RTC Timer re-initialization

RE: RTC Timer re-initialization

ovrebekk — Tue, 02 Nov 2021 10:13:22 GMT

Hi David

The peripheral registers are running off of the 16MHz peripheral clock, so you should be able to access these much faster than the LFCLK, but the RTC module itself might not read the value until the next tick happens.

I notice from the PRESCALER section in the product specification that you can only write the PRESCALER value while the RTC is stopped, and quite possibly you need to wait for one LFCLK period after stopping the RTC in order to ensure it is fully stopped.

The CC registers does not have this limitation, and can be changed at any time.

While it is a bit more work to change the PRESCALER value, this is not something that is usually done dynamically. Most applications use a static PRESCALER value, and simply change the CC values instead when changing the wakeup interval.

Best regards
Torbjørn

RE: RTC Timer re-initialization

TigerD — Mon, 01 Nov 2021 14:00:51 GMT

Hi Torbjorn,

I am good and you can close the case for me. Thank you for your help.

But just as feedback, I have tested that CC[] can be dynamically reinitialized, except prescaler. Although the re-initialization is in the ISR, but the interrupt and tasks are stopped before initialization. So it shouldn't matter. Also RTC is on LFCLK, but the registers for RTC is on another CLK as CPU? I validated prescaler by read-out into a variable. I am not sure why CC can be reinitialized while Prescaler not. As I said, I am good and I can fix the prescaler all the time, and just change CC values.

Best Regards,

David Zhou

RE: RTC Timer re-initialization

ovrebekk — Mon, 01 Nov 2021 13:46:42 GMT

Hi David

Sounds good. Unless you have more questions I will consider the case resolved then 🙂

Best regards
Torbjørn

RE: RTC Timer re-initialization

TigerD — Mon, 01 Nov 2021 12:48:00 GMT

Hi Torbjorn

Thank you so much for your code review, explanation, and solutions for me.

I now realized that RTC in on LFCLK, it needs time to settle or for prescaler, etc. to take effective. I will resolve it my self based on my application.

Best Regards,

David

RE: RTC Timer re-initialization

ovrebekk — Mon, 01 Nov 2021 12:40:08 GMT

Hi David

I can't see any obvious mistake in the code, other than the fact that it is a bit unorthodox to reconfigure the entire peripheral and it's interrupt inside its own interrupt handler.

What happens if you rather try just to clear the RTC timer when the EVENTS_COMPARE[1] interrupt happens, and only do a full re-initialize every 10 seconds or so (either by using a separate timer to schedule this, or by checking a counter inside the interrupt and wait for the counter to reach a certain value)?

Also, you could try just to run the rtc_initialize() function from main rather than interrupt context, by setting a flag in the interrupt handler that you can check from the main loop.

Best regards
Torbjørn

RE: RTC Timer re-initialization

TigerD — Mon, 01 Nov 2021 12:32:42 GMT

Hi Dmitry,

Thank you for your help. I believe this is the case. I will test it out.

Best Regards,

David Zhou

RE: RTC Timer re-initialization

Dmitry — Sat, 30 Oct 2021 10:09:35 GMT

Hi,

if I understood your question right - RTC is clocked from LFCLK and thus has up to 46 usec delay from writing to TASK register to actual task execution. Try to add nfr_delay_us(50) after stopping timer - this will ensure that PRESCALER is written when timer is stopped.