Merge pull request #1926 from nophead/acceleration_fix

Fixed jump in speed at high accelerations on axes with lots of steps
This commit is contained in:
Scott Lahteine 2015-04-18 03:09:54 -07:00
commit 67a4471324

View file

@ -54,7 +54,7 @@ static unsigned int cleaning_buffer_counter;
locked_z2_motor = false;
#endif
// Counter variables for the bresenham line tracer
// Counter variables for the Bresenham line tracer
static long counter_x, counter_y, counter_z, counter_e;
volatile static unsigned long step_events_completed; // The number of step events executed in the current block
@ -66,7 +66,7 @@ volatile static unsigned long step_events_completed; // The number of step event
static long acceleration_time, deceleration_time;
//static unsigned long accelerate_until, decelerate_after, acceleration_rate, initial_rate, final_rate, nominal_rate;
static unsigned short acc_step_rate; // needed for deccelaration start point
static unsigned short acc_step_rate; // needed for deceleration start point
static char step_loops;
static unsigned short OCR1A_nominal;
static unsigned short step_loops_nominal;
@ -205,8 +205,14 @@ volatile signed char count_direction[NUM_AXIS] = { 1, 1, 1, 1 };
// intRes = longIn1 * longIn2 >> 24
// uses:
// r26 to store 0
// r27 to store the byte 1 of the 48bit result
#define MultiU24X24toH16(intRes, longIn1, longIn2) \
// r27 to store bits 16-23 of the 48bit result. The top bit is used to round the two byte result.
// note that the lower two bytes and the upper byte of the 48bit result are not calculated.
// this can cause the result to be out by one as the lower bytes may cause carries into the upper ones.
// B0 A0 are bits 24-39 and are the returned value
// C1 B1 A1 is longIn1
// D2 C2 B2 A2 is longIn2
//
#define MultiU24X32toH16(intRes, longIn1, longIn2) \
asm volatile ( \
"clr r26 \n\t" \
"mul %A1, %B2 \n\t" \
@ -237,6 +243,11 @@ volatile signed char count_direction[NUM_AXIS] = { 1, 1, 1, 1 };
"lsr r27 \n\t" \
"adc %A0, r26 \n\t" \
"adc %B0, r26 \n\t" \
"mul %D2, %A1 \n\t" \
"add %A0, r0 \n\t" \
"adc %B0, r1 \n\t" \
"mul %D2, %B1 \n\t" \
"add %B0, r0 \n\t" \
"clr r1 \n\t" \
: \
"=&r" (intRes) \
@ -313,7 +324,7 @@ void enable_endstops(bool check) { check_endstops = check; }
// The trapezoid is the shape the speed curve over time. It starts at block->initial_rate, accelerates
// first block->accelerate_until step_events_completed, then keeps going at constant speed until
// step_events_completed reaches block->decelerate_after after which it decelerates until the trapezoid generator is reset.
// The slope of acceleration is calculated with the leib ramp alghorithm.
// The slope of acceleration is calculated using v = u + at where t is the accumulated timer values of the steps so far.
void st_wake_up() {
// TCNT1 = 0;
@ -469,7 +480,7 @@ ISR(TIMER1_COMPA_vect) {
if ((current_block->steps[A_AXIS] != current_block->steps[B_AXIS]) || (TEST(out_bits, A_AXIS) == TEST(out_bits, B_AXIS))) {
if (TEST(out_bits, X_HEAD))
#else
if (TEST(out_bits, X_AXIS)) // stepping along -X axis (regular cartesians bot)
if (TEST(out_bits, X_AXIS)) // stepping along -X axis (regular Cartesian bot)
#endif
{ // -direction
#ifdef DUAL_X_CARRIAGE
@ -714,7 +725,7 @@ ISR(TIMER1_COMPA_vect) {
unsigned short step_rate;
if (step_events_completed <= (unsigned long)current_block->accelerate_until) {
MultiU24X24toH16(acc_step_rate, acceleration_time, current_block->acceleration_rate);
MultiU24X32toH16(acc_step_rate, acceleration_time, current_block->acceleration_rate);
acc_step_rate += current_block->initial_rate;
// upper limit
@ -737,7 +748,7 @@ ISR(TIMER1_COMPA_vect) {
#endif
}
else if (step_events_completed > (unsigned long)current_block->decelerate_after) {
MultiU24X24toH16(step_rate, deceleration_time, current_block->acceleration_rate);
MultiU24X32toH16(step_rate, deceleration_time, current_block->acceleration_rate);
if (step_rate > acc_step_rate) { // Check step_rate stays positive
step_rate = current_block->final_rate;