SAADC Noise issue

RE: SAADC Noise issue

Kenneth — Mon, 15 Feb 2021 10:35:46 GMT

You would need to uninit the saadc and re-configure it with only one channel with oversampling and back and forth in such case unfortunately.

Kenneth

RE: SAADC Noise issue

KellyJust — Sat, 13 Feb 2021 18:08:04 GMT

Thanks Kenneth have been able to successfully implement oversampling and remove the noise. Now another issue is that I have to use another analog input and I read somewhere that scan mode and oversampling won't work together. Is there a way around it I only require oversampling for one channel for the other it is not required and I am also happy to read it in blocking mode if that avoids the scan mode.

RE: SAADC Noise issue

Kenneth — Sat, 13 Feb 2021 13:33:14 GMT

The Tacq should be set depending on the source resistance you are measuring:
https://infocenter.nordicsemi.com/topic/com.nordic.infocenter.nrf52832.ps.v1.1/saadc.html#concept_qh3_spp_qr

However there are some issues when using short Tacq:
https://infocenter.nordicsemi.com/topic/errata_nRF52832_Rev2/ERR/nRF52832/Rev2/latest/anomaly_832_74.html

https://infocenter.nordicsemi.com/topic/errata_nRF52832_Rev2/ERR/nRF52832/Rev2/latest/anomaly_832_86.html

https://infocenter.nordicsemi.com/topic/errata_nRF52832_Rev2/ERR/nRF52832/Rev2/latest/anomaly_832_150.html

https://infocenter.nordicsemi.com/topic/errata_nRF52832_Rev2/ERR/nRF52832/Rev2/latest/anomaly_832_178.html

https://infocenter.nordicsemi.com/topic/errata_nRF52832_Rev2/ERR/nRF52832/Rev2/latest/anomaly_832_212.html

So typically using 5us (or higher) is preferred.

RE: SAADC Noise issue

KellyJust — Sat, 13 Feb 2021 12:50:08 GMT

Thanks and Tacq can be set at a minimum of 3us right and thus 1khz should be possible?

RE: SAADC Noise issue

Kenneth — Sat, 13 Feb 2021 12:44:07 GMT

The max sampling rate will likely be 1/((TOVERSAMPLE*Tacq)+(interrupt delay+processing)+(possible higher level interrupt that may cause the interrupt to be delayed)).

So presuming 1/((256 * 10us) + (~50) + (~150)) = 362Hz.

RE: SAADC Noise issue

KellyJust — Fri, 12 Feb 2021 16:12:52 GMT

Thanks, Kenneth oversampling seems to be working. Can let me know what is the maximum sampling rate that can be achieved to get averaged value when oversampling at 256?

RE: SAADC Noise issue

Kenneth — Thu, 11 Feb 2021 22:55:03 GMT

Hi,

There are for instance thermal noise that will impact the measurement, so for a single sample you will not get a fixed value no. Though thermal noise for instance is random, so if you enable oversampling you will be able to remove much of the noise because it will even out.

Kenneth

RE: SAADC Noise issue

KellyJust — Thu, 11 Feb 2021 18:09:13 GMT

Here are the results with 2xAA batteries as VDD source and a single AA battery connected directly between analog pin and ground. Still quite a bit of variation. Have even tried with gain 1_6. Is this deviation normal, does the adc ever give out a single non changing output?

RE: SAADC Noise issue

Kenneth — Thu, 11 Feb 2021 11:56:52 GMT

Try to connect a AA battery directly to the analog input, this should be a very low noise source. For VDD you can use 2xAA batteries for test.

Kenneth

RE: SAADC Noise issue

KellyJust — Thu, 11 Feb 2021 11:11:53 GMT

I am using the VDD pin from the NRF52DK (powered by USB from laptop) to power and gnd my custom NRF52832 circuit. Don't unfortunately have access to a scope to measure the supply wave form.

RE: SAADC Noise issue

awneil — Thu, 11 Feb 2021 00:32:14 GMT

So that's roughly a 2% variation.

So how stable is your supply, and how good is your grounding?