C28335 ADC conversion fails

Hi Johan,

Johan W���lger said:

Is this a normal behavior that I should just handle, or if I should put more effort into investigating the cause?

This is to get around the ADC first sample issue that exist in some of the C2000 MCUs, such as the F28027. 
According to the Errata:

Other than this case you should not face any issues!

Johan W���lger said:

In my project the AD conversion seems to fail once in a while (between 15 minutes and an hour between failures).

Can you describe in detail how you're able to grab this ADC result? Using graphs?

Regards,

Gautam

Hi Gautam,

Thank you for your reply.

My interpretation of your answer is that this sporadic failure of AD conversion is not to be considered a normal behavior.

I am working on a F28335, so I guess the problem you are referring to for the F28027 should not apply. Also, the problem I am having is not related to the converted values or the first value. This appears to occur spontaneously, and what happens is that 1ms after I have started the AD conversion, the ISR still has not been executed.

This is what my ADC code looks like.

In this case CPU_FRQ_150MHZ == 1. I trigger the conversion by calling ADCDrv_StartADC, wait for 1ms and then check if the conversion is done by calling boADCDrv_ADCReady. When the problem occurs, it looks like the ISR hasn't been executed, because boADCDrv_ADCReady returns false.

/****************************************************************
* Private defines
****************************************************************/
#if (CPU_FRQ_150MHZ)     // Default - 150 MHz SYSCLKOUT
  #define ADC_MODCLK 0x3 // HSPCLK = SYSCLKOUT/2*ADC_MODCLK2 = 150/(2*3)   = 25.0 MHz
#endif
#if (CPU_FRQ_100MHZ)
  #define ADC_MODCLK 0x2 // HSPCLK = SYSCLKOUT/2*ADC_MODCLK2 = 100/(2*2)   = 25.0 MHz
#endif
#define ADC_CKPS   0x1   // ADC module clock = HSPCLK/2*ADC_CKPS   = 25.0MHz/(1*2) = 12.5MHz
#define ADC_SHCLK  0xf   // S/H width in ADC module periods                        = 16 ADC clocks

/****************************************************************
* Private function prototypes
****************************************************************/ 
interrupt void ADCDrv_ISR(void);

/****************************************************************
* Private global variables
****************************************************************/	
static uint16_t m_boChannelsUpdated;

/************************************************************************************
* Initialize the ADC
*************************************************************************************/
void ADCDrv_InitADC(void)
{	
	/* Specific clock setting */
   	EALLOW;
    SysCtrlRegs.HISPCP.all = ADC_MODCLK;	/* HSPCLK = SYSCLKOUT/ADC_MODCLK */
    EDIS;
   
	/* Init the ADC */
	InitAdc();  
	
    /* ADC timing setup */
    AdcRegs.ADCTRL1.bit.ACQ_PS = ADC_SHCLK;
    AdcRegs.ADCTRL3.bit.ADCCLKPS = ADC_CKPS;
      	
	/* Configure ADC */
	AdcRegs.ADCTRL1.bit.SEQ_CASC = 0x1; 	// Setup cascaded sequencer mode
   	AdcRegs.ADCMAXCONV.all = 0x000f;        // Setup 16 conv's
   	
	AdcRegs.ADCCHSELSEQ1.bit.CONV00 = 0x0;  // Setup conv from ADCINA0  
	AdcRegs.ADCCHSELSEQ1.bit.CONV01 = 0x1;  // Setup conv from ADCINA1
	AdcRegs.ADCCHSELSEQ1.bit.CONV02 = 0x2;  // Setup conv from ADCINA2
	AdcRegs.ADCCHSELSEQ1.bit.CONV03 = 0x3;  // Setup conv from ADCINA3
	AdcRegs.ADCCHSELSEQ2.bit.CONV04 = 0x4;  // Setup conv from ADCINA4
	AdcRegs.ADCCHSELSEQ2.bit.CONV05 = 0x5;  // Setup conv from ADCINA5
	AdcRegs.ADCCHSELSEQ2.bit.CONV06 = 0x6;  // Setup conv from ADCINA6
	AdcRegs.ADCCHSELSEQ2.bit.CONV07 = 0x7;  // Setup conv from ADCINA7
	AdcRegs.ADCCHSELSEQ3.bit.CONV08 = 0x8;  // Setup conv from ADCINB0
	AdcRegs.ADCCHSELSEQ3.bit.CONV09 = 0x9;  // Setup conv from ADCINB1
	AdcRegs.ADCCHSELSEQ3.bit.CONV10 = 0xA;  // Setup conv from ADCINB2
	AdcRegs.ADCCHSELSEQ3.bit.CONV11 = 0xB;  // Setup conv from ADCINB3
	AdcRegs.ADCCHSELSEQ4.bit.CONV12 = 0xC;  // Setup conv from ADCINB4
	AdcRegs.ADCCHSELSEQ4.bit.CONV13 = 0xD;  // Setup conv from ADCINB5
	AdcRegs.ADCCHSELSEQ4.bit.CONV14 = 0xE;  // Setup conv from ADCINB6
	AdcRegs.ADCCHSELSEQ4.bit.CONV15 = 0xF;  // Setup conv from ADCINB7
		
   	AdcRegs.ADCTRL2.bit.INT_ENA_SEQ1 = 1;  	// Enable SEQ1 interrupt (every EOS)
    AdcRegs.ADCTRL2.bit.RST_SEQ1 = 1;       /* Reset SEQ1 */

    /* Interrupt setup */
    EALLOW;
    PieVectTable.ADCINT = &ADCDrv_ISR;
    EDIS;

    IER |= M_INT1;
    PieCtrlRegs.PIEIER1.bit.INTx6 = 1;

    /* Reset channel status */
   	m_boChannelsUpdated = FALSE;
}

/************************************************************************************
* Start the ADC sequence
*
*************************************************************************************/
void ADCDrv_StartADC( void )
{	
    m_boChannelsUpdated = FALSE;        /* Waiting for new sequence */
	AdcRegs.ADCTRL2.bit.SOC_SEQ1 = 1;	/* Enable SOC to start SEQ1 */
}


bool_t boADCDrv_ADCReady( void )
{
    return( m_boChannelsUpdated );
}

/************************************************************************************
* The ISR for ADC
*************************************************************************************/
interrupt void ADCDrv_ISR( void )
{
  	/* Reinitialize for next ADC sequence */
  	AdcRegs.ADCTRL2.bit.RST_SEQ1 = 1;       /* Reset SEQ1 */
  	AdcRegs.ADCST.bit.INT_SEQ1_CLR = 1;     /* Clear INT SEQ1 bit */
  	PieCtrlRegs.PIEACK.all = PIEACK_GROUP1; /* Acknowledge interrupt to PIE   */
	
    m_boChannelsUpdated = TRUE;             /* SAmpled data now available */
}

Best regards,

Johan

Thank you Bill,

This is very useful information. Right now I can not run the system in the debugger because of a problem with booting a slave DSP when the F28335 is running in the debugger. Once that is solved I have a bunch of ideas about what could go wrong. I'm going to start looking at the timing of th RTOS. I wouldn't be surprised if what should be a 1ms period just isn't at the point where the ADC fails.

What I needed most of all was an indication about whether this is a normal behavior that error handling should just deal with, or if there is reason to start investigating the cause.

I will post my findings in this thread as soon as possible.

Best regards,

Johan