1 ACCEPTED SOLUTION
Accepted Solutions
Hi Jan,
I have already tried you idea, but I gave it another go. Like the alternative using PWM1 it did produce one pulse, but when I did another StartPWM it did not produce a pulse it just stayed low.
I then took a snap shot of the Timer registers pre the start function both on the first time thru' then the second and compared them.
They were identical with the exception of the BTDR MOE bit Master output enable, being set. Resetting this bit prior to the second run made no difference.
So we had the timer in an identical state as read by STM32CubeIDE SFR monitor yet behaving differently from an intial run to repeated runs. Hmmm!! I thought to myself. What to do?
I made a copy of the timer initialisation function and then placed it before the StartPWM function and voila it works. I then used a process of elimination to determine what part of the initialisation function "did the job."
Turns out the code I added to reverse the polarity of the complementary output must be executed before each StartPWM:
FRAGMENT OF TIM 1 INIT CODE
sBreakDeadTimeConfig.Break2Polarity = TIM_BREAK2POLARITY_HIGH;
sBreakDeadTimeConfig.Break2Filter = 0;
sBreakDeadTimeConfig.AutomaticOutput = TIM_AUTOMATICOUTPUT_DISABLE;
if (HAL_TIMEx_ConfigBreakDeadTime(&htim1, &sBreakDeadTimeConfig) != HAL_OK)
{
Error_Handler();
}
/* Set the complementary output for Ch 3 active Polarity High*/
htim1.Instance->CCER = 0xd00;
HAL_TIM_MspPostInit(&htim1);
HAL_NVIC_SetPriority(TIM1_UP_IRQn, 5, 0);
HAL_NVIC_EnableIRQ(TIM1_UP_IRQn);
So I placed this before the StartPWM function and voila, it works.
PULSE GENERATION FUNCTION
/*
*****************************************************************************
* PENDULUM IMPULSE GENERATOR *
*****************************************************************************
*/
__inline void impulseGen(const float pulseWidth_ms)
{
uint16_t pulseWidthCount;
pulseWidthCount = (uint16_t)(ms_to_count_tim1 * pulseWidth_ms);
MX_TIM1_Init();
htim1.Instance->ARR = (uint32_t)(pulseWidthCount + 1);
htim1.Instance->CCR3 = 1; /* pulseWidthCount; */
/* Set the complementary OP polarity for ch 3
* will not do repeat runs correctly without this */
htim1.Instance->CCER = 0xd00;
HAL_TIM_PWM_Start(&htim1, TIM_CHANNEL_3);
triggerLED_Start(200, IMP_LED_BLUE, false);
}
The observant programmer looking at the Timer registers would have noted that this value 0xd00 was set in the register before each StartPWM command, so why do I have to set it before every start as it appears to be already set from initialisation. I also inspected the memory location with the memory browser and found the value 0xd00 was set prior to each start command.
Could this be a silicon glitch? The register has the correct value, but somehow it ignores it.
Regards
Rob
Timer 1 SFR's
Initial Run
Repeat Runs
Hi Jan,
I have already tried you idea, but I gave it another go. Like the alternative using PWM1 it did produce one pulse, but when I did another StartPWM it did not produce a pulse it just stayed low.
I then took a snap shot of the Timer registers pre the start function both on the first time thru' then the second and compared them.
They were identical with the exception of the BTDR MOE bit Master output enable, being set. Resetting this bit prior to the second run made no difference.
So we had the timer in an identical state as read by STM32CubeIDE SFR monitor yet behaving differently from an intial run to repeated runs. Hmmm!! I thought to myself. What to do?
I made a copy of the timer initialisation function and then placed it before the StartPWM function and voila it works. I then used a process of elimination to determine what part of the initialisation function "did the job."
Turns out the code I added to reverse the polarity of the complementary output must be executed before each StartPWM:
FRAGMENT OF TIM 1 INIT CODE
sBreakDeadTimeConfig.Break2Polarity = TIM_BREAK2POLARITY_HIGH;
sBreakDeadTimeConfig.Break2Filter = 0;
sBreakDeadTimeConfig.AutomaticOutput = TIM_AUTOMATICOUTPUT_DISABLE;
if (HAL_TIMEx_ConfigBreakDeadTime(&htim1, &sBreakDeadTimeConfig) != HAL_OK)
{
Error_Handler();
}
/* Set the complementary output for Ch 3 active Polarity High*/
htim1.Instance->CCER = 0xd00;
HAL_TIM_MspPostInit(&htim1);
HAL_NVIC_SetPriority(TIM1_UP_IRQn, 5, 0);
HAL_NVIC_EnableIRQ(TIM1_UP_IRQn);
So I placed this before the StartPWM function and voila, it works.
PULSE GENERATION FUNCTION
/*
*****************************************************************************
* PENDULUM IMPULSE GENERATOR *
*****************************************************************************
*/
__inline void impulseGen(const float pulseWidth_ms)
{
uint16_t pulseWidthCount;
pulseWidthCount = (uint16_t)(ms_to_count_tim1 * pulseWidth_ms);
MX_TIM1_Init();
htim1.Instance->ARR = (uint32_t)(pulseWidthCount + 1);
htim1.Instance->CCR3 = 1; /* pulseWidthCount; */
/* Set the complementary OP polarity for ch 3
* will not do repeat runs correctly without this */
htim1.Instance->CCER = 0xd00;
HAL_TIM_PWM_Start(&htim1, TIM_CHANNEL_3);
triggerLED_Start(200, IMP_LED_BLUE, false);
}
The observant programmer looking at the Timer registers would have noted that this value 0xd00 was set in the register before each StartPWM command, so why do I have to set it before every start as it appears to be already set from initialisation. I also inspected the memory location with the memory browser and found the value 0xd00 was set prior to each start command.
Could this be a silicon glitch? The register has the correct value, but somehow it ignores it.
Regards
Rob
Timer 1 SFR's
Initial Run
Repeat Runs
Hi
I did some more checking and found that I was getting a "runt" pulse at the start of the sequence viz:
The cause of this pulse is the TIM_CCxChannelCmd which sets the CCER register at line 12.
void TIM_CCxChannelCmd(TIM_TypeDef *TIMx, uint32_t Channel, uint32_t ChannelState)
{
uint32_t tmp;
/* Check the parameters */
assert_param(IS_TIM_CC1_INSTANCE(TIMx));
assert_param(IS_TIM_CHANNELS(Channel));
tmp = TIM_CCER_CC1E << (Channel & 0x1FU); /* 0x1FU = 31 bits max shift */
/* Reset the CCxE Bit */
// TIMx->CCER &= ~tmp; /* debug rjg */
/* Set or reset the CCxE Bit */
TIMx->CCER |= (uint32_t)(ChannelState << (Channel & 0x1FU)); /* 0x1FU = 31 bits max shift */
}Commenting out this line removed the runt pulse:
The final code I ended up with is attached.
Of course modifying HAL might create some nasty side effects for other timers and configs, which I will have to test for in the full system.
