ADC issue (strange things happen when measuring)

RE: ADC issue (strange things happen when measuring)

Sébastien Despont — Wed, 04 Jun 2014 15:01:12 GMT

I Ole, I cann't understand why the measure would be more accurate if the 16MHz is enabled. COuld you explain? Or maybe you would say that the time spent to execute a measure is more precise with 16MHz? Thank you in advance for explaining that.

RE: ADC issue (strange things happen when measuring)

Marius R. — Mon, 10 Mar 2014 17:00:00 GMT

I don't know what is going on. Either there was an issue in my project or maybe mdk mixed something up. But I got it running with xour world-class help. Thank you very much ole.

RE: ADC issue (strange things happen when measuring)

Ole Morten — Mon, 10 Mar 2014 16:06:56 GMT

I just tested your code, and I can't see any problems with it, so I suspect there is something else going on here. I edited my answer above to include a zip with a code project with your code. Can you test it and see what results you get with it?

RE: ADC issue (strange things happen when measuring)

Marius R. — Thu, 06 Mar 2014 12:26:37 GMT

I removed (ADC_CONFIG_INPSEL_AnalogInputOneThirdPrescaling << ADC_CONFIG_INPSEL_Pos) and at first it was not working but after just recompiling and downloading afterwards a couple of times it is working now. This makes me wondering a little bit.

As far as I know with the voltage levels I apply to the AIN pins this setting is necessary, isnt it?
Is there an issue causing this compiler behavior? I use MDK+S110 v6.0 + SDK5.1.

RE: ADC issue (strange things happen when measuring)

Marius R. — Thu, 06 Mar 2014 12:06:57 GMT

Sadly I do enable the adc in main()

RE: ADC issue (strange things happen when measuring)

Ole Morten — Thu, 06 Mar 2014 11:14:59 GMT

You seem to lack actually enabling the ADC, by setting the ENABLE register. Try setting this before starting the conversion, and then disabling it afterwards.

Also, the ADC will be more precise if you start the external 16 MHz clock before the reading. If you use the softdevice, this can be done by calling sd_clock_hfclk_request(), and then _release() afterwards. If you don't use the softdevice, the same can be done with the TASKS_HFCLKSTART and HFCLKSTOP in the CLOCK peripheral.

Edit: I just added your code to the ble_app_template from the SDK, and it seems to work as expected, as long as I don't have anything else connected to the pins in question. Remember that AIN2-AIN4 are P0.01 to P0.03, so if you have the default setup, those will be bound high externally, on the nRFgo motherboard.

I've attached the code I used.

devzone-marius-adc.zip

RE: ADC issue (strange things happen when measuring)

Marius R. — Thu, 06 Mar 2014 10:18:48 GMT

I forgot to mention, PIN_CNF is set like: nrf_gpio_cfg_input(1,NRF_GPIO_PIN_NOPULL); nrf_gpio_cfg_input(2,NRF_GPIO_PIN_NOPULL); nrf_gpio_cfg_input(3,NRF_GPIO_PIN_NOPULL); for AIN2,3,4. So the input buffers should be connected. I think thats right?