = <15000>;
};

&lfxo {
	load-capacitors = "internal";
	load-capacitance-femtofarad = <17000>;
};
&hfxo {
	load-capacitors = "internal";
	load-capacitance-femtofarad = <15000>;
};

3.3 Check the radio transmission power configuration in the {build}/l4_e2/zephyr/.config folder.

CONFIG_BT_CTLR_TX_PWR_0=y 

CONFIG_BT_CTLR_TX_PWR_0=y 

The radio TX power is 0 dBm.

 You can find more options using Kconfig search:

3.4 The connection parameters and estimated average current consumption should look like this:

• Connection interval: 1000 ms
• Peripheral accuracy: 50 ppm
• Central accuracy (estimated): 50 ppm
• TX power: 0 dBm

4. Estimating current consumption during the connection phase.

For an empty PDU packet, the total receive and transmit size is 10 B:

• 1B Preamble
• 4B Access address
• 2B PDU header
• 3B CRC

4.1 Estimate the current consumption using the Online Power Profiler.

The estimated total average current consumption is 5.3 µA.

4.2 Estimate the data exchange time and window widening.

Using 2 Mbps PHY spends around 40 µs on receiving/transmitting.

The easiest part is the peripheral transmission, which sends 10 B and takes 40 µs. For the receiving time, you can identify the calculated receiving time (40 µs) and the window widening (WW) part. You can estimate what the WW part should be for a given clock accuracy and the connection interval using the following formula:

WW = 50(periph clock accuracy) +50(central clock accuracy) /1000000 * 1 s (connection interval) + 2 * 16 µs= 132 µs

The Online Power Profile (OPP) tool shows a similar value. Note that the tool is for evaluation purposes only. The estimations include additional hardware-specific properties, and actual results can also differ by device by approximately 5%.

5. Verify the power and timing estimations during the connection phase.

When you have your estimations ready, see how they align with the practice.

5.1 Build and flash the application to the DK.

5.2 Establish a connection to the DK.

Open the nRF Connect for Mobile application on your smartphone. In the app, press the Play/SCAN icon in the upper right corner. Your smartphone now acts as a central device, scanning for any Bluetooth LE devices that are advertising in its presence. Several Bluetooth-enabled devices around you will start appearing. Choose the one called NORDIC_LBS and tap Connect.

5.3 (Optional) Sniffing the communication

You do not need any additional hardware to complete this course, other than the PPK2 and an nRF54L Series DK.

However, if you have an nRF52840 DK or dongle, or any BT sniffer you can also check the connection data. For the nRF52840 DK and dongle, the procedure for setting up and using the sniffer is described in Bluetooth LE Fundamentals Course: Lesson 6 Bluetooth LE Sniffer. Use the sniffer to verify if the payload and transmission time match the OPP configuration. You can also figure out the low-frequency clock accuracy of the central, and whether it matches the value (50 ppm) assumed in the OPP.

See the connection indication event sent by the central. You can confirm that the Sleep Clock Accuracy (SCA) is 50 ppm or better. Note that SCA can vary depending on the device.

See the packages being sent periodically. There is transmission in both directions, and there are 10 B in the Bluetooth LE Link Layer (including preamble). The transmission takes 44 µs for a single packet, which confirms what was estimated in the OPP.

5.4 Current consumption verification using the Power Profiler app.

Start measuring the current consumption using the PPK2, just like in previous lessons. Note that the connection interval is only changed from the default 30 ms to 1 s after five seconds from the connection request. In the app, you can confirm that the connection interval is correct and that the average current consumption is quite close to the previous estimate.

The idle current is 2.55 µA, which is a bit below the estimate in the OPP.

5.4 Identifying transmission phases.

When you take a closer at the communication, you should see something like this:

There are three main parts of the communication event.

• Crystal ramp-up (followed by standby time). The application works using HFINT. Enabling the external crystal is required for the radio activity.
• Radio RX (including also starting the radio and WW).
• Radio TX.

6. (Optional ) Observing radio activity using external pins.

This step requires an oscilloscope or logic analyzer, and is therefore optional.

The PPK2 is a measurement tool for current measurement rather than precisely showing the events in time. For this purpose, you can use a logic analyzer or an oscilloscope. In this step, you will measure the radio transceiver activity using a logic analyzer or an oscilloscope. In the nRF Connect SDK, the MPSL is responsible for the Bluetooth LE communication. It has a feature that allows debugging the hardware behavior.

6.1 To enable MPSL debugging, add the following line in the prj.conf file:

CONFIG_MPSL_PIN_DEBUG=y

CONFIG_MPSL_PIN_DEBUG=y

6.2 Build and flash the firmware to your DK.

You can use Kconfig search to check the default value of the CONFIG_MPSL_PIN_DEBUG_RADIO_READY_AND_DISABLED_PIN and identify the events and corresponding pins. See especially the radio transceiver events. Notice that pin P1.11 on an nRF54L Series device is toggled in sync with the activity of the radio transceiver.

6.3 Connect a probe to the pin and look at the communication event.

Looking at the signal, you can identify all TX and RX phases, with the radio switch (~150 µs) phase. The RX window is ~360 µs. The period starts when the radio is enabled and contains WW (until the package arrives), the receiving process, and completes when the transceiver is stopped. The radio switch and TX phases can be directly matched with the estimated values in the OPP.

7. Using an internal RC oscillator during the connection phase.

The use of the default and most common source for a low-frequency clock are already covered. As an alternative, you can use the internal RC oscillator as a low-frequency source. The benefits of both solutions were covered in Clock sources. When using an internal oscillator and calculating the power budget, you need to take into account the following factors:

• Calibration procedure
• Longer RX period (due to window widening procedure)
• Higher run current for internal RC: +0.6 µA

7.1 Enable internal RC oscillator.

To enable an internal RC oscillator, add the following line to the prj.conf file:

CONFIG_CLOCK_CONTROL_NRF_K32SRC_RC=y

CONFIG_CLOCK_CONTROL_NRF_K32SRC_RC=y

7.2 Build and flash the firmware to your DK.

The internal oscillator comes with reduced accuracy. It requires a calibration procedure performed periodically to achieve and keep this level of declared accuracy. You need to know which low-frequency clock accuracy to use for the current consumption estimation. See the low-frequency clock settings in the Kconfig search.

Clock accuracy (CLOCK_CONTROL_NRF_ACCURACY_PPM):

Calibration period (CLOCK_CONTROL_NRF_CALIBRATION_PERIOD):

Clock summary:

• 250 ppm accuracy
• 4 s/8 s calibration period (depends on temperature)

7.3 You now have all the parameters for the OPP to estimate the current consumption, including the parameters already known from the previous steps of this exercise:

• 50 ppm central SCA
• 1000 ms connection interval
• 2 Mbps PHY
• 0 dBm TX power

The total average current consumption is 7.6 µA (5.3 µA using an external crystal oscillator), so the difference is more than 2 µA. The window widening area is also much larger. You can compare it with the WW from the previous steps. The rest of the communication stays on a similar level.

7.4 In the Power Profiler app, after programming the application, you can see the calibration procedure every 4 s and three connection events due to the 1 s connection interval.

The calibration procedure consumes a lot of energy and CPU time:

7.5 You can also confirm the increased idle current (due to internal RC run current): 3.28 µA (2.55 µA for external crystal):

7.6 As a summary, the average current consumption is 8.3 µA, which is pretty close to the OPP estimations.

7.7 (Optional) Analyzing the activity of the radio transceiver.

Use the scope or logic analyzer to see how the RX window (including WW) changed when using the internal RC:

You can identify when the radio is in receiving mode (~550 µs), the radio mode switch (~150 µs), and transmission (~40 µs). These values are quite close to the OPP estimations.

Part 2 – Bluetooth LE advertising

8. Configuring the advertising interval.

In this part, you will observe the current consumption during the advertisement process.

8.1 Define the advertising setting by adding the following snippet in main.c.

#define BT_LE_ADV_CONN_100                                                                      \
	BT_LE_ADV_PARAM(BT_LE_ADV_OPT_CONN, BT_GAP_ADV_FAST_INT_MIN_2, BT_GAP_ADV_FAST_INT_MIN_2,     \
			NULL)

8.2 Apply defined settings for the advertising process.

Replace the call to bt_le_adv_start() in main.c, with the following code line

int err = bt_le_adv_start(BT_LE_ADV_CONN_100, ad, ARRAY_SIZE(ad), sd, ARRAY_SIZE(sd));

The advertising interval is now: 100 ms

8.3 To use the default low-frequency clock source, disable the internal oscillator in the prj.conf file:

CONFIG_CLOCK_CONTROL_NRF_K32SRC_RC=n

CONFIG_CLOCK_CONTROL_NRF_K32SRC_RC=n

8.4 Build and flash the firmware to your DK.

9. Estimating the current consumption for the internal oscillator and external crystal.

9.1 Find the size of the advertisement payload.

To estimate the current consumption, you need to know the advertisement payload, which is defined in main.c file. It contains the following information:

• Flags: 1 B len + 1 B type + 1 B data = BT_LE_AD_GENERAL | BT_LE_AD_NO_BREDR
• Device name: ( 1 B len + 1 B type + 10 B data + “Nordic_LBS”

static const struct bt_data ad[] = {
	BT_DATA_BYTES(BT_DATA_FLAGS, (BT_LE_AD_GENERAL | BT_LE_AD_NO_BREDR)),
	BT_DATA(BT_DATA_NAME_COMPLETE, DEVICE_NAME, DEVICE_NAME_LEN),
};

static const struct bt_data ad[] = {
	BT_DATA_BYTES(BT_DATA_FLAGS, (BT_LE_AD_GENERAL | BT_LE_AD_NO_BREDR)),
	BT_DATA(BT_DATA_NAME_COMPLETE, DEVICE_NAME, DEVICE_NAME_LEN),
};

The payload is 15 B.

9.2 Find out how to select a low-frequency clock source. Like in the connection part of the exercise, you see that the default clock source is an external crystal with 50 ppm accuracy.

9.3 Check the default setting of the radio transmission power in the {build}/l4_e1/zephyr/.config file.

CONFIG_BT_CTLR_TX_PWR_0=y

CONFIG_BT_CTLR_TX_PWR_0=y

The radio TX power is 0 dBm.

9.4 Now you have all the data that you need to estimate the current consumption for your device. Use the Online Power Profiler to estimate the average current consumption with the the following values:

• 100 ms advertising interval
• 15 B advertisement payload
• 50 ppm external crystal driving low-frequency clock
• 0 dBm Radio TX power

9.5 Estimate the current consumption for the internal RC oscillator.

The total current consumption is approximately one µA higher when using an external crystal. As the window widening is applied only when the connection is established, the increased current consumption is caused by higher run current and the calibration procedure.

10. Verifying the estimations.

10.1 Confirm that your application works as expected. Verify the advertising interval using the PPK2 to measure this.

10.2 (Optional) Payload verification.

When you capture the proper advertisement packet using Bluetooth LE sniffer, you can look into the timings and the structure.

The payload and packet transmission time match your design and estimation. It is 15 B Payload + 16 B Bluetooth frame (headers, addresses) = 31 B = 248 b. Using 1 Mbps, it gives 248 µs transmission time.

10.3 Measure the average current consumption.

Look into average current consumption. The application uses an external crystal by default.

10.4 To configure the project to use an internal oscillator again, enable the following Kconfig option in the prj.conf file:

CONFIG_CLOCK_CONTROL_NRF_K32SRC_RC=y

CONFIG_CLOCK_CONTROL_NRF_K32SRC_RC=y

10.5 Build and flash the firmware to your DK and measure the current consumption:

The current consumption has increased slightly. The measured values are not far from the estimations.