I've been trying to configure & run hardware timer over SAML21 MCU to generate a 100ms delay i.e. ISR is supposed to hit at every 100ms. But it is observed that after starting the timer ISR is hitting at every 10us and changing the Prescaler & Compare register values isn't creating any difference in the 10us interval. Please review my code and let me know where I'm doing wrong.
I'm trying to configure Timer1(TC1) in 16bit mode, using GCLK_GENERATOR_1 as its clock source running at 8MHz frequency(CPU Main Clock:16MHz). The timer is expected to cause overflow interrupt every 100ms.
TcCount16 *tc_hw1 = NULL; /* Pointer to TC1 hardware registers Initilized later */
void init_timer1(void)
{
struct tc_module tc_inst1;
struct tc_config conf_tc1;
tc_get_config_defaults(&conf_tc1);
conf_tc1.clock_source = GCLK_GENERATOR_1;
conf_tc1.clock_prescaler = TC_CLOCK_PRESCALER_DIV64; /* 8MHz/64 = 125KHz*/
conf_tc1.reload_action = TC_RELOAD_ACTION_GCLK;
conf_tc1.counter_size = TC_COUNTER_SIZE_16BIT;
conf_tc1.count_direction = TC_COUNT_DIRECTION_UP;
conf_tc1.counter_16_bit.value = 0x0000;
/** Rest of the settings are used as defaults **/
while (tc_init(&tc_inst1, TC1, &conf_tc1) != STATUS_OK){
}
tc_set_top_value(&tc_inst1, 12500); /* Set counter compare top value */
/* Enable interrupt & Set Priority */
tc_hw1 = &(tc_inst1.hw->COUNT16); /* Initialize pointer to TC1 hardware register */
tc_hw1->INTENSET.reg |= TC_INTFLAG_OVF; /* Enable Overflow Interrupt */
NVIC_SetPriority(TC1_IRQn, 2);
NVIC_EnableIRQ(TC1_IRQn);
tc_enable(&tc_inst1); /*Start The TIMER*/
}
void TC1_Handler(void)
{
if((tc_hw1->INTFLAG.reg) & (TC_INTFLAG_OVF))
{
port_pin_toggle_output_level(PIN_PB03);
}
system_interrupt_clear_pending(SYSTEM_INTERRUPT_MODULE_TC1);
}
Debugger Information: I can see that the timer register is configured correctly but the COUNT register is not incrementing itself every time I pause to capture the debug info it shows only 0x0000 values.
Please help. Thanks!
I resolved the issue actually it's kind of mandatory to clear the timer OVF bit in the INTFLAG register. So the interrupt handler should've been like this:
void TC1_Handler(void)
{
if((tc_hw1->INTFLAG.reg) & (TC_INTFLAG_OVF))
{
tc_hw1->INTFLAG.reg = TC_INTFLAG_OVF; /*Clears the flag by writing 1 to it*/
port_pin_toggle_output_level(PIN_PB03);
}
system_interrupt_clear_pending(SYSTEM_INTERRUPT_MODULE_TC1); /*Not necessarily needed*/
}
Related
I have code written by someone else (for a dsPIC33FJ128MC706A) - which initializes UART1 & PWMs and uses a busy wait in the U1Rx interrupt to get 14 bytes, and timer1 interrupts to update PWM registers if new data arrives (debugged - it enters the _U1RXInterrupt when data is sent)
I tried modifying it to use DMA instead of busy wait. I changed nothing in the initialization, but called the following function after initialization of UART:
unsigned char received_data1 [0x0f] __attribute__((space(dma)));
unsigned char received_data2 [0x0f] __attribute__((space(dma)));
void dma_init(void)
{
DMA0REQ = 0x000b; // IRQSEL = 0b0001011 (UART1RX))
DMA0PAD = (volatile unsigned int) &U1RXREG;
DMA0STA = __builtin_dmaoffset(received_data1);
DMA0STB = __builtin_dmaoffset(received_data2);
DMA0CNT = 0xe;//15 bytes in
DMA0CON = 0x0002; //continuous ping-pong + post-increment
IFS0bits.DMA0IF = 0;
IEC0bits.DMA0IE = 1;
DMA0CONbits.CHEN = 1;
}
This is based on example 22-10 in the FRM with a few slight changes (UART1 receiver instead of UART2). I can't seem to find other examples that don't use library functions (which I didn't even know existed up until I came across those examples).
(Also, to be consistent with the code I got, I put in _ISR after void instead of making interrupt an attribute, but that didn't change anything)
In the example, there was no interrupt for the UART1 receiver - but according to the FRM, it should be enabled. Made sense to me that I don't need to clear the interrupt flag myself (it would kind-of defeat the purpose of DMA if I needed to use the CPU for it).
And yet, when I send data, the breakpoint on DMA0Interrupt doesn't trigger, but the one on the default interrupt does, and I have PWMIF=1 and U1RXIF=1. I also tried commenting out the default ISR and adding a _U1ErrInterrupt even though examining the flags I didn't notice any error flags being raised, just the normal U1RXIF.
I don't understand why this would happen. I haven't changed the PWM code at all, and the PWM flag isn't raised in the original code when I place a breakpoint in the _U1RxInterrupt. I don't even understand how a DMA set-up could cause a PWM error.
The U1RXIF flag seems to be telling me the DMA didn't handle the incoming byte (I assume it's the DMA that clears the flag), which once again relates to the question "what did I do wrong in this DMA setup"
UART initialization code:
in a general initialization function (called by main())
TRISFbits.TRISF2 = 1; /* U1RX */
TRISFbits.TRISF3 = 0; /* U2TX */
also called by main:
void UART1_init_USB (void)
{
/* Fcy 7.3728 MHZ */
U1MODE = 0x8400; /* U1ATX Transmit status and control register */
U1STA = 0x2040; /* Receve status and control register Transmit disabled*/
U1BRG = 0x000f; /* Baund rate control register 115200 bps */
IPC2bits.U1RXIP = 3; /* Receiver Interrupt Priority bits */
IEC0bits.U1RXIE = 1; /* Enable RS232 interrupts */
IFS0bits.U1RXIF = 0;
}
(comments by original code author)
Every time I enable the timer it instantly activates the Interrupt. No matter how I try to prescale it. only ARR seems to work but 16 bit with 0,5MHz clock gives me ~160ms maneuver.
#define SYSCLK_FREQ 524288
void timer_init(uint16_t detonation_delay_ms);
int main(void){
RCC->APB2ENR = RCC_APB2ENR_TIM22EN;
TIM22->PSC = (SYSCLK_FREQ/1000)-1;
NVIC_EnableIRQ(TIM22_IRQn);
NVIC_SetPriority(TIM22_IRQn,4);
}
/* calling function */
timer_init(65535);
/* calling function */
void timer_init(uint16_t detonation_delay_ms){
TIM22->CR1 &= ~TIM_CR1_CEN;
TIM22->SR=0;
TIM22->ARR = detonation_delay_ms;
TIM22->CR1 |= TIM_CR1_CEN;
TIM22->SR = 0;
}
void TIM22_IRQHandler(void){
TIM22->CR1 &= ~TIM_CR1_CEN;
TIM22->SR=0;
GPIOB->BSRR = GPIO_BSRR_BS_7;
}
I wish that calling function makes the timer tick till the called value in milisec. But no matter how I set it up it ends up with no scaled timer and instant interrupt after calling it.
Correct way to do it?
TIM22->DIER = TIM_DIER_UIE;
TIM22->ARR = 65535-detonation_delay_ms;
TIM22->EGR = TIM_EGR_UG;
TIM22->CR1 |= TIM_CR1_OPM | TIM_CR1_CEN;
TIM22->SR=0;
Do not delay in interrupts
you enable the timer then set the ARR which is wrong - first set ARR and prescaller, then generate the UG event using the EGR register, then enable the timer.
Works like a charm. Just because I got help here will describe for future interested people.
The way to get the interrupt working for the timers is to generate interrupt 'by hand' once. It's okay to do it because you can control what's happening during the interrupt by a single 'if'.
/* TIMER Enable */
RCC->APB2ENR = RCC_APB2ENR_TIM22EN;
I had a problem with the above declaration, dunno why but it wasn't working after declaring some more modules before it. Had to put it on higher on the list. Manual does not say why it happened.
/* Configure TIM22 interrupt */
TIM22->PSC = (SYSCLK_FREQ/1000)-1; /* Prescaler TIMERA 22 */
TIM22->DIER = TIM_DIER_UIE;
TIM22->CNT = 0;
TIM22->EGR = TIM_EGR_UG;
NVIC_EnableIRQ(TIM22_IRQn); /* Zalaczenie przerwania od TIMER'a */
NVIC_SetPriority(TIM22_IRQn,4); /* Ustawienie priorytetu przerwania od TIMER'a */
The prescaler meant to be 1ms so I divided 524288 of my core speed. Then enabling the interrupt, resetting the counter to make sure it starts from 0 and then manually generating the interrupt. And it does the interrupt 'loop' once but with a single 'if' and variable, I can control what it does.
So what I do, I am calling a function that set the clock and enables the count inside another function enable = 1; timer_init(ms);
Then comes the function call
void timer_init(uint16_t ms)
{
TIM22->CNT = 65535-ms;
TIM22->CR1 |= TIM_CR1_CEN;
}
void TIM22_IRQHandler(void)
/* Up-Counter milisec */
{
if(TIM22->CR1 & TIM_CR1_CEN) {TIM22->CR1 &= ~TIM_CR1_CEN;}
if(TIM22->SR & TIM_SR_UIF) {
if(enable == 1){
GPIOB->BSRR = GPIO_BSRR_BS_7;
}
}
TIM22->SR=0;
}
And the interrupt.
Thanks a lot!Have fun with registers!
I am working on an auto-parking car robot and I am using 8 (hc-sr04) ultrasonic sensors (2 at each side) but the problem is that I am using atmega32 which has limited resources only 3 external interrupts and 3 timers (and even if using interrupts somehow works I might run into risk to have two interrupts triggered at the same time).
I am using this sensor : http://ram-e-shop.com/oscmax/catalog/product_info.php?products_id=907
I've tried using digital I/O pins with polling procedure but it didn't work.
here is the code for polling procedure:
unsigned int read_sonar(){
int dist_in_cm = 0;
init_sonar(); // Setup pins and ports
trigger_sonar(); // send a 10us high pulse
while(!(ECHO_PIN & (1<<ECHO_BIT))){ // while echo pin is still low
USART_Message("echo pin low\r\n");
trig_counter++;
uint32_t max_response_time = SONAR_TIMEOUT;
if (trig_counter > max_response_time){ // SONAR_TIMEOUT
return TRIG_ERROR;
}
}
TCNT1=0; // reset timer
TCCR1B |= (1<<CS10); // start 16 bit timer with no prescaler
TIMSK |= (1<<TOIE1); // enable overflow interrupt on timer1
overFlowCounter=0; // reset overflow counter
sei(); // enable global interrupts
while((ECHO_PIN & (1<<ECHO_BIT))){ // while echo pin is still high
USART_Message("echo pin high\r\n");
if (((overFlowCounter*TIMER_MAX)+TCNT1) > SONAR_TIMEOUT){
USART_Message("timeout");
return ECHO_ERROR; // No echo within sonar range
}
};
TCCR1B = 0x00; // stop 16 bit timer with no prescaler
cli(); // disable global interrupts
no_of_ticks = ((overFlowCounter*TIMER_MAX)+TCNT1); // counter count
dist_in_cm = (no_of_ticks/(CONVERT_TO_CM*CYCLES_PER_US)); // distance in cm
return (dist_in_cm );}
This method doesn't work if I want to read all sensors at the same time, because it gets stuck in the loop for a while.
I also tried using freeRTOS to build a task that checks the state of pins like every 1msec but this won't be a time accurate.
any help?
Assuming that You use internal clock which is 8MHz I would try to handle this inside timer overflow interrupt and would use whole port to connect the sensors.
Use Timer in normal mode or CTC mode (which I find quite intuitive) to ensure periodical interrupts. Set the appropriate period. Remember that the clock has pretty low frequency so don't exaggerate (I think that 0,25 ms will fit).
Connect the sensors to one port, e.g. PORTB. This is a nice situation because ATmega32 has 4 ports with pins numbered from 0-7 and you use 8 sensors so the register for the specific port can cover all of the pins and You can use one read to get states of all of the pins.
Implement the logic:
volatile uint8_t sensors_states;
volatile uint8_t read_flag = 0;
ISR(TIMER0_OVF_vect)
{
sensors_states = PORTB;
read_flag = 1;
}
int main()
{
// Initialize peripherals ...
// You must assume on your own how much time could the pin be held
// in the same state. This is important because the number must not
// be bigger than max value for the type of the array
uint8_t states_time[8] = {0, 0, 0, 0, 0, 0, 0, 0};
uint8_t prev_sensors_states = PORTB;
while(1)
{
// Wait until the flag will be set in the ISR
if(read_flag)
{
for(uint8_t i = 0, mask = 0x80 ; i < 8 ; i ++, mask >>= 1)
{
states_time[i]++;
// Compare the previous state and present state on each pin
uint8_t state = mask & sensors_states;
if((mask & prev_sensors_states) != state)
{
// Here you can use the state of the pin and the duration of that state.
// Remember that when 'state' is > 0 it means that previous state of the
// pin was '0' and if if 'state' is == 0 then the previous pin state
// was '1' (negation).
do_something_with_pin_change(states_time[i], state);
states_time[i] = 0;
}
}
// Save the previous states of the pins
prev_sensors_states = sensors_states;
// Clear the flag to await next data update
read_flag = 0;
}
}
}
If You will try to use FreeRTOS You could use ulTaskNotifyTake and vTaskNotifyGiveFromISR, instead of using read_flag, to implement a simple mechanism which will notify a task from the interrupt that the port has been read. The processor will go into idle state for a while and you could then invoke a sleep function to minimize power consumption.
I don't know what You want to do with this data so I've invoked do_something_with_pin_change function to indicate the point where You can use the data.
To sum up for this solution You would only use one interrupt and of course 8 pins.
Can a interrupt flag be set by the code as in the example below or is that line just an error of thinking? This is just the main function. Below this code snipet are the interrupt it self, is it correct and nessasary to clear the interrupt flag in the end of the code?
if(duty != (uint8_t) (SOFT_PWM_PERIOD - 1))
{
// Request an immediate interrupt if the timer counter has
// already the initial period. This helps minimize glitches
// when changing duty cycles
if(duty < TMR4)
PIR3bits.TMR4IF = 1;
// Finally (re-)start the timer
T4CON =
0 << 3 | // 1x post-scaler
1 << 2 | // Active
2 /* << 0 */; // 16x pre-scaler
IPR3bits.TMR4IP = 1; // TMR4 Overflow Interrupt Priority bit High
PIE3bits.TMR4IE = 1; // TMR4 Overflow Interrupt Enable bit
}
The Intrrupt code ->
// Deal with PWM timer interrupts. Add this to the high-priority interrupt handler.
void SoftPWM_Interrupt(void)
{
volatile uint8_t _SoftPWM_Toggle; // Is this variable really accessed by both the ISR and mainline functions? (C.G)
/* Has a flank been reached yet? */
if(PIR3bits.TMR4IF)
{
/* Alternate between the low and high periods */
PR4 ^= _SoftPWM_Toggle;
/* Try to deal gracefully with the new period already having been reached. */
/* The hardware timer works by checking if TMR4 = PR4 when it is time to increase */
/* counter, in which case TMR4 is reset instead. Thus if is already TMR4 > PR4 due to */
/* interrupt latency then we've missed the period and an extra interrupt is needed. */
/* First acknowledging the flag and then conditionally setting it is necessary to */
/* avoid a race between reading TMR4 and changing the flag. */
/* Finally the the TMR4 > PR4 test is actually implemented as skip if TMR4 < PR4 + 1 */
/* but the increment cannot overflow since the interrupt won't be used for 0% or 100% */
/* duty cycles */
PIR3bits.TMR4IF = 0;
_asm
INCF PR4,0,ACCESS
CPFSLT TMR4,ACCESS
_endasm
/* if(TMR4 > PR4) */
PIR3bits.TMR4IF = 1; // Cant only the harware set this flag? (C.G)
/* Finally toggle the output pin */
SOFT_PWM_PIN ^= 1;
/*Important?*/
PIR3bits.TMR4IF = 0;
}
}
Yes you can set an interrupt flag by soft. But IMO it is not really a good practice to do it...
If you really want the behavior of your ISR to be executed in the normal context, why don't you externalize your ISR code in a function that you can call in your main function?
About the interrupt flag, if you don't clear it, the ISR will be executed in loop and you will never go back to your main program.
I generated my code from STM32CubeMx and wanted to generate a update event every 1µs. I work with the internal clock at 48MHz, which should be with Prescaler:0 and Autoreload:47 result to 1µs.
I use a STM32F030 with TrueStudio V.9.0.0
generated code
/* TIM17 init function */
static void MX_TIM17_Init(void)
{
LL_TIM_InitTypeDef TIM_InitStruct;
/* Peripheral clock enable */
LL_APB1_GRP2_EnableClock(LL_APB1_GRP2_PERIPH_TIM17);
/* TIM17 interrupt Init */
NVIC_SetPriority(TIM17_IRQn, 3);
NVIC_EnableIRQ(TIM17_IRQn);
TIM_InitStruct.Prescaler = 0;
TIM_InitStruct.CounterMode = LL_TIM_COUNTERMODE_UP;
TIM_InitStruct.Autoreload = 47;
TIM_InitStruct.ClockDivision = LL_TIM_CLOCKDIVISION_DIV1;
TIM_InitStruct.RepetitionCounter = 0;
LL_TIM_Init(TIM17, &TIM_InitStruct);
LL_TIM_EnableARRPreload(TIM17);
}
I added in my init:
LL_TIM_EnableIT_UPDATE(TIM17);
LL_TIM_EnableCounter(TIM17);
In the IRQ_Handler i toggle a PIN:
void TIM17_IRQHandler(void)
{
/* USER CODE BEGIN TIM17_IRQn 0 */
LL_GPIO_TogglePin(LED_D2_2_GPIO_Port,LED_D2_2_Pin);
/* USER CODE END TIM17_IRQn 0 */
/* USER CODE BEGIN TIM17_IRQn 1 */
/* USER CODE END TIM17_IRQn 1 */
}
After flashing my device with the code it generates a Signal with Frequency 889kHz with Pulsewidth of 564ns measured with Oscilloscope. Changes on Prescaler or Autoreload does not affect this output, it stays right away at T_Pulse=564ns or F=889kHz.
Any idea what I am missing here?
Register output from debugging:
CR1:0x81 CR2:0
DIER:0x01 SR:0x03
CCMR1_O/I:0 CCER:0
PSC:0 ARR:0x2f
RCR:0 CCR1:0
BDTR:0 DCR:0
DMAR:0x81
The solution was the clearance of the FLAG UIE in TIM17->SR in the IRQ_Handler.
I was stuck permanently in the IRQ routine.