My project focuses on minimizing power consumption while transmitting data from sensors connected to a peripheral via BLE to a central device. By sending the data as ManufacturerSpecificData within an Advertise packet without establishing a connection, significant reductions in power consumption can be achieved.
‘System_ON_Sleep of XIAO BLE - #31 by msfujino’
Since it spends most of its time in sleep mode, it is important that the sleep current value is as low as possible. Additionally, although transmission occurs only occasionally, it is also important that the current consumption during advertising is low.
I have previously reported on sleep current and advertising current for XIAO C3, C6, S3, nRF52840, and MG24 with built-in BLE functionality. With the release of the nRF54L15, I compared its advertising current with that of the nRF52840 and MG24.
‘Comparison of Sleep Current for nRF54L15, nRF52840, and MG24’
‘BLE Advertising Current Comparison, ESP32C3,C6,S3, MG24, nRF52840’
This report compares the average current consumption of my project using a sequence where it advertises 10 times per second and then sleeps for 59 seconds. Since it does not connect, it sends the same data 10 times to increase reception reliability. Furthermore, assuming short-range operation, the transmission power was set to 4dBm.
Only the Arduino IDE, available Board Support Package, and Seeed-provided board definitions for VsCode+PlatformIO were used. Dedicated chip IDEs or low-power coprocessors were not considered.
The current waveforms during advertising and sleep modes, along with a consumption current comparison graph, are shown. Within the limited scope of my project’s assumptions, the conclusions are:
For simple BLE applications, the nRF54L15 cannot be said to be overwhelmingly superior to the nRF52840.
While the MG24 has a larger transmit pulse current, its sleep current is extremely low. Therefore, it may be advantageous over the nRF52840 and nRF54L15 for applications where the transmission frequency exceeds 15 minutes.
I hope this proves helpful to those interested.
The code used for the experiments is attached. AdvertisingCurrent.zip (102.8 KB)
Thanks for the information. I am surprised to see the NRF is higher current in sleep? I’m not surprised at the MG24 transmit current - are you not using the full 19.5dBm? I get around 150mA for that… Useful for a device that is a few hundred meters from my nearest router.
And Great work as per Usual.
Even the Nordic folks would agree with these findings in the realm or the 52840, The only reason for the RISC co,LowPower processor is for the LOWEST power consumption per compute unit in the industry.
New LiB’s and SoftDevice are currently in DEV to unleash the next level, Stay tuned
When it hits, batteries can be swapped to super capacitors Things like ambient light to recharge.
IMO , it all comes down to how much transmitting you do as it has the greatest effect on the overall numbers. I found Less is more, with Xmit power unless I’m on the edge of the distance.
I like the pace for which it’s picking up… nRF54L15
New discoveries make it interesting again.
Absolutely, here it’s part of the Express dev course released a few weeks ago I completed it a few days ago, You verify the results with the nRF54L15 DK and the PPK2 under (test.hex) files in the last part, . it’s all explained in the excellent documentation they provide for all there stuff. IMO. (code , HEX is included)
Things like ambient light to recharge.
PJ 
