Behavior when a single PPI channel triggers 2 timer tasks

RE: Behavior when a single PPI channel triggers 2 timer tasks

snoopy20 — Fri, 26 Jun 2020 18:50:21 GMT

Need clarification on this.

RE: Behavior when a single PPI channel triggers 2 timer tasks

joeyh — Wed, 26 Jun 2019 08:22:32 GMT

About your answer to #2, it seems to be contradictory to the example provided here:

https://devzone.nordicsemi.com/f/nordic-q-a/34547/ppi-configuration-to-count-pulses

In the example, it has code like this:

// Capture counter timer using PPI to make sure this is not delayed by BLE interrupt
err_code = nrf_drv_ppi_channel_assign(ppi_channel_2, timer_read_compare_event_addr, timer_count_capture_task_addr);

// Clear counter timer using PPI to make sure this is not delayed by BLE interrupt
err_code = nrf_drv_ppi_channel_fork_assign(ppi_channel_2, timer_count_clear_task_addr);

It works as intended by the code, which is to capture the counter, then clear it.

I am confused.

Joseph

RE: Behavior when a single PPI channel triggers 2 timer tasks

XavierYin — Thu, 14 Dec 2017 10:08:47 GMT

It nice that at least there is deterministic behavior. Thanks for this clear answer from Nordic.

RE: Behavior when a single PPI channel triggers 2 timer tasks

MartinBL — Thu, 14 Dec 2017 10:04:55 GMT

Hi,

The word from the HW architects is that:

No, it does not make a difference whether you assign the main event before or after the fork event.
Unfortunately the TASKS_CLEAR will have priority over TASKS_CAPTURE and the value captured in the CC register will be 0.

RE: Behavior when a single PPI channel triggers 2 timer tasks

MartinBL — Wed, 13 Dec 2017 10:01:02 GMT

Hi, I have asked for a comment from the designers, but it might take a few days.