piotr/edgetx-gx12
1
0
Fork 0
Code Issues 2 Pull Requests Actions Packages Projects Releases Wiki Activity

[avr boards] Minor fix to legacy problem with the very first PPM

Browse Source 
output frame, where correct pulse polarity was not always set 
until the second call to setupPulses, from within the PPM mode 
pulse train interrupt handler.

I tested proper function in PPM mode on the logic analyser 
before committing.
This commit is contained in:
gruvin@gmail.com
2012-11-11 03:31:45 +00:00
parent 340043a73e
commit 434c071bda
1 changed files with 10 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/pulses_avr.cpp
+10 -14
View File
@@ -67,7 +67,7 @@ void startPulses()
#endif
{
// TODO g: There has to be a better place for this bug fix
OCR1B = 0xffff; /* Prevent any PPM_PUT pin toggle before the TCNT1 interrupt
OCR1B = 0xffff; /* Prevent any PPM_OUT pin toggle before the TCNT1 interrupt
fires for the first time and sets up the pulse period.
*** Prevents WDT reset loop. */
}
@@ -96,12 +96,9 @@ uint8_t *pulses2MHzWPtr = pulses2MHz;
#ifndef SIMU
uint8_t g_ppmPulsePolarity = 0;
ISR(TIMER1_COMPA_vect) //2MHz pulse generation (BLOCKING ISR)
{
static uint8_t pulsePol = 0; /* The very first call to this handler will toggle PPM_out high, then
call setupPulses() and initialise pulsePol. This is allowed in favour
of minimal ongoing PPM toggle latency. */
uint8_t dt = TCNT1L; // record Timer1 latency for DEBUG stats display
#ifdef DSM2_SERIAL
@@ -118,13 +115,13 @@ ISR(TIMER1_COMPA_vect) //2MHz pulse generation (BLOCKING ISR)
#if !defined(PCBGRUVIN9X)
// Original bitbang for PPM
if (s_current_protocol != PROTO_NONE) {
if (pulsePol) {
if (g_ppmPulsePolarity) {
PORTB |= (1<<OUT_B_PPM); // GCC optimisation should result in a single SBI instruction
pulsePol = 0;
g_ppmPulsePolarity = 0;
}
else {
PORTB &= ~(1<<OUT_B_PPM);
pulsePol = 1;
g_ppmPulsePolarity = 1;
}
}
#else
@@ -134,13 +131,13 @@ ISR(TIMER1_COMPA_vect) //2MHz pulse generation (BLOCKING ISR)
// Toggle bit: Can't read PPM_OUT I/O pin when OC1B is connected (on the ATmega2560 -- can on ATmega64A!)
// so need to use pusePol register to keep track of PPM_out polarity.
if (s_current_protocol != PROTO_NONE) {
if (pulsePol) {
if (g_ppmPulsePolarity) {
TCCR1A = (3<<COM1B0); // SET the state of PB6(OC1B)/PPM_out on next TCNT1==OCR1B
pulsePol = 0;
g_ppmPulsePolarity = 0;
}
else {
TCCR1A = (2<<COM1B0); // CLEAR the state of PB6(OC1B)/PPM_out on next TCNT1==OCR1B
pulsePol = 1;
g_ppmPulsePolarity = 1;
}
}
#endif
@@ -150,8 +147,6 @@ ISR(TIMER1_COMPA_vect) //2MHz pulse generation (BLOCKING ISR)
pulses2MHzRPtr += sizeof(uint16_t); // non PPM protocols use uint8_t pulse buffer
if (*((uint16_t*)pulses2MHzRPtr) == 0) {
pulsePol = g_model.pulsePol;
#if defined(PCBGRUVIN9X)
TIMSK1 &= ~(1<<OCIE1A); // stop reentrance (disable Timer1 interrupt)
#else
@@ -905,7 +900,8 @@ void setupPulses()
break;
#endif
default:
default: // standard PPM protocol
g_ppmPulsePolarity = g_model.pulsePol;
#if defined(PCBGRUVIN9X)
OCR5A = (uint16_t)0x7d * (45+g_model.ppmFrameLength-lastMixerDuration-2/*1ms*/);
TCNT5 = 0;