Random behaviour while using ADC on STM32f030cc

[aatif shaikh]

Hello.

I'm using STM32f030CC in which we have to use 2 ADC channels.

Following is the Initialization part.

  ADC_InitTypeDef stADC_InitStructure;
  GPIO_InitTypeDef GPIO_InitStructure;
 
  RCC_AHBPeriphClockCmd(RCC_AHBPeriph_GPIOB , ENABLE);
	
	
	 /********************* GPIO INITIALIZATION***************/
	GPIO_InitStructure.GPIO_Pin = GPIO_Pin_1 | GPIO_Pin_0;
  GPIO_InitStructure.GPIO_Mode = GPIO_Mode_AN;
  GPIO_InitStructure.GPIO_Speed = GPIO_Speed_2MHz;
  GPIO_InitStructure.GPIO_OType = GPIO_OType_PP;
  GPIO_InitStructure.GPIO_PuPd = GPIO_PuPd_NOPULL;
  GPIO_Init(GPIOB, &GPIO_InitStructure); 
  /********************* ADC1 INITIALIZATION***************/
  RCC_APB2PeriphClockCmd(RCC_APB2Periph_ADC1, ENABLE);
  /*ADC Deinit*/
  ADC_DeInit(ADC1);
 
  /*ADC Structure Initialization*/
  ADC_StructInit(&stADC_InitStructure);
 
  stADC_InitStructure.ADC_Resolution = ADC_Resolution_12b;
  stADC_InitStructure.ADC_ContinuousConvMode = DISABLE;;
	stADC_InitStructure.ADC_ExternalTrigConvEdge = ADC_ExternalTrigConvEdge_None;
  stADC_InitStructure.ADC_DataAlign = ADC_DataAlign_Right;
	stADC_InitStructure.ADC_ScanDirection=ADC_ScanDirection_Upward;
  ADC_Init(ADC1, &stADC_InitStructure);
 
  
  /* Enable the ADC peripheral */
  ADC_Cmd(ADC1, ENABLE);    
	
  
  /* Wait the ADRDY flag */
  while(!ADC_GetFlagStatus(ADC1, ADC_FLAG_ADRDY)); 
  //while(ADC_GetCalibrationStatus(ADC1) == SET);

Now in the phantom loop we're continuously sampling the ADC channels sequentially as follows.

if(uiSamplingCount <= NO_OF_SAMPLES)
{
    val1 = fnA();
    val2 = fnB();
  uiSamplingCount++;
}
else
{ 
    uiSamplingCount=0;
   /* Averaging the values along with calibration */
    /* Print the values */
}

Definition for the above functions are as follows:

void fnA(void)
{
  /* Configure channel   */
  ADC_ChannelConfig(ADC1, ADC_Channel_9, ADC_SampleTime_55_5Cycles);
 
  /* Start the conversion   */
  ADC_StartOfConversion(ADC1);
 
  /* Wait until conversion completion   */
  while(ADC_GetFlagStatus(ADC1, ADC_FLAG_EOC) == RESET);
 
  /* Get the conversion value   */
  return ADC_GetConversionValue(ADC1);
}
void fnB(void)
{
  /* Configure channel   */
  ADC_ChannelConfig(ADC1, ADC_Channel_8, ADC_SampleTime_55_5Cycles);
 
  /* Start the conversion   */
  ADC_StartOfConversion(ADC1);
 
  /* Wait until conversion completion   */
  while(ADC_GetFlagStatus(ADC1, ADC_FLAG_EOC) == RESET);
 
  /* Get the conversion value   */
  return ADC_GetConversionValue(ADC1);
}

Now the issue is we're getting offset values when we use the above sequence.

Second test case:

We're reading only individual channel value and printing the same.

Output is proper.

Third test case:

fnB() is executed first and then fnA() ie Channel 8 first and then Channel 9.

In this case after 5 to 6 loops it stats giving proper values.

Fourth test case:

At a time we're either taking 1000 samples from Channel 8 or 1000 samples from Channel 9 with a delay of 4 secs in between them.

My question is, Is there a dependency between the channels we're using ?

Is there an issue of some settling time in ADC so some error in the sequence that we're missing?

Regards,

Aatif.

@Community member​ 

@Community member​ 

@Community member​ 

@Pavel A.​ 

@henry.dick​ 

[AvaTar]

I don't understand why you initialize the ADC each time before taking a sample. This is waste of time and code space.

Ideally, you configure two channels at once including the calibration, and then trigger + read in a loop.

> My question is, Is there a dependency between the channels we're using ?

This depends.

The sampling time must correlate to the ADC input circuitry. In other words, the input must be able to fully charge/discharge the S&H capacitor in the configured sample time. Otherwise, you will experience a "bleeding" effect, i.e. channel values depend on the value (S&H charge) of the previous channel.

So, if there is any dependency, it is in your interface hardware and/or ADC configuration.

[Tesla DeLorean]

Can we please stop with this habit of calling out forum members in initial posts.

[droid nk]

@Community member​ I guess this is done maybe because of the fact that the reply rate on ST's forum is less compared to other forums. Maybe. 🙂

[droid nk]

@Community member​ 

>I don't understand why you initialize the ADC each time before taking a sample.

I guess the initialization is done only once. By initialization did you mean this:

ADC_ChannelConfig(ADC1, ADC_Channel_9, ADC_SampleTime_55_5Cycles);

[Tesla DeLorean]

If providing support paid better I might care more?

I review things here, it's not necessary to flag posts

[AvaTar]

I have really trouble finding posts I viewed and participated hours or days ago.

The sorting criteria of the forum's "Latest" list is somehow a mystery, and there is no personalized view.

Up to now, I would not say the forum platform changed for the better.

> I guess the initialization is done only once. By initialization did you mean this: ...

Even if you call the fnA() and fnB() function only once, one of them is basically redundant.

You can configure the two pins and two channels at once.

[aatif shaikh]

Hello.

But if we don't write ADC_ChannelConfig(ADC1, ADC_Channel_8, ADC_SampleTime_55_5Cycles); and

ADC_ChannelConfig(ADC1, ADC_Channel_9, ADC_SampleTime_55_5Cycles); every time how can we known the return value after conversion is of which channel?

Regards,

Aatif.

[AvaTar]

I highly suggest to read the (whole) ADC section of the Reference Manual. It will answer your questions.