Problem with ADC precision

RE: Problem with ADC precision

awneil — Sat, 29 Jun 2019 09:27:19 GMT

You should be able to avoid the int32 overflow by casting to 64-bit ...

Or, as it's all compile-time constants, just work it out before hand.

RE: Problem with ADC precision

p_t — Sat, 29 Jun 2019 07:32:04 GMT

In here multiplication was the reason which overflowed the int32, thats why i suggested doing division first . but thats true that if result is int type then it will lose its fraction part. Thanks for the correction.

RE: Problem with ADC precision

awneil — Thu, 27 Jun 2019 13:30:54 GMT

[quote userid="75788" url="~/f/nordic-q-a/49054/problem-with-adc-precision/195067"](600000/16384) * 6 = 219.7265625[/quote]

Louis Hatte: remember that 'C' will do this in integer arithmetic

So it will first calculate 600000/16384 = 36 and then multiply by 6 to give 192 - which is a significant error!

With integer maths, always do the multiplication first;

ie, 600000 * 6 = 3600000 and then dividing by 16384 gives 219 - which is much closer.

😃

Of course, you have to be careful of overflow ...

You might want to note this thread about using the FPU:

https://devzone.nordicsemi.com/f/nordic-q-a/49062/fpu-power-consumption

RE: Problem with ADC precision

Louis Hatte — Thu, 27 Jun 2019 13:16:15 GMT

Thank for the trick p_t It works now with 14BIT resolution.
Yes haakonsh, my source wasn't really 1.00V but 0.940V

RE: Problem with ADC precision

haakonsh — Thu, 27 Jun 2019 11:32:17 GMT

Are you using multiple SAADC channels?

RE: Problem with ADC precision

haakonsh — Thu, 27 Jun 2019 11:03:36 GMT

My guess is that you've probably not calibrated the offset, or your 1V source is not really 1.00V.

RE: Problem with ADC precision

awneil — Thu, 27 Jun 2019 09:28:22 GMT

[quote userid="80791" url="~/f/nordic-q-a/49054/problem-with-adc-precision"]When we are generating a 1V voltage[/quote]

How are you generating it?

How precise is it?

[quote userid="80791" url="~/f/nordic-q-a/49054/problem-with-adc-precision"]We get values around 940 000 µV instead of 1 000 000µV.[/quote]

Assuming your "1V" really is 1 000 000 uV, that's an error of 60 000 uV - which is 6%

In the nrf52840 SAADC electrical specs, the gain error is 3-4%

So it doesn't seem that you're wildly off ?

https://infocenter.nordicsemi.com/topic/ps_nrf52840/saadc.html?cp=3_0_0_5_22_9#unique_1950216936

RE: Problem with ADC precision

p_t — Thu, 27 Jun 2019 09:24:23 GMT

(600000/16384) * 6 = 219.7265625

try,

adc_uv = adc_value * 219.7265625;

or make adc_uv unsigned32

shouldnt get negative values after that