BluePill PWM Generation Example Code Not Working #891
-
|
Hi! #define PWM_CHANNEL 2
#define PWM_FREQ_HZ 50
#define PWM_MS_PER_PERIOD (1000/PWM_FREQ_HZ)
#include "PwmThread.hpp"
#include <modm/board.hpp>
constexpr uint16_t overflow = 0xffff;
constexpr uint16_t minPwm = (overflow / PWM_MS_PER_PERIOD) * 1.f; // 1 ms pulse width
constexpr uint16_t targetPwm = (overflow / PWM_MS_PER_PERIOD) * 1.5f; // 1.5 ms pulse width
constexpr uint16_t maxPwm = (overflow / PWM_MS_PER_PERIOD) * 2.f; // 2 ms pulse width
bool PwmThread::run()
{
PT_BEGIN();
MODM_LOG_INFO << "PWM_FREQ_HZ: " << PWM_FREQ_HZ << modm::endl;
MODM_LOG_INFO << "PWM_MS_PER_PERIOD: " << PWM_MS_PER_PERIOD << modm::endl;
// Code used to test that we are in the correct GPIO
if(test_gpio == true)
{
// TEST GPIO PIN
GpioB7::setOutput();
GpioB7::set();
}
// Code used to initialize PWM
else
{
// TEST PWM
pwmInit();
}
while (true)
{
if(test_gpio == false)
{
// TEST PWM
Timer4::setCompareValue(PWM_CHANNEL, targetPwm);
}
PT_YIELD();
}
PT_END();
}
void PwmThread::pwmInit()
{
Timer4::connect<GpioB7::Ch2>();
Timer4::enable();
Timer4::setMode(Timer4::Mode::UpCounter);
// 170 MHz / 52 / 2^16 = approx. 50 Hz -> NUCLEO G474RE
// 72Mhz / X / 2^16 = approx. 50 Hz -> Blue Pill
// X = 72MHz/(50Hz*2^16) = 21
Timer4::setPrescaler(21);
Timer4::setOverflow(overflow);
Timer4::configureOutputChannel(PWM_CHANNEL, Timer4::OutputCompareMode::Pwm, 0);
Timer4::applyAndReset();
Timer4::start();
Timer4::enableOutput();
}
uint16_t PwmThread::getValue()
{
return Timer4::getValue();
}
In the oscilloscope I just get a flat line. I even started doubting about my own existence so that is why you can find the private variable I adapted the prescaler value so it generates a 50Hz frequency in the BluePill too. However, this is the result I get: I managed to get my hands on a Nucleo-G474RE and with minimal changes (commenting the RTT lines, changing the LED name and adjusting the project.xml accordingly) I managed to get a PWM with the same code: I call the I checked the Datasheet and the Reference Manual for the STM32F103 but I was not able to find an obvious answer for my mistake. Any idea what I am doing wrong? Thanks! |
Beta Was this translation helpful? Give feedback.
Replies: 6 comments
-
|
cc @rleh @chris-durand have more experience in this area of the HAL. |
Beta Was this translation helpful? Give feedback.
-
|
It was totally non-obvious for me at first but I think I have found it after digging through the reference manual and the modm GPIO source code. My first suspicion was right that this has to do with the odd F1 pin remapping design. On all STM32 devices except F1 there is only one alternate function mode for both inputs and outputs. The pin direction is controlled by the peripheral. On STM32F1 you have to manually select whether your alternate function pin is an input or output: For input alternate functions standard input mode has to be selected. In case of the timer the modm API has no way to tell which one you want by writing The modm API will only set the pin to alternate function output in case it already has been configured as output before: (The code refers to the bits from the reference manual shown above). If I am right putting Another thing seems odd. The modm API has |
Beta Was this translation helpful? Give feedback.
-
|
Wow, It worked. Thank you so much! |
Beta Was this translation helpful? Give feedback.
-
The problem is that you can also use the timer channel in input capture mode and thus requires input. I think the solution is to document the Timer API properly, because there are a number of other pitfalls like forgetting to |
Beta Was this translation helpful? Give feedback.
-
|
ok! then for the time being I could just prepare a PWM modm example for the BluePill and adding a comment linking to this issue in the |
Beta Was this translation helpful? Give feedback.
-
|
Done! #897 |
Beta Was this translation helpful? Give feedback.
It was totally non-obvious for me at first but I think I have found it after digging through the reference manual and the modm GPIO source code. My first suspicion was right that this has to do with the odd F1 pin remapping design.
On all STM32 devices except F1 there is only one alternate function mode for both inputs and outputs. The pin direction is controlled by the peripheral.
On STM32F1 you have to manually select whether your alternate function pin is an input or output:
For input alternate functions standard input mode has to be selected.
In case of the timer the modm API has no way to tell which one you want by writing
Timer4::connect<GpioB7::Ch2>();since the timer pin can be e…