Timer1 NRF_TIMER1->CC[1] not having an effect

RE: Timer1 NRF_TIMER1->CC[1] not having an effect

Mahesh Venkitachalam — Fri, 17 Jul 2015 09:36:45 GMT

OK, got it. Thanks.

RE: Timer1 NRF_TIMER1->CC[1] not having an effect

Susheel Nuguru — Fri, 17 Jul 2015 09:24:09 GMT

Note that Timer works at 16Mhz CLK (same as Cortex M0). When you put 100 in CC, you expect that your interrupt latency is less than 100 cycles(12uS). This might not be true, so you are missing few interrupts.

RE: Timer1 NRF_TIMER1->CC[1] not having an effect

Mahesh Venkitachalam — Fri, 17 Jul 2015 09:12:48 GMT

Does the accuracy vary with the value set in CC? When I set it to 100, I get a pin period of 24.8 us. The expected value is 2 * 100 * 62 ns = 12.4 us, isn't it? I seem to be missing a factor of 2.

RE: Timer1 NRF_TIMER1->CC[1] not having an effect

Mahesh Venkitachalam — Fri, 17 Jul 2015 08:47:45 GMT

OK, thanks.

RE: Timer1 NRF_TIMER1->CC[1] not having an effect

Susheel Nuguru — Fri, 17 Jul 2015 07:49:54 GMT

I do not understand that dependency. Make sure that you are not using timer1 for something else.

RE: Timer1 NRF_TIMER1->CC[1] not having an effect

Mahesh Venkitachalam — Fri, 17 Jul 2015 07:05:55 GMT

Great! I am seeing 0.138 ms. I appreciate your patience in explaining this to me. One more question - I had to call start_timer() after starting BLE advertizing, or it doesn't advertize. Is that expected?

RE: Timer1 NRF_TIMER1->CC[1] not having an effect

Susheel Nuguru — Fri, 17 Jul 2015 05:46:41 GMT

With prescalar 0, 1 timer tick is 62 nano second long. If you put 1000 in CC, it should generate event in 62 micro second. Since you are toggling pin at every event, your pin period will be 125 micro second.

RE: Timer1 NRF_TIMER1->CC[1] not having an effect

Mahesh Venkitachalam — Fri, 17 Jul 2015 03:53:57 GMT

Aryan, could you please comment on the observed period/frequency of the signal at P0.21? With prescaler set to 0, and CC[1] set to 1000, what is the expected pulse frequency?

RE: Timer1 NRF_TIMER1->CC[1] not having an effect

Mahesh Venkitachalam — Fri, 17 Jul 2015 01:54:44 GMT

Thanks, Aryan. I guess calling NRF_TIMER1->TASKS_CLEAR = 1; manually also would work? So what is NRF_TIMER1->EVENTS_COMPARE[1] = 0; doing? It's clearing the compare event, but it doesn't effect the timer counting?

RE: Timer1 NRF_TIMER1->CC[1] not having an effect

Susheel Nuguru — Thu, 16 Jul 2015 16:55:40 GMT

you have the enable the SHORT that clears the the TIMER value after CC[0] event is triggered. Else no matter what you set in CC register, the timer will go circular over 16bit CC register