STM32F070 ADC Reference voltage problem

[ASz]

Hi Everyone!

I have read a lot of topics, articles about the F070 ADC settings, calibration, etc., but I couldn't solve my issue. I'm using ADC with DMA right now, but the problem is the same with the interrupt and the polling method. In my code I'm using 3 channels (2 simple ADCs and the Vref). The result of the Vref channel is about 3.4V but I measure 3.31V with voltmeter (the other channels have this kind of problem too). Calibration is used by every 4th cycle and I have a 100nF capacitor between the Vcc and the GND and input filter as well.

I hope You can help.

Thanks in advance.

Andras

[T J]

I had the same problem and calibration fixed it with 'F072 and 'F091

are you driving any heavy loads from another pin ? this can affect the internal rail voltage.

[ASz]

No.

Right now I'm using just the ADC pins, but my final project will use 2 PWMs and UART.

The calibration helped me a little bit, but I get a 50-70mV difference between the voltmeter's and ADC's value on the ADC pins too (not just Vref).

[T J]

50-70mV you have a problem.

when did you change the battery in the meter last ? <- this may be your issue, do you have another meter ?

every bit helps improve the figure...

a few years ago, I paid a little extra to get a really good Volt Meter. ita an AMProbe 37XR-A

nothing special, 4 digit and works very well.

anyhow I am reading 1-2mV difference on the analog input pins and 2-3-4mV on the Vdda calculation.

it moves around after the re-calibration takes place, cal factor 41-42...

what sampling speed are you running ?

the highest few are not good... I use the slowest speed.

I beef up the Caps all round.. so on Vdda 1u and 0.1u

lots of 0.1u on the board, if I have a switcher, I use some ferrites and inductors too..

do you have a switcher power supply ?, this can induce some DC offsets when the noise is rectified in the pin "input stage protection diodes", inside the processor..

just guessing of course...

[ASz]

In my final setup the MCU is working from a battery with a voltage regulator but now I'm using a DC-DC regulator to get 3.3V.

Now I replaced that regulator with the battery (It is on 3.6V now) and the ADC says its 3.67V. I've made a value correction with a 70mV substraction and it's seems to be right (according to the other values).

I've tested with another multimeter but it says it's 3.57V so I'm going to get a better multimeter and after that I will check it again.

[T J]

after a power up, you can see the Calibration is off, then after the first cal, all the values seem to settle quite nicely.

(a 4 digit meter is much better than a 3 1/2 digit... )

[ASz]

I have measured with another multimeter but the result is the same :\

My VCC and VCCA are common (without any cap), could it cause any problem?

I have included my main.c

[ASz]

And this is my ADC settings in the CubeMX

[ASz]

..

[T J]

this is not quite the same ...

your code is tiny,

I cant understand why my code is 2000 +lines for a short project,

yours is under 20...

void process_ADCcalibration(void) {
	
 //sprintf(string,"ADC Recal is Progressing\n\r");
 //puts (string);
 
	
// check DMA and ADC have stopped;
/*
		Calibration software procedure
		1. Ensure that ADEN=0
		2. Set ADCAL=1
		3. Wait until ADCAL=0
		4. The calibration factor can be read from bits 6:0 of ADC_DR.
				ADC Calibration code example
				// (1) Ensure that ADEN = 0 
				// (2) Clear ADEN 
				// (3) Launch the calibration by setting ADCAL
				// (4) Wait until ADCAL=0
*/
	if((ADC1->CR & ADC_CR_ADEN) != 0)				 // (1) 
 {
 ADC1->CR &= (uint32_t)(~ADC_CR_ADEN); // (2)
 }
 ADC1->CR |= ADC_CR_ADCAL; 							// (3)
 while((ADC1->CR & ADC_CR_ADCAL) != 0) // (4)
 {
 checkBackgroundServices();
 // For robust implementation, add here time-out management 
 }
 ADC_CalibrationFactor = (ADC1->DR); 	//& 0x7F);
 ADChasBeenRecalibrated = true;
 
 HAL_ADC_Start_DMA(&hadc, (uint32_t *)ADC_DMABuffer, ADC_ChannelCount); //from main, adc-start
		
} 

[T J]

here is one of my boards setup