RE: Timer/Counter example using Zephyr drivers for nRF52832

Maria Gilje — Mon, 02 Oct 2023 13:27:33 GMT

Hi Mike!

Great that you found a solution which meets your requirements. Thank you for sharing!

BR,

Maria

RE: Timer/Counter example using Zephyr drivers for nRF52832

Mike Austin (LPI) — Wed, 27 Sep 2023 06:17:29 GMT

Hi Maria,

Got this sorted now.

Just needed to change:

while (!sampling_complete) {
	k_usleep(1);
}

to:

while (!sampling_complete) {
	k_yield();
}

I'm now able to get my sampling speed up to ~ 40kHz, which is more than sufficient

Regards,

Mike

RE: Timer/Counter example using Zephyr drivers for nRF52832

Maria Gilje — Thu, 21 Sep 2023 14:45:55 GMT

Hi Mike,

I will look into this soon. Thank you for your patience.

Regards,

Maria

RE: Timer/Counter example using Zephyr drivers for nRF52832

Mike Austin (LPI) — Thu, 21 Sep 2023 07:23:17 GMT

OK, have made a bit of progress on this, but still floundering along not really sure what I'm doing.

I solved the initial problem with this:

#define TIMER DT_NODELABEL(timer0)
const struct device *counter_dev = DEVICE_DT_GET(TIMER);

So, now I have a counter that seems to work. When I say "work", I mean that I can get it to trigger a callback at the value of ticks that I specify in the counter struct:

struct counter_top_cfg counter_config = {
	.ticks = 32000, // Equivalent to 2msec
	.callback = counter_callback,
	.user_data = NULL,
	.flags = COUNTER_TOP_CFG_RESET_WHEN_LATE | COUNTER_CONFIG_INFO_COUNT_UP,
};

I am using a button on the DK to trigger the start of the sampling, and my counter_callback function, that is getting triggered every 32,000 ticks looks like this:

void counter_callback(const struct device *dev, void *user_data)
{
	printk("Getting sample %d\n", sample_count);
	gpio_pin_set_dt(&sample_adc, HIGH);
	if (sample_count < MAX_ADC_SAMPLES) {
		err = adc_read(dev_adc, &sequence);
		gpio_pin_set_dt(&sample_adc, LOW);
		if (err != 0) {
			printk("ADC reading failed with error %d.\n", err);
			return;
		}
		adc_samples[sample_count] = sample_buffer;
		sample_count++;
	}
	else {
		sampling_complete = true;
	}
}

I'm basically aiming to take up to MAX_ADC_SAMPLES (currently set to 32) and when this has been achieved, I set sample_complete to true. In my main loop, I am doing this:

	while (1) {
		if (measure_adc) {

			sample_count = 0;
			sampling_complete = false;
			counter_start(counter_dev);
			while (!sampling_complete) {
				k_usleep(1);
				//printk("sampling_complete = %d\n", sampling_complete);
			}
			printk("Sampling completed\n");
			err = counter_stop(counter_dev);
			if (err) {
				printk("Unable to stop counter (err %d)\n", err);
			}
			printk("Counter stopped\n");

			measure_adc = false;
			sample_count=0;
			sampling_complete = false;

			for (i=0;i<MAX_ADC_SAMPLES;i++) {
				//printk("sample: %d	value: %d	sample time: %d\n", i, adc_samples[i], adc_sample_time[i]);
				printk("sample: %d	value: %d\n", i, adc_samples[i]);
			}
		}
		k_usleep(100);
	}

So the button press asserts measure_adc. That triggers the start of the counter, which triggers the sampling. There are a few problems I am noticing though:

1. If I reduce the value of .ticks in my counter_config structure (ultimately I need the sampling period to be 312usec), then the check on the value of sample_complete in my main loop doesn't seem to get tested, so I never stop the counter and hence stop the sampling. It just goes on sampling forever

2. The same thing seems to occur if I remove the printk("Getting sample %d\n", sample_count); in my counter_callback function. I'm presuming this is somehow adding a delay that, in the same way as not reducing the ticks value too much, allows the check on sample_complete to be done

3. Again, in a similar vain, if I don't put a suitable delay in the while(!sampling_complete) loop, this doesn't seem to work correctly.

Can anyone see what I'm doing wrong? As I said, ultimately I need to be able to trigger a sampling sequence of 32 samples taken at 312usec intervals, and store this into an array that I can then use in my RMS and FFT calcs later on in my firmware.

Best regards,

Mike