Having problems with STM32F7 TIM1 PWM Mode 1 Output pins

[anonymous.8]

Posted on March 10, 2018 at 18:41

Hi all. I am using an STM32F767VGT6 in a custom board. I am trying to use TIM1 to generate four independent PWM output pulses on each of its CH1 - CH4 output pins.

I have successfully implemented this elsewhere in my system using one channel only, CH1, on TIM10. My TIM1 Timer init code is almost identical, yet I cannot get the PWM pulses to appear on the PB9, PB11, PB13 and PB14 pins.

If I switch on the TIM1 Update and CC interrupts and poke the pins in the interrupt handlers, I can get the correct waveforms, but I shouldn't have to do that, the TIM1 peripheral should change the pins by itself with no interrupts needed.

So I know the pins, when set as regular GPIO pins, and also the board, all work. The problem is that when I set those pins to be GPIO_AF1_TIM1 and switch off the interrupts which are not needed for any other purpose, I don't get any PWM waveform on any of those CH1 - CH4 pins.

Apart from the fact that TIM1 is an advanced timer with lots of bells and whistles which I don't need, and TIM10 is a regular simpler timer, I can't figure out what I am missing in my TIM1 init code to make it work just like my TIM10 does elsewhere.

I have attached my TIM1 Init code here. Can anyone see what is wrong? Thanks.

[Tesla DeLorean · Accepted solution]

Posted on March 10, 2018 at 18:47

On Advanced TIM (ie 1 and 😎 you must explicitly enable PWM Output (Input)

void TIM_CtrlPWMOutputs(TIM_TypeDef* TIMx, FunctionalState NewState)

{  /* Check the parameters */  assert_param(IS_TIM_LIST4_PERIPH(TIMx));  assert_param(IS_FUNCTIONAL_STATE(NewState));

  if (NewState != DISABLE)

  {    /* Enable the TIM Main Output */    TIMx->BDTR |= TIM_BDTR_MOE;  }  else  {    /* Disable the TIM Main Output */    TIMx->BDTR &= (uint16_t)~TIM_BDTR_MOE;  }}

Also ARR needs to be written as an N-1 value, marks the last value on N for 0 thru N-1

   // Setup 16ms period using the AutoReload register

   TIM1->ARR = 16000 - 1;

[anonymous.8]

Posted on March 10, 2018 at 19:03

Clive, thank you so much. Wow, that was a fast reply too! Your suggestion does indeed solve the problem.

Thanks, David.