Tiva Digital Comparator Problem

Hi Nathan,

In the main code, the ADCSequenceConfigure is on ALWAYS TRIGGER Mode, but the ADC Sequencer is being enabled later. I would make a change to enable the Trigger after the ADC Sequencer is configured and Enabled.

Also once the Digital Comparator is enabled, you may want to clear the Stale Data using ADCComparatorReset function. There is more detail in the datasheet for ADCDCRIC register.

Also you may want to check the Errata for Digital Comparators in TM4C123 series to make sure that it is not what you are seeing.

Amit

Amit,

I'm not quite sure what you mean by enabling the trigger after configuring and enabling the sequence. Isn't the trigger set in my call to configure my sequence? I made the change to reset the comparator and now it will trigger from all events. However only the 2nd, 4th, and 6th comparator interrupts are placed in the correct bit. The 1st interrupt is in the 5th spot, the 5th is in the 1st spot etc. Here is my updated code:

	ADCSequenceDisable( ADC0_BASE, 0 );

	// Set up the ADC trigger to be from the processor (software)
	ADCSequenceConfigure( ADC0_BASE, 0, ADC_TRIGGER_ALWAYS, 0 );

	// Set up the ADC interrupt steps
	ADCSequenceStepConfigure( ADC0_BASE, 0, 0, XOUT_CHANNEL | ADC_CTL_CMP0 );
	ADCSequenceStepConfigure( ADC0_BASE, 0, 1, XOUT_CHANNEL | ADC_CTL_CMP1 );
	ADCSequenceStepConfigure( ADC0_BASE, 0, 2, YOUT_CHANNEL | ADC_CTL_CMP2 );
	ADCSequenceStepConfigure( ADC0_BASE, 0, 3, YOUT_CHANNEL | ADC_CTL_CMP3 );
	ADCSequenceStepConfigure( ADC0_BASE, 0, 4, ZOUT_CHANNEL | ADC_CTL_CMP4 );
	ADCSequenceStepConfigure( ADC0_BASE, 0, 5, ZOUT_CHANNEL | ADC_CTL_CMP5 | ADC_CTL_END);

	// Configure the comparators
	ADCComparatorConfigure( ADC0_BASE, 0, ADC_COMP_INT_LOW_ONCE );
	ADCComparatorConfigure( ADC0_BASE, 1, ADC_COMP_INT_HIGH_ONCE );
	ADCComparatorConfigure( ADC0_BASE, 2, ADC_COMP_INT_LOW_ONCE );
	ADCComparatorConfigure( ADC0_BASE, 3, ADC_COMP_INT_HIGH_ONCE );
	ADCComparatorConfigure( ADC0_BASE, 4, ADC_COMP_INT_LOW_ONCE );
	ADCComparatorConfigure( ADC0_BASE, 5, ADC_COMP_INT_HIGH_ONCE );

	ADCComparatorRegionSet( ADC0_BASE, 0, 1000, 3000 );
	ADCComparatorRegionSet( ADC0_BASE, 1, 1000, 3000 );
	ADCComparatorRegionSet( ADC0_BASE, 2, 1000, 3000 );
	ADCComparatorRegionSet( ADC0_BASE, 3, 1000, 3000 );
	ADCComparatorRegionSet( ADC0_BASE, 4, 1000, 3000 );
	ADCComparatorRegionSet( ADC0_BASE, 5, 1000, 3000 );

	ADCComparatorReset(ADC0_BASE, 0, true, true);
	ADCComparatorReset(ADC0_BASE, 1, true, true);
	ADCComparatorReset(ADC0_BASE, 2, true, true);
	ADCComparatorReset(ADC0_BASE, 3, true, true);
	ADCComparatorReset(ADC0_BASE, 4, true, true);
	ADCComparatorReset(ADC0_BASE, 5, true, true);

	// Enable each ADC sequence
	ADCSequenceEnable( ADC0_BASE, 0 );

	while( true )
	{
		uint32_t intStatus = ADCComparatorIntStatus( ADC0_BASE );

		if( intStatus != 0 )
		{
			ADCComparatorIntClear(ADC0_BASE, intStatus);
		}
	}

Hi Nathan

I was referring to the API Call for enabling the trigger.

This has to be done after the Sequencer is Enabled.

However the issue at hand is when ZOUT is LOW, the Interrupt for Digital Comparator XOUT when LOW is getting set. What about Interrupt for 3rd digital comparator. Does it fire correctly?

Amit

I think i may be seeing the same problem.  I'm trying to use 4 channels and 8 comparators in a single sequence to look for low and high conditions.  

If I use a single channel with two comparators(two steps) in a single sequence I get the results I expects.  I connect the channel to a voltage supply and can consistently cause high and low interrupts.  

As soon as I try to add any additional channels the behavior becomes strange.  If I add two additional steps using two different comparators(same sequence or even a different sequence) on a different channel I lose the expected behavior on both channels.  If I ground my second channel my first channel will no longer see my high or low conditions.  If I tie both channels to my voltage source I can then, once again, cause my high and low conditions to trigger.  If I leave one pin floating and the other tied to my voltage source I can once again cause my high and low interrupts but only at a voltage higher than the voltage I programmed the comparators to trigger on.  

It almost seems as if there is some averaging between channels.  The more channels I add the higher the threshold to trigger my high comparator becomes.  I need to monitor 4 channels separately for both high and lows.  Is there something that would cause this behavior?

For the record I was never able to find a solution to this issue. We ended up changing the design to work around the issue.

Looks like my Launchpad has Channel 3 and Channel 0 shorted somehow(PE0 and PE3).  I decided to use channels 1 & 0(PE2 & PE3) but mistakenly setup PE3 and PE0.  It worked as I intended despite setting up the wrong pin.  After fixing this mistake it still worked.  Looks like PE0 and PE3 are acting as one.  Not sure if it's a flaw in the launchpad in general or just my board.  I programmed our own board which contains a Tiva(slightly different version) and it works correctly.

I did also running into the Digital comparator mistake but that was mentioned in the erratta.  Not explained well but easy to work around.

I am having exactly the same issues with my Tiva TM4c123GH6PM. I have configured all 12 Analog Input Channels. The first 4 analog inputs are fed to SS0 of ADC0 and remaining 8 inputs are fed to SS0 of ADC1. And then they go to their respective Digital Comparators.

Strangely, the Analog Pins are somehow coupled internally. When I change voltage levels on one analog pin, I get interrupts on several pins. I can't find any solution to this problem in the Errata.

I have discovered another unexpected behavior. I am also using PWM on PB6 and PB7. When I start the PWM, The Digital Comparator linked to PD0 starts giving me interrupts every now and then. I used a multimeter to check the voltage and it showed ~ 3V after apprpx. every 1 sec.
My PWM frequency is 10 KHz btw. There are all sorts of strange things going on with my launchpad. Can anybody help?

Hello Zaigam,

There is a 0Ohm Resistor on the board (R9) which connected PB6 to PD0. You would need to remove the same.

Tim, All

I checked the board and there is no short or low impedance path between PE0 and PE3.

Regards

Amit

I call this the setup routine from inside of my ADC handler thread.  The pins are already configured with the TIVA Pin tool.  

The call to enable is:

// Set up the Low Voltage Alarm processor.

EnableLowVoltageAlarm( ADC1_BASE, 32 );

I am using ADC0 to read voltages on 8 different channels.  I wanted the power fail interrupts to be very quick, but not respond to spikes.  Since ADC1 was idle, I use that for the power monitor.

The second parameter to EnableLowVoltageAlarm controls the hardware averaging.   I selected 32 as I am monitoring 3 channels, as well as a dummy channel as required by Errata ADC#04.  So a basic conversion takes 32*4 micro seconds. That lets me respond in 128 us, but still filters out the occasional noise glitch that may occur on the power rails.  

The event handler comes from product code and could be simplified.   Some of the functions are from FreeRTOS.  The following code adds more comments.  Sorry about the lack of indentation.

/** ***************************************************************************
* LVAlarmService
*
* This ISR handles the low volage alarms generated by the digital comparator.
*
**************************************************************************** */

// Silicon Revisions 6 & 7 have bug in DCComparator interrupt. ADC#04
// The bits are shifted by 1.
// If they fix this bug, the following defines must change:

#define APS_LV_FAULT 2
#define BAT_LV_FAULT 4
#define VOUT_LV_FAULT 8

void LVAlarmHandler(void)
{
uint32_t DCStatus;
SourceEvents_t Event;

// Flag used by FreeRTOS to control IRQ Chaining..
// Not relavent to basic operation
long xHigherPriorityTaskWoken = false;

// Get the cause into DCStatus.
// Errata ADC#04 says events are shifted.
// The defines handle the shift for now.
DCStatus = ADCComparatorIntStatus( ADC1_BASE );

//
// Clear all the ADC Digital Comparator interrupts that are set
//
ADCComparatorIntClear( ADC1_BASE, DCStatus );

// check for APS disconnect
if ( DCStatus & APS_LV_FAULT )
{
// Application processing of low power
// If APS fails, turn on Battery immediately.
// use PIN object which abstracts GPIO

EnBattPower.Set();
Event = SourceRemovedFromAps;

// FreeRTOS call to send a message to the handler.
xQueueSendToFrontFromISR ( EventQue,
&Event,
&xHigherPriorityTaskWoken);
}

// check for Battery disconnect
if ( DCStatus & BAT_LV_FAULT )
{
// If Batt fails, turn on APS
EnBattPower.Set();
Event = SourceRemovedFromBat;
xQueueSendToFrontFromISR ( EventQue,
&Event,
&xHigherPriorityTaskWoken);
}

// FreeRTOS method of IRQ chaining.
if( xHigherPriorityTaskWoken == pdTRUE)
{
portYIELD_FROM_ISR(true);
}
}