Inconsistent ADC Behaviour with DFU

RE: Inconsistent ADC Behaviour with DFU

Jørgen Holmefjord — Wed, 13 Sep 2017 11:43:31 GMT

You will allways go through the bootloader when restarting the chip, to check for a valid application. If you are setting the pins in the bootloader code, this will be retained when executing the application.

RE: Inconsistent ADC Behaviour with DFU

eyigiter — Wed, 13 Sep 2017 09:31:49 GMT

Hi Jorgen,

Sorry, I didn't notice your suggestion while writing my comment. Yes, changing pin conffiguration solved the issue. I don't know if it is an expected behaviour but it seems the chip carries the pin configuration from DFU code to the application code. After correcting the pin configuration in app code, all looks fine. Thanks for all your help.

RE: Inconsistent ADC Behaviour with DFU

eyigiter — Wed, 13 Sep 2017 09:04:49 GMT

Hi Jorgen,

I found the root cause of the problem but it looked like a bug on SDK and I still need a solution for this.

I noticed that I set the pin 1 as output in bootloader code as follows:

nrf_gpio_cfg_output(1); nrf_gpio_pin_set(1);

But I am expecting the application code to reset this pin as input in my main code with the following code:

nrf_gpio_cfg_input(1,NRF_GPIO_PIN_NOPULL);

But it doesn't. Do you have any idea why it is not reset as input in application code or may you offer a workaround for me?

Regards

RE: Inconsistent ADC Behaviour with DFU

Jørgen Holmefjord — Wed, 13 Sep 2017 08:00:48 GMT

Could you try configuring the GPIO pull settings using the GPIO pin number in the call to nrf_gpio_cfg_input()? The define you are using will not evaluate to the correct GPIO pin number (AIN2 = P0.01):

NRF_ADC_CONFIG_INPUT_2 = ADC_CONFIG_PSEL_AnalogInput2  = (4UL)

Try changing from this:

nrf_gpio_cfg_input(NRF_ADC_CONFIG_INPUT_2,NRF_GPIO_PIN_NOPULL);

To this:

nrf_gpio_cfg_input(1,NRF_GPIO_PIN_NOPULL);

RE: Inconsistent ADC Behaviour with DFU

eyigiter — Wed, 13 Sep 2017 06:30:36 GMT

Also may it still be a hardware issue when the problem occurs when I transmit the package over DFU only? Without DFU, it works as expected.

Edit: In my previous messages, I have posted the following image and marked the HIGH value as expected analog value but now I noticed that it is not correct. Expected analog value must be at 1V level, but in the image it is about 3V.

pasteboard.co/GK23FHl.jpg

So to summarize the issue:

When we load the application on top of S130 by using nrfGo, it works as expected (we can measure the analog signal about 1V)
When we load the application on top of S130+DFU by using NRF Toolbox, analog signal goes to 3V level during normal operation which should have been 1V actually, and draws a wavesquare each time ADC sampling occurs. Once the ADC sampling stopped, it goes to 3V level again. May it be an internal pullup-pulldown issue during ADC sampling?

RE: Inconsistent ADC Behaviour with DFU

eyigiter — Wed, 13 Sep 2017 03:29:19 GMT

Actually I have a DK but it is really painful to revise the code to make it working on DK. There are too many inputs and outputs to make the code working as intended. Is there another way to understand the reason of voltage drop om AIN2 during ADC sampling?

RE: Inconsistent ADC Behaviour with DFU

Jørgen Holmefjord — Tue, 12 Sep 2017 19:47:14 GMT

Do you have nRF51 DK that you can test the firmware on, just to make sure no hardware issues are causing this?

RE: Inconsistent ADC Behaviour with DFU

eyigiter — Tue, 12 Sep 2017 14:07:29 GMT

Here is a more detailed waveform of the AIN2 measurement:

link text

RE: Inconsistent ADC Behaviour with DFU

eyigiter — Tue, 12 Sep 2017 11:42:11 GMT

Hi Jorgen,

I am using a custom board and I don't use P0.01 for any other purposes. It is only for measuring an analog input. When I check the signal with ossiloscope, I can see the expected waveform on the pin. However as soon as the sampling started it is voltage goes down as shared in the previous image link and returns to the expected level again. It is repeated at all sampling cycle.

RE: Inconsistent ADC Behaviour with DFU

Jørgen Holmefjord — Tue, 12 Sep 2017 11:29:45 GMT

Are you testing this on the nRF51 DK, or some custom board? Have you configured GPIO P0.01 (AIN2) for any other purposes in the code?

RE: Inconsistent ADC Behaviour with DFU

eyigiter — Tue, 12 Sep 2017 09:50:12 GMT

Hi Jorgen,

It is bootloader_secure in SDK_12.1.0 and I am using NRF Toolbox to transport the PKG.

RE: Inconsistent ADC Behaviour with DFU

Jørgen Holmefjord — Tue, 12 Sep 2017 09:36:55 GMT

Which SDK version and DFU transport are you using?

RE: Inconsistent ADC Behaviour with DFU

eyigiter — Tue, 12 Sep 2017 09:00:25 GMT

An additional information: Problem is solved if I connect the analog circuit to NRF_ADC_CONFIG_INPUT_4 instead of NRF_ADC_CONFIG_INPUT_2. So the problem looks related to AIN2 when chip has DFU code. I am still looking for a solution :(

RE: Inconsistent ADC Behaviour with DFU

eyigiter — Mon, 11 Sep 2017 13:05:44 GMT

Following is my ADC setup:

nrf_gpio_cfg_input(NRF_ADC_CONFIG_INPUT_2,NRF_GPIO_PIN_NOPULL);

adc_channel_config.config.config.reference = NRF_ADC_CONFIG_REF_SUPPLY_ONE_THIRD
adc_channel_config.config.config.input = NRF_ADC_CONFIG_SCALING_INPUT_ONE_THIRD;
adc_channel_config.config.config.ain = NRF_ADC_CONFIG_INPUT_2;

nrf_drv_adc_channel_enable(&adc_channel_config);

I am getting ADC samples at every 1ms. The issue occurs only if DFU is programmed to the chip. (S130+DFU+Application) By the way, you can see the waveform of NRF_ADC_CONFIG_INPUT_2 once the ADC is enabled where it is stable until ADC started.

link text

RE: Inconsistent ADC Behaviour with DFU

Jørgen Holmefjord — Mon, 11 Sep 2017 08:53:02 GMT

Could you elaborate on your statement "are fluctuating in a large margin"? How is the ADC setup, and what sample rate are you running at? Is the issue only with DFU, or also with other BLE activity?