mirror of
https://github.com/eddyem/stm32samples.git
synced 2025-12-06 10:45:11 +03:00
159 lines
5.3 KiB
C
159 lines
5.3 KiB
C
/*
|
|
* This file is part of the nitrogen project.
|
|
* Copyright 2023 Edward V. Emelianov <edward.emelianoff@gmail.com>.
|
|
*
|
|
* This program is free software: you can redistribute it and/or modify
|
|
* it under the terms of the GNU General Public License as published by
|
|
* the Free Software Foundation, either version 3 of the License, or
|
|
* (at your option) any later version.
|
|
*
|
|
* This program is distributed in the hope that it will be useful,
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
* GNU General Public License for more details.
|
|
*
|
|
* You should have received a copy of the GNU General Public License
|
|
* along with this program. If not, see <http://www.gnu.org/licenses/>.
|
|
*/
|
|
|
|
#include "adc.h"
|
|
#ifdef EBUG
|
|
#include "proto.h"
|
|
#endif
|
|
|
|
/**
|
|
* @brief ADCx_array - arrays for ADC channels with median filtering:
|
|
* ADC1:
|
|
* 0..9 - AIN0..9 (ADC1_IN1..10)
|
|
* 10 - internal Tsens - ADC1_IN16
|
|
* ADC2:
|
|
* 11 - AINext - (ADC2 in 1)
|
|
*/
|
|
static uint16_t ADC_array[NUMBER_OF_ADC_CHANNELS*9];
|
|
|
|
TRUE_INLINE void calADC(ADC_TypeDef *chnl){
|
|
// calibration
|
|
// enable voltage regulator
|
|
chnl->CR = 0;
|
|
chnl->CR = ADC_CR_ADVREGEN_0;
|
|
// wait for 10us
|
|
uint16_t ctr = 0;
|
|
while(++ctr < 1000){nop();}
|
|
// ADCALDIF=0 (single channels)
|
|
if((chnl->CR & ADC_CR_ADEN)){
|
|
chnl->CR |= ADC_CR_ADSTP;
|
|
chnl->CR |= ADC_CR_ADDIS;
|
|
}
|
|
chnl->CR |= ADC_CR_ADCAL;
|
|
while((chnl->CR & ADC_CR_ADCAL) != 0 && ++ctr < 0xfff0){};
|
|
chnl->CR = ADC_CR_ADVREGEN_0;
|
|
// enable ADC
|
|
ctr = 0;
|
|
do{
|
|
chnl->CR |= ADC_CR_ADEN;
|
|
}while((chnl->ISR & ADC_ISR_ADRDY) == 0 && ++ctr < 0xfff0);
|
|
}
|
|
|
|
TRUE_INLINE void enADC(ADC_TypeDef *chnl){
|
|
// ADEN->1, wait ADRDY
|
|
chnl->CR |= ADC_CR_ADEN;
|
|
uint16_t ctr = 0;
|
|
while(!(chnl->ISR & ADC_ISR_ADRDY) && ++ctr < 0xffff){}
|
|
chnl->CR |= ADC_CR_ADSTART; /* start the ADC conversions */
|
|
}
|
|
|
|
/**
|
|
* ADC1 - DMA1_ch1
|
|
* ADC2 - DMA2_ch1
|
|
*/
|
|
// Setup ADC
|
|
void adc_setup(){
|
|
RCC->AHBENR |= RCC_AHBENR_ADC12EN; // Enable clocking
|
|
ADC12_COMMON->CCR = ADC_CCR_TSEN | ADC_CCR_CKMODE; // enable Tsens, HCLK/4
|
|
calADC(ADC1);
|
|
calADC(ADC2);
|
|
// ADC1: channels 1-10,16; ADC2: channel 1
|
|
ADC1->SMPR1 = ADC_SMPR1_SMP1 | ADC_SMPR1_SMP2 | ADC_SMPR1_SMP3 | ADC_SMPR1_SMP4 | ADC_SMPR1_SMP5
|
|
| ADC_SMPR1_SMP6 | ADC_SMPR1_SMP7 | ADC_SMPR1_SMP8 | ADC_SMPR1_SMP9 ;
|
|
ADC1->SMPR2 = ADC_SMPR2_SMP10 | ADC_SMPR2_SMP16;
|
|
// 11 conversions in group: 1...10->16
|
|
ADC1->SQR1 = (1<<6) | (2<<12) | (3<<18) | (4<<24) | (NUMBER_OF_ADC1_CHANNELS-1);
|
|
ADC1->SQR2 = (5<<0) | (6<<6) | (7<<12) | (8<<18) | (9<<24);
|
|
ADC1->SQR3 = (10<<0)| (16<<6);
|
|
ADC2->SMPR1 = ADC_SMPR1_SMP1;
|
|
ADC2->SQR1 = (1<<6) | (NUMBER_OF_ADC2_CHANNELS-1);
|
|
// configure DMA for ADC
|
|
ADC1->CFGR = ADC_CFGR_CONT | ADC_CFGR_DMAEN | ADC_CFGR_DMACFG;
|
|
ADC2->CFGR = ADC_CFGR_CONT | ADC_CFGR_DMAEN | ADC_CFGR_DMACFG;
|
|
DMA1_Channel1->CPAR = (uint32_t) (&(ADC1->DR));
|
|
DMA1_Channel1->CMAR = (uint32_t)(ADC_array);
|
|
DMA1_Channel1->CNDTR = NUMBER_OF_ADC1_CHANNELS * 9;
|
|
DMA1_Channel1->CCR |= DMA_CCR_MINC | DMA_CCR_MSIZE_0 | DMA_CCR_PSIZE_0 | DMA_CCR_CIRC;
|
|
DMA1_Channel1->CCR |= DMA_CCR_EN;
|
|
|
|
DMA2_Channel1->CPAR = (uint32_t) (&(ADC2->DR));
|
|
DMA2_Channel1->CMAR = (uint32_t)(&ADC_array[ADC2START]);
|
|
DMA2_Channel1->CNDTR = NUMBER_OF_ADC2_CHANNELS * 9;
|
|
DMA2_Channel1->CCR |= DMA_CCR_MINC | DMA_CCR_MSIZE_0 | DMA_CCR_PSIZE_0 | DMA_CCR_CIRC;
|
|
DMA2_Channel1->CCR |= DMA_CCR_EN;
|
|
|
|
enADC(ADC1);
|
|
enADC(ADC2);
|
|
}
|
|
|
|
/**
|
|
* @brief getADCval - calculate median value for `nch` channel
|
|
* @param nch - number of channel
|
|
* @return
|
|
*/
|
|
uint16_t getADCval(int nch){
|
|
register uint16_t temp;
|
|
#define PIX_SORT(a,b) { if ((a)>(b)) PIX_SWAP((a),(b)); }
|
|
#define PIX_SWAP(a,b) { temp=(a);(a)=(b);(b)=temp; }
|
|
uint16_t p[9];
|
|
int addr = nch, adval = NUMBER_OF_ADC1_CHANNELS;
|
|
if(nch >= NUMBER_OF_ADC1_CHANNELS){
|
|
adval = NUMBER_OF_ADC2_CHANNELS;
|
|
addr += ADC2START - NUMBER_OF_ADC1_CHANNELS;
|
|
}
|
|
for(int i = 0; i < 9; ++i, addr += adval) // first we should prepare array for optmed
|
|
p[i] = ADC_array[addr];
|
|
PIX_SORT(p[1], p[2]) ; PIX_SORT(p[4], p[5]) ; PIX_SORT(p[7], p[8]) ;
|
|
PIX_SORT(p[0], p[1]) ; PIX_SORT(p[3], p[4]) ; PIX_SORT(p[6], p[7]) ;
|
|
PIX_SORT(p[1], p[2]) ; PIX_SORT(p[4], p[5]) ; PIX_SORT(p[7], p[8]) ;
|
|
PIX_SORT(p[0], p[3]) ; PIX_SORT(p[5], p[8]) ; PIX_SORT(p[4], p[7]) ;
|
|
PIX_SORT(p[3], p[6]) ; PIX_SORT(p[1], p[4]) ; PIX_SORT(p[2], p[5]) ;
|
|
PIX_SORT(p[4], p[7]) ; PIX_SORT(p[4], p[2]) ; PIX_SORT(p[6], p[4]) ;
|
|
PIX_SORT(p[4], p[2]) ;
|
|
#undef PIX_SORT
|
|
#undef PIX_SWAP
|
|
#ifdef EBUG
|
|
DBG("val: "); printu(p[4]); newline();
|
|
#endif
|
|
return p[4];
|
|
}
|
|
|
|
// get voltage @input nch (V)
|
|
float getADCvoltage(int nch){
|
|
float v = getADCval(nch) * 3.3;
|
|
v /= 4096.f; // 12bit ADC
|
|
#ifdef EBUG
|
|
DBG("v="); printf(v); newline();
|
|
#endif
|
|
return v;
|
|
}
|
|
|
|
// return MCU temperature (degrees of celsius)
|
|
float getMCUtemp(){
|
|
// make correction on Vdd value
|
|
int32_t ADval = getADCval(ADC_TSENS);
|
|
float temperature = ADval - (float) *TEMP30_CAL_ADDR;
|
|
temperature *= (110.f - 30.f);
|
|
temperature /= (float)(*TEMP110_CAL_ADDR - *TEMP30_CAL_ADDR);
|
|
temperature += 30.f;
|
|
#ifdef EBUG
|
|
DBG("t="); printf(temperature); newline();
|
|
#endif
|
|
return(temperature);
|
|
}
|