Power consumption of DFU at start and end of upload

[Thomas N.]

Dear Nordic Team,

We are building a very low power wearable with the nRF52810, only powered by a small solar cell and a super-capacitor.

We would like to offer firmware updates for our product, but unfortunately a long and high-power spike at the beginning and end of the DFU process prohibits this. The power spike is too long to be buffered by the 220uF ceramic capacitor and too high for the 200 Ohm internal resistance of the Super-Capacitor.

Here you see the power spike at the beginning of the upload, measured with an oscilloscope and a 22 Ohm serial resistor:

... and a much longer spike at the end (i guess during rebooting)

The test setup was:

• nrf52DK pca10040
• SDK 15.2, Softdevice v6.1.0
• Secure Bootloader example (pca10040_ble) with a reduced connection interval to min: 100 - max: 200
• DFU package ble_app_buttonless_dfu_without_bonds_s132.zip from the secure_dfu_test_images/ble/nrf52832 folder

My questions:

1. Can you reproduce this measurement?
2. Can you suggest a way to reduce this long power spikes by software?

Best regards & thanks in advance, Thomas

P.s.: i had a look at existing posts in this forum, the only relevant one i found was this one. The answer didn't help with our problem though.

[Simonr]

Hi Thomas

This seems like rather normal currents for a DFU update, and your options are rather limited I'm afraid. What's causing these currents are decryption/encryption, as well as hash validation of firmware and flash write in the end. What you could do is remove all security from the DFU update, as this is what draws the bulk of the current here, but I'm guessing that's not a viable solution.

Alternatively, you will have to find a peripheral chip that is able to do this in HW for less overall current consumption, for this I don't have any suggestions sadly.

Best regards,

Simon