Calculate atan2 in CCS (limit from -180 to 180 degree)

For general tips on floating point math like this, please see this wiki article.  

I presume the type float16 is a 32-bit floating point type.  Consider changing to a 64-bit type.  Be sure to use compiler version 15.12.0.LTS or later, so you can use C99 64-bit math functions like atan2l.

Thanks and regards,

-George

I use CCS the version 6.1.2 , so I can say it the latest version

The microcontroller has data of 32 bits, so it's useless when you use 64 bits

Maxime Torrelli said:

The microcontroller has data of 32 bits, so it's useless when you use 64 bits

That's not correct.  You may need your final results to be 32-bit, but you can compute intermediate results with 64-bit types.

Thanks and regards,

-George

When I use C2000 compiler version 15.12.0.LTS with no options, I get the correct value with the following test case, so there's nothing wrong with the library function.  Please show us your complete command line options.  You can find this in the build console window.

#include <stdio.h>
#include <math.h>

#define PI 3.14159

int main()
{
    double sinteta = -0.6557275;
    double costeta =  0.7352034;
    double gAngle0;

    gAngle0 = (atan2(sinteta,costeta)*180)/PI;

    printf("sinteta %f\n", sinteta);
    printf("costeta %f\n", costeta);
    printf("atan2(sinteta, costeta) %f\n", atan2(sinteta,costeta));
    printf("gAngle0 %f\n", gAngle0);
    printf("gAngle  %ld\n", (long)gAngle0 % 180);
}

sinteta -0.655728
costeta 0.735203
atan2(sinteta, costeta) -0.728322
gAngle0 -41.729786
gAngle  -41

Here is all of my code. I modified it from the example project  "Example_2806xAdcSoc " in the controlSuite.

#include "DSP28x_Project.h"      // Device Headerfile and Examples Include File
#include "math.h"



#define Adc_ValeurBits     4096  // 12 bits
#define Adc_ValeurTension  3.3  // 0 - 3.3 V
#define Adc_Offset1        1.557  // Sin offset
#define Adc_Ampli1         0.4553
#define Adc_Offset2        1.574  // Cos offset
#define Adc_Ampli2         0.4375
#define PI                 3.14159265359
#define Adc_CalibBits      3
#define Adc_CalibVerifier  0.05   // 5% erreur


// Prototype statements for functions found within this file.
__interrupt void adc_isr(void);
void Adc_Config(void);

// Global variables used in this example:
float16 gVoltage0[3];
float16 gVoltage1[3];
float16 sinteta;
float16 costeta;
float16 gAngle0;
float16 gVerifier;
int16 gAngle;
Uint16 gcount;



main()
{

// Step 1. Initialize System Control:
// PLL, WatchDog, enable Peripheral Clocks
// This example function is found in the F2806x_SysCtrl.c file.
   InitSysCtrl();


// Step 2. Initialize GPIO:
// This example function is found in the F2806x_Gpio.c file and
// illustrates how to set the GPIO to it's default state.
// InitGpio();  // Skipped for this example

// Step 3. Clear all interrupts and initialize PIE vector table:
// Disable CPU interrupts
   DINT;

// Initialize the PIE control registers to their default state.
// The default state is all PIE interrupts disabled and flags
// are cleared.
// This function is found in the F2806x_PieCtrl.c file.
   InitPieCtrl();

// Disable CPU interrupts and clear all CPU interrupt flags:
   IER = 0x0000;
   IFR = 0x0000;

// Initialize the PIE vector table with pointers to the shell Interrupt
// Service Routines (ISR).
// This will populate the entire table, even if the interrupt
// is not used in this example.  This is useful for debug purposes.
// The shell ISR routines are found in F2806x_DefaultIsr.c.
// This function is found in F2806x_PieVect.c.
   InitPieVectTable();

// Interrupts that are used in this example are re-mapped to
// ISR functions found within this file.
   EALLOW;  // This is needed to write to EALLOW protected register
   PieVectTable.ADCINT1 = &adc_isr;
   EDIS;    // This is needed to disable write to EALLOW protected registers

// Step 4. Initialize all the Device Peripherals:
// This function is found in F2806x_InitPeripherals.c
// InitPeripherals(); // Not required for this example
   InitAdc();  // For this example, init the ADC
   AdcOffsetSelfCal();




// Step 5. User specific code, enable interrupts:

// Enable ADCINT1 in PIE
   PieCtrlRegs.PIEIER1.bit.INTx1 = 1;	// Enable INT 1.1 in the PIE
   IER |= M_INT1; 						// Enable CPU Interrupt 1
   EINT;          						// Enable Global interrupt INTM
   ERTM;          						// Enable Global realtime interrupt DBGM


// Configure ADC
	EALLOW;
    AdcRegs.ADCCTL2.bit.ADCNONOVERLAP  = 1;	// Enable non-overlap mode
	AdcRegs.ADCCTL1.bit.INTPULSEPOS	   = 1;	// ADCINT1 trips after AdcResults latch
	AdcRegs.INTSEL1N2.bit.INT1E        = 1;	// Enabled ADCINT1
	AdcRegs.INTSEL1N2.bit.INT1CONT     = 0;	// Disable ADCINT1 Continuous mode
    AdcRegs.INTSEL1N2.bit.INT1SEL 	   = 1;    // setup EOC1 to trigger ADCINT1 to fire
    AdcRegs.ADCSAMPLEMODE.bit.SIMULEN0 = 1 ;  //Simulating mode
    AdcRegs.ADCSOC0CTL.bit.CHSEL 	   = 2;    // set SOC0 channel select to ADCINA2 + ADCINB2
    AdcRegs.ADCSOC0CTL.bit.TRIGSEL 	   = 5;    // set SOC0 start trigger on EPWM1A, due to round-robin SOC0 converts first then SOC1
	AdcRegs.ADCSOC0CTL.bit.ACQPS 	   = 6;	// set SOC0 S/H Window to 7 ADC Clock Cycles, (6 ACQPS plus 1)
	EDIS;

// Assumes ePWM1 clock is already enabled in InitSysCtrl()452;
   EPwm1Regs.ETSEL.bit.SOCAEN	= 1;		// Enable SOC on A group
   EPwm1Regs.ETSEL.bit.SOCASEL	= 4;		// Select SOC from CMPA on upcount
   EPwm1Regs.ETPS.bit.SOCAPRD 	= 1;		// Generate pulse on 1st event
   EPwm1Regs.CMPA.half.CMPA 	= 0;	// Set compare A value
   EPwm1Regs.TBPRD 				= 113;  // Set period ADC for ePWM1  - 400kHz  - 2.5 us
   EPwm1Regs.TBCTL.bit.CTRMODE 	= 0;		// count up and start


   gcount= 0;
   // Wait for ADC interrupt
   for(;;)
   {


   }

}


__interrupt void  adc_isr(void)
{
  // Mesurer les signaux : 20 periode ADC pour prendre toutes les valeurs - 50 us
  gVoltage0[gcount] = AdcResult.ADCRESULT0;   //Sin
  gVoltage1[gcount] = AdcResult.ADCRESULT1;   //Cos
  // Mesurer la deuxieme fois
  if (gcount == 1)
   {
	  // Valeurs moyennes
	  gVoltage0[2] = (gVoltage0[0] +gVoltage0[1]) /2;
	  gVoltage1[2] = (gVoltage1[0] +gVoltage1[1]) /2 ;
      // Deduire les valeurs en tension
	  gVoltage0[2] = (gVoltage0[2]/Adc_ValeurBits)*Adc_ValeurTension - Adc_Offset1;
	  gVoltage1[2] = (gVoltage1[2]/Adc_ValeurBits)*Adc_ValeurTension - Adc_Offset2;
	  // Valeurs de cosinus et sinus
	  sinteta  = gVoltage0[2]/Adc_Ampli1;
	  costeta =  gVoltage1[2]/Adc_Ampli2;
	  // Calibration des valeurs des cosinus
	  gVerifier =   (sinteta)*(sinteta) + (costeta)*(costeta);
	  if ( (gVerifier > (1+Adc_CalibVerifier)) || (gVerifier <(1-Adc_CalibVerifier)))
		{
		  if (costeta < 0)
		  {
			  costeta  =  - sqrt(1-sinteta*sinteta);
	      }
		  else costeta  =  sqrt(1-sinteta*sinteta) ;
		}
      // Calcul d'angle
	  gAngle0 =  (atan2(sinteta,costeta)*180)/PI;
	  gAngle =  ((int16) gAngle0) % 180 ;
	  gcount = 0;
   }
  else gcount++;


  AdcRegs.ADCINTFLGCLR.bit.ADCINT1 = 1;		//Clear ADCINT1 flag reinitialize for next SOC
  PieCtrlRegs.PIEACK.all = PIEACK_GROUP1;   // Acknowledge interrupt to PIE

  return;
}

I cant find out the problem, maybe the example has problems itself

So we use 64-bit type to calculate and after, shift the bit to get the value of 32-bit type ?

I think that I find out the problem, in the preference window, I cant find C2000 compiler version 15.12.0.LTS existing. So maybe the installation has the problem. So how I can update my compiler ?

I used "Install the new software" with the option "Code Generation Tools update", but the site of ti refused my connection. Do you have any solution ?

Maxime Torrelli said:

I used "Install the new software" with the option "Code Generation Tools update", but the site of ti refused my connection. Do you have any solution ?

There are multiple sites to choose from for updating the compiler.  Try another one.  See this screen shot below.

Thanks and regards,

-George

Looking at your screenshot, I see you are using C2000 compiler version 6.4.6. Using this version, my test case operates as expected, so we have not yet established that upgrading to 15.12.0.LTS will improve your example.

On C2000, the type "double" is only 32 bits. I don't see any 64-bit arithmetic in your example.

Please show us your complete command line options from the build console window.

Like you said, somehow after updating the compiler v15, it changed nothing. In fact, it updated value of Angle0 faster, but sended numbers out of range. Really  weird.

Here is what exists in the build console with compiler v6

**** Build of configuration Debug for project Example_2806xAdcSoc ****

"C:\\ti\\ccsv6\\utils\\bin\\gmake" -k all 
'Building file: ../Example_2806xAdcSoc.c'
'Invoking: C2000 Compiler'
"C:/ti/ccsv6/tools/compiler/ti-cgt-c2000_6.4.6/bin/cl2000" -v28 -ml -mt --cla_support=cla0 --float_support=fpu32 --vcu_support=vcu0 --include_path="C:/ti/ccsv6/tools/compiler/ti-cgt-c2000_6.4.6/include" --include_path="/packages/ti/xdais" --include_path="C:/ti/controlSUITE/device_support/f2806x/v151/F2806x_headers/include" --include_path="C:/ti/controlSUITE/device_support/f2806x/v151/F2806x_common/include" --include_path="C:/ti/controlSUITE/libs/math/IQmath/v160/include" --include_path="C:/ti/controlSUITE/libs/math/FPUfastRTS/V100/include" -g --define="_DEBUG" --define="LARGE_MODEL" --quiet --verbose_diagnostics --diag_warning=225 --diag_suppress=10063 --issue_remarks --output_all_syms --cdebug_asm_data --preproc_with_compile --preproc_dependency="Example_2806xAdcSoc.pp"  "../Example_2806xAdcSoc.c"
'Finished building: ../Example_2806xAdcSoc.c'
' '
'Building target: Example_2806xAdcSoc.out'
'Invoking: C2000 Linker'
"C:/ti/ccsv6/tools/compiler/ti-cgt-c2000_6.4.6/bin/cl2000" -v28 -ml -mt --cla_support=cla0 --float_support=fpu32 --vcu_support=vcu0 -g --define="_DEBUG" --define="LARGE_MODEL" --quiet --verbose_diagnostics --diag_warning=225 --diag_suppress=10063 --issue_remarks --output_all_syms --cdebug_asm_data -z -m"Example_2806xAdcSoc.map" --stack_size=0x200 --warn_sections -i"C:/ti/ccsv6/tools/compiler/ti-cgt-c2000_6.4.6/lib" -i"C:/ti/ccsv6/tools/compiler/ti-cgt-c2000_6.4.6/include" -i"C:/ti/controlSUITE/device_support/f2806x/v151/F2806x_common/lib" -i"C:/ti/controlSUITE/libs/math/IQmath/v160/lib" -i"C:/ti/controlSUITE/libs/math/FPUfastRTS/V100/lib" --reread_libs --verbose_diagnostics --issue_remarks --xml_link_info="Example_2806xAdcSoc_linkInfo.xml" --entry_point=code_start --rom_model -o "Example_2806xAdcSoc.out" -l"rts2800_fpu32_fast_supplement.lib" -l"rts2800_fpu32.lib" "./Example_2806xAdcSoc.obj" "./F2806x_Adc.obj" "./F2806x_CodeStartBranch.obj" "./F2806x_DefaultIsr.obj" "./F2806x_GlobalVariableDefs.obj" "./F2806x_PieCtrl.obj" "./F2806x_PieVect.obj" "./F2806x_SysCtrl.obj" "./F2806x_usDelay.obj" "C:/ti/controlSUITE/device_support/f2806x/v151/F2806x_common/cmd/28069_RAM_lnk.cmd" "C:/ti/controlSUITE/device_support/f2806x/v151/F2806x_headers/cmd/F2806x_Headers_nonBIOS.cmd"  -l"IQmath_fpu32.lib" 
'Finished building target: Example_2806xAdcSoc.out'
' '

**** Build Finished ****

They are all default build configurations of example "Example_2806xAdcSoc" in controlSUITE. I have just modified the code

Did you really implant your code in a microcontroller ? because when I use your code calculating atan2 with a microcontroller and with different programs, it doesn't work. But if I run seperately it , without implantation, like you have done, it works. So maybe the problem is that CCS compiles in wrong way for microcontroller.

If you noticed, I have said the value of variable is extended by its data type. So i don't know if data type is also a problem

No, I ran the executable under a simulator so that I could use printf. This demonstrates that the compiler is generating correct code for atan2. The compiler doesn't generate code differently if the code will be embedded.

The correct value for gAngle0 (approximately 41) easily fits in range for a 32-bit float variable, so there should be no issues with whether you use float, double, or long double.

Perhaps CCS is displaying the value of gAngle0 incorrectly. Please show us the memory window view of gAngle0 when CCS shows its value as 1.23e+09.

I run my code with the new compiler v15, and here is the memory window for all variables.

The strange here is the microcontroller use the wrong number of Angle0 to calculate Angle.