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

Throttle trace with P1-3 should now be ok

Browse Source 
Problem with recursive mixers and Flight Phases
This commit is contained in:
bsongis
2012-02-24 18:08:27 +00:00
parent 16d35bfd99
commit a094fd9ba2
1 changed files with 28 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/open9x.cpp
+28 -12
View File
@@ -1216,7 +1216,8 @@ void perOut(int16_t *chanOut, uint8_t phase)
//========== MIXER LOOP ===============
mixWarning = 0;
for(uint8_t i=0;i<MAX_MIXERS;i++){
for(uint8_t i=0; i<MAX_MIXERS; i++) {
MixData *md = mixaddress( i ) ;
if((md->destCh==0) || (md->destCh>NUM_CHNOUT)) break;
@@ -1237,7 +1238,7 @@ void perOut(int16_t *chanOut, uint8_t phase)
uint8_t swTog;
//swOn[i]=false;
if(!getSwitch(md->swtch,1)){ // switch on? if no switch selected => on
if (!getSwitch(md->swtch,1)) { // switch on? if no switch selected => on
swTog = swOn[i];
swOn[i] = false;
if(md->srcRaw!=MIX_MAX && md->srcRaw!=MIX_FULL) continue;// if not MAX or FULL - next loop
@@ -1291,8 +1292,8 @@ void perOut(int16_t *chanOut, uint8_t phase)
if (diff && (md->speedUp || md->speedDown)) {
//rate = steps/sec => 32*1024/100*md->speedUp/Down
//act[i] += diff>0 ? (32768)/((int16_t)100*md->speedUp) : -(32768)/((int16_t)100*md->speedDown);
//-100..100 => 32768 -> 100*83886/256 = 32768, For MAX we divide by 2 sincde it's asymmetrical
if(tick10ms) {
//-100..100 => 32768 -> 100*83886/256 = 32768, For MAX we divide by 2 since it's asymmetrical
if (tick10ms) {
int32_t rate = (int32_t)DEL_MULT*2048*100;
if(md->weight) rate /= abs(md->weight);
// TODO port optim er9x by Mike
@@ -1358,7 +1359,7 @@ void perOut(int16_t *chanOut, uint8_t phase)
// at the end chans[i] = chans[i]/100 => -1024..1024
// interpolate value with min/max so we get smooth motion from center to stop
// this limits based on v original values and min=-1024, max=1024 RESX=1024
//printf("chans%d=%d\n", i, chans[i]);fflush(stdout);
int32_t q = chans[i];// + (int32_t)g_model.limitData[i].offset*100; // offset before limit
chans[i] /= 100; // chans back to -1024..1024
@@ -1403,7 +1404,12 @@ void perMain()
static uint16_t fp_act[MAX_PHASES] = {0};
static uint16_t delta = 0;
static uint8_t s_fade_flight_phases = 0;
static uint8_t s_last_phase = 255;
static uint8_t s_last_phase = 255;
#ifdef PCBV4
static int16_t s_fp_chans[NUM_CHNOUT][MAX_PHASES];
#else
static int8_t s_fp_chans[NUM_CHNOUT][MAX_PHASES];
#endif
uint8_t phase = getFlightPhase();
if (s_last_phase != phase) {
@@ -1431,11 +1437,21 @@ void perMain()
int32_t sum_chans512[NUM_CHNOUT] = {0};
int32_t weight = 0;
for (uint8_t p=0; p<MAX_PHASES; p++) {
if (s_fade_flight_phases & (1<<p)) {
perOut(next_chans512, p);
perOut(next_chans512, p); // TODO temporary fix for issue 80. I think th9x has a better algorithm for dealing with it!
// printf("perOut(%d - %d)=>%d\n", p, fp_act[p], next_chans512[2]);
if (s_fade_flight_phases & (1<<p)) {
for (uint8_t i=0; i<NUM_CHNOUT; i++) {
#ifdef PCBV4
chans[i] = ((int32_t)s_fp_chans[i][p]) << 2;
#else
chans[i] = ((int32_t)s_fp_chans[i][p]) << 10;
#endif
}
perOut(next_chans512, p);
for (uint8_t i=0; i<NUM_CHNOUT; i++) {
#ifdef PCBV4
s_fp_chans[i][p] = chans[i] >> 2;
#else
s_fp_chans[i][p] = chans[i] >> 10;
#endif
sum_chans512[i] += (int32_t)next_chans512[i] * fp_act[p];
}
weight += fp_act[p];
@@ -1454,7 +1470,7 @@ void perMain()
for (uint8_t i=0; i<NUM_CHNOUT; i++) {
cli();
g_chans512[i] = next_chans512[i];
g_chans512[i] = next_chans512[i];
sei();
}
@@ -1477,7 +1493,7 @@ void perMain()
val = RESX + g_chans512[g_model.thrTraceSrc-NUM_POTS-1];
}
else {
val = calibratedStick[g_model.thrTraceSrc+NUM_STICKS-1];
val = RESX + calibratedStick[g_model.thrTraceSrc+NUM_STICKS-1];
}
val /= (RESX/16); // calibrate it