ADC scan mode problem, STM32F407

[rgrimes]

Posted on June 16, 2014 at 17:35The original post was too long to process during our migration. Please click on the attachment to read the original post.

[Tesla DeLorean]

Posted on June 16, 2014 at 17:57

Not sure I have time to go over this completely, but the most obvious issues are:

tinit.TIM_Period = (84000000 / 553) - 1; // 152 KHz Should be

tinit.TIM_Period = (84000000 / 152000) - 1; // 152 KHz

(number of ticks at 84MHz for a 152KHz period) Stream ONE so IF1, not IF0

void DMA2_Stream1_IRQHandler(void)

{

  if(DMA_GetITStatus(DMA2_Stream1, DMA_IT_HTIF1)) {

    DMA_ClearITPendingBit(DMA2_Stream1, DMA_IT_HTIF1);

    STM_EVAL_LEDOff(LED2);

  }

  if(DMA_GetITStatus(DMA2_Stream1, DMA_IT_TCIF1)) {

    DMA_ClearITPendingBit(DMA2_Stream1, DMA_IT_TCIF1);

    STM_EVAL_LEDOn(LED2);

  }

}

Describes only half the buffer, the point is to describe the whole buffer and have interrupts at each half (HT/TC)

dinit.DMA_BufferSize = SAMPLES_BUFFER_HALF;

Should be

dinit.DMA_BufferSize = SAMPLES_BUFFER;

TIM2 is on APB1 NOT AHB1

// Enable peripheral clocks

  RCC_AHB1PeriphClockCmd(  RCC_AHB1Periph_DMA2  | RCC_AHB1Periph_GPIOA

                         | RCC_AHB1Periph_GPIOC | RCC_AHB1Periph_GPIOF

                         | RCC_APB1Periph_TIM2, ENABLE);

[rgrimes]

Posted on June 16, 2014 at 18:22

Hi Clive1,

1. Yeah, I blew the period - not worth explaining.2. Yeah, missed the correct stream number for the ADC3 DMA ISR.3. Buffer size, yeah.4. And of course, the biggie!  That TIM2 peripheral clock mistake!  One lousy character - though it is more than one bit - and everything is broken!Thanks to you, everything now appears to be working correctly!Best regards,Bob

[rgrimes]

Posted on June 16, 2014 at 18:43

Not sure it matters to my app, but the TIM2 interrupt doesn't appear to be happening. Is it possible to generate interrupts using

TIM_IT_Trigger

? Or am I missing yet something else? Excerpts:

NVIC_InitTypeDef ninit;
TIM_TimeBaseInitTypeDef tinit;
. . .
RCC_APB1PeriphClockCmd(RCC_APB1Periph_TIM2, ENABLE);
. . .
ninit.NVIC_IRQChannel = TIM2_IRQn;
ninit.NVIC_IRQChannelPreemptionPriority = 0;
ninit.NVIC_IRQChannelSubPriority = 0;
ninit.NVIC_IRQChannelCmd = ENABLE;
NVIC_Init(&ninit);
. . .
TIM_TimeBaseStructInit(&tinit);
tinit.TIM_Period = (84000000 / 152000) - 1; // 152 KHz
tinit.TIM_Prescaler = 0;
tinit.TIM_ClockDivision = 0;
tinit.TIM_CounterMode = TIM_CounterMode_Up;
TIM_TimeBaseInit(TIM2, &tinit);
TIM_SelectOutputTrigger(TIM2, TIM_TRGOSource_Update); // ADC_ExternalTrigConv_T2_TRGO
TIM_ITConfig(TIM2, TIM_IT_Trigger, ENABLE);
TIM_Cmd(TIM2, ENABLE);
. . .
void TIM2_IRQHandler(void)
{
if (TIM_GetITStatus(TIM2, TIM_IT_Trigger) != RESET) {
TIM_ClearITPendingBit(TIM2, TIM_IT_Trigger);
STM_EVAL_LEDToggle(LED3);
}
}

[Tesla DeLorean]

Posted on June 16, 2014 at 18:43

If 152 KHz is critical, you could run the processor at 152 MHz rather than 168 MHz, although this would be problematic for using USB.

[Tesla DeLorean]

Posted on June 16, 2014 at 18:49

I don't know, I'd perhaps just use the Update rather than the Trigger, but 152 KHz is getting a bit on the high side to be interrupting unless you have too. I'd probably check the pin was configure right, and then see if can toggle a pin directly with one of the TIM2_CHx pins.

You should also be able to measure the TIM2 frequency indirectly via the HT/TC interrupts. A glance suggests they get called at 1 KHz, so a toggle should give 500 Hz at the LED

[rgrimes]

Posted on June 16, 2014 at 19:38

Yeah, you're right - that is too fast!  I really just wanted to confirm that TIM2 was running, because at the time I wasn't sure what was broken.  So, forget about it.

Thanks for your help!