TM4C123GE6PM: ADC triggered by timer

Dear Bob, yo were right about the ADC interrupt, I appreciate a lot the example, actually I wanted to make a simple example reading the adc and then move forward to the uDMA transfer since I do not understand a lot how it works. But the example make it much easy.

Dear cb1_mobile

First thanks a lot for the advices. I have check and fixed each the points you mentioned. Now it is working (I think) the main problem lied in the ADC Interrupt (rookie mistake), I set a breakpoint to watch the variable and it update the value correctly.

But since I need it goes to 350kHz I am toggling a pin to measure it in the oscilloscope to test, two issues shows up

a. when I am debugging the frecuency falls to ~10Hz, perhaps is normal, make sense to me that the refresh interval of the CCS is in milliseconds and it affects. When I stop the debug session the frequency is "normal" and goes to (b).

b. The frecuency goes to 388kHz instead of 350kHz, could be that I am not toggling the pin properly?

Here is the code with the changes

uint32_t muestra[1];

void ADC0IntHandler(void);

void main(void)
{
uint32_t ui32Period;

SysCtlPeripheralReset(SYSCTL_PERIPH_ADC0);
SysCtlPeripheralReset(SYSCTL_PERIPH_TIMER1);

SysCtlClockSet(SYSCTL_SYSDIV_2_5 | SYSCTL_USE_PLL | SYSCTL_OSC_MAIN |
SYSCTL_XTAL_16MHZ);

SysCtlPeripheralEnable(SYSCTL_PERIPH_GPIOF);
GPIOPinTypeGPIOOutput(GPIO_PORTF_BASE, GPIO_PIN_1|GPIO_PIN_2|GPIO_PIN_3);
SysCtlPeripheralEnable(SYSCTL_PERIPH_TIMER1);

// ***Timer Initialization***

TimerConfigure(TIMER1_BASE, TIMER_CFG_A_PERIODIC);
ui32Period = (SysCtlClockGet()/350000) /2 ;
TimerLoadSet(TIMER1_BASE, TIMER_A, ui32Period -1);
TimerEnable(TIMER1_BASE, TIMER_A);
TimerControlTrigger(TIMER1_BASE, TIMER_A, true);

// ***ADC Initialization***

SysCtlPeripheralEnable(SYSCTL_PERIPH_ADC0); // Enable clock to the ADC module 0
SysCtlDelay(10);
SysCtlPeripheralEnable(SYSCTL_PERIPH_GPIOE);
GPIOPinTypeADC(GPIO_PORTE_BASE, GPIO_PIN_3); // AIN0

IntDisable(INT_ADC0SS3);
ADCIntDisable(ADC0_BASE, 3);
ADCSequenceDisable(ADC0_BASE, 3);

ADCSequenceConfigure(ADC0_BASE, 3, ADC_TRIGGER_TIMER, 0);

ADCSequenceStepConfigure(ADC0_BASE, 3, 0, ADC_CTL_CH0 | ADC_CTL_IE |
ADC_CTL_END);

ADCSequenceEnable(ADC0_BASE, 3);
ADCIntClear(ADC0_BASE, 3);
ADCIntEnable(ADC0_BASE, 3);
IntEnable(INT_ADC0SS3);

IntMasterEnable(); // Enable processor interrupts

while(1)
{
}
}

void ADC0IntHandler(void) {

// Clear the interrupt status flag.
ADCIntClear(ADC0_BASE, 3);

ADCSequenceDataGet(ADC0_BASE, 3, muestra);

if(GPIOPinRead(GPIO_PORTF_BASE, GPIO_PIN_1))
{
GPIOPinWrite(GPIO_PORTF_BASE, GPIO_PIN_1, 0);
}
else
{
GPIOPinWrite(GPIO_PORTF_BASE, GPIO_PIN_1, 2);
}
}

Greetings and thank you.

Israel Rebolledo said:

The frequency goes to 388kHz instead of 350kHz, could be that I am not toggling the pin properly?

That's both a valuable observation - AND - a strange one!     When you "impose a toggle function" - it is expected that such would, "Add Execution Time" - thus 'LOWERING'  your frequency - NOT 'INCREASING' it - don't  you agree?     Is it possible that you've "Reversed your reporting results" - and that "Adding the GPIO Toggle" - has indeed yielded the (lower) frequency?     A faster means of "GPIO Toggle" is via the use of an 'X-OR' family Operator...

Beyond that - the clock source for your timer - (to me) is unspecified.    I usually avoid the internal (not especially precise) oscillator.   (love how marketing 'over-promotes!')     That difference in frequency is around 11% - you'll have to check (if you employed the internal oscillator) to see if that 11% is "in range."

I've undergone dental surgery earlier this morning (not that I'm normally - less loopy) - but noted your, "Divide of SysClock by 350K - and then (again) by 2.    Perhaps - due to drug influence - I find that "divide by 2" suspect...    (meaning your frequency should be even higher!)

Has your scope been recently calibrated - and as I've written (often) here - have you removed the scope's Ground Probe - so that a proper HF Measurement (minus over & under-shoot) can be achieved?   (you attach a very short wire to the probe's grounded "barrel" - which greatly reduces such over/under-shoot.)

Greetings,

Returns the MASTER - we've "Sent out the hounds" - Two of Five returned (frost-bitten paws)!

Three of your (neat) points I'll grant you - yet the "X-OR" toggle - for sure - executes faster... (should that be an objective - otherwise - your 4 'function detection' explains why (you) receive the (proper) BIG Bucks...

Note that your 'frozen tundra' yielded ... 77°F (best kind of degrees) - followed by ... 31°F - the (very) next day. Hard to find (anyone) not made Sick. OH Canada - indeed!

Agree with cb1_mobile, those are total punchlines. I did what you recommend and the frequency has no issues now. The divide by 2 was also wrong, I left that division from a workbook's example, it was needed to get 50 per cent of the interrupt.

I will check the clock source instruction, although I have not seen it on the board examples that is why I relied on the internal oscillator. Many thanks!

If you employ the "internal" oscillator - recognize that "precision" may be 'more in the (accepting) eye of marketing' - than the cold stare of a "precision seeking" Client-User...