Conditions that lead to long flash write/erase times

[Pascal Künzi]

I am wondering what circumstances lead to the big difference of the write and erase times of the NVMC. Since the difference between the max. value is a multiple of the min. value are there certain things to pay attention or avoid?

Thanks in advance

[Sigurd]

Are you using a SoftDevice?

[Pascal Künzi]

Yes the S132. Is it only due to the restricted access of the NVMC when using a SoftDevice?

[Sigurd]

When using SoftDevice you should use the Flash memory API. Use sd_flash_write() to write and sd_flash_page_erase() to erase.

The API will schedule flash access in between radio events automatically, to avoid conflicts between the radio and the flash operations. The way this works is that you don't really know when the flash access will happen, but you will get an event after the flash operation is complete telling you whether or not the flash operation was successful.

In cases where flash access fails you can typically just retry the operation, but if you have so aggressive connection parameters that there is no room for flash operations (mainly a problem for flash erase operations) you might have to either change the connection parameters or temporarily disconnect the BLE link.

[Pascal Künzi]

Thank you! So the SoftDevice is the main reason for those max. times. Without SoftDevice they should be close to min value. Is that correct?

[Sigurd]

If you are referring to the values in the PS, Chapter 11.8.1 Flash programming, these min and max values are best/worst case values with only the NVMC. The maximum values are a result of tests performed at very high temperatures, and at low voltages. While the min values are a result of testing at ideal temperatures and ideal voltages.