1
0
mirror of https://github.com/MarlinFirmware/Marlin.git synced 2024-11-29 06:46:42 +00:00
MarlinFirmware/Marlin/src/module/planner.h
2020-05-12 00:10:04 -05:00

957 lines
34 KiB
C++

/**
* Marlin 3D Printer Firmware
* Copyright (c) 2020 MarlinFirmware [https://github.com/MarlinFirmware/Marlin]
*
* Based on Sprinter and grbl.
* Copyright (c) 2011 Camiel Gubbels / Erik van der Zalm
*
* 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/>.
*
*/
#pragma once
/**
* planner.h
*
* Buffer movement commands and manage the acceleration profile plan
*
* Derived from Grbl
* Copyright (c) 2009-2011 Simen Svale Skogsrud
*/
#include "../MarlinCore.h"
#if HAS_JUNCTION_DEVIATION
// Enable this option for perfect accuracy but maximum
// computation. Should be fine on ARM processors.
//#define JD_USE_MATH_ACOS
// Disable this option to save 120 bytes of PROGMEM,
// but incur increased computation and a reduction
// in accuracy.
#define JD_USE_LOOKUP_TABLE
#endif
#include "motion.h"
#include "../gcode/queue.h"
#if ENABLED(DELTA)
#include "delta.h"
#endif
#if ABL_PLANAR
#include "../libs/vector_3.h" // for matrix_3x3
#endif
#if ENABLED(FWRETRACT)
#include "../feature/fwretract.h"
#endif
#if ENABLED(MIXING_EXTRUDER)
#include "../feature/mixing.h"
#endif
#if HAS_CUTTER
#include "../feature/spindle_laser_types.h"
#endif
#if ENABLED(DIRECT_STEPPING)
#include "../feature/direct_stepping.h"
#define IS_PAGE(B) TEST(B->flag, BLOCK_BIT_IS_PAGE)
#else
#define IS_PAGE(B) false
#endif
// Feedrate for manual moves
#ifdef MANUAL_FEEDRATE
constexpr xyze_feedrate_t _mf = MANUAL_FEEDRATE,
manual_feedrate_mm_s { _mf.x / 60.0f, _mf.y / 60.0f, _mf.z / 60.0f, _mf.e / 60.0f };
#endif
#if IS_KINEMATIC && HAS_JUNCTION_DEVIATION
#define HAS_DIST_MM_ARG 1
#endif
enum BlockFlagBit : char {
// Recalculate trapezoids on entry junction. For optimization.
BLOCK_BIT_RECALCULATE,
// Nominal speed always reached.
// i.e., The segment is long enough, so the nominal speed is reachable if accelerating
// from a safe speed (in consideration of jerking from zero speed).
BLOCK_BIT_NOMINAL_LENGTH,
// The block is segment 2+ of a longer move
BLOCK_BIT_CONTINUED,
// Sync the stepper counts from the block
BLOCK_BIT_SYNC_POSITION
// Direct stepping page
#if ENABLED(DIRECT_STEPPING)
, BLOCK_BIT_IS_PAGE
#endif
};
enum BlockFlag : char {
BLOCK_FLAG_RECALCULATE = _BV(BLOCK_BIT_RECALCULATE)
, BLOCK_FLAG_NOMINAL_LENGTH = _BV(BLOCK_BIT_NOMINAL_LENGTH)
, BLOCK_FLAG_CONTINUED = _BV(BLOCK_BIT_CONTINUED)
, BLOCK_FLAG_SYNC_POSITION = _BV(BLOCK_BIT_SYNC_POSITION)
#if ENABLED(DIRECT_STEPPING)
, BLOCK_FLAG_IS_PAGE = _BV(BLOCK_BIT_IS_PAGE)
#endif
};
#if ENABLED(LASER_POWER_INLINE)
typedef struct {
uint8_t status, // See planner settings for meaning
power; // Ditto; When in trapezoid mode this is nominal power
#if ENABLED(LASER_POWER_INLINE_TRAPEZOID)
uint8_t power_entry; // Entry power for the laser
#if DISABLED(LASER_POWER_INLINE_TRAPEZOID_CONT)
uint8_t power_exit; // Exit power for the laser
uint32_t entry_per, // Steps per power increment (to avoid floats in stepper calcs)
exit_per; // Steps per power decrement
#endif
#endif
} block_laser_t;
#endif
/**
* struct block_t
*
* A single entry in the planner buffer.
* Tracks linear movement over multiple axes.
*
* The "nominal" values are as-specified by gcode, and
* may never actually be reached due to acceleration limits.
*/
typedef struct block_t {
volatile uint8_t flag; // Block flags (See BlockFlag enum above) - Modified by ISR and main thread!
// Fields used by the motion planner to manage acceleration
float nominal_speed_sqr, // The nominal speed for this block in (mm/sec)^2
entry_speed_sqr, // Entry speed at previous-current junction in (mm/sec)^2
max_entry_speed_sqr, // Maximum allowable junction entry speed in (mm/sec)^2
millimeters, // The total travel of this block in mm
acceleration; // acceleration mm/sec^2
union {
abce_ulong_t steps; // Step count along each axis
abce_long_t position; // New position to force when this sync block is executed
};
uint32_t step_event_count; // The number of step events required to complete this block
#if EXTRUDERS > 1
uint8_t extruder; // The extruder to move (if E move)
#else
static constexpr uint8_t extruder = 0;
#endif
TERN_(MIXING_EXTRUDER, MIXER_BLOCK_FIELD); // Normalized color for the mixing steppers
// Settings for the trapezoid generator
uint32_t accelerate_until, // The index of the step event on which to stop acceleration
decelerate_after; // The index of the step event on which to start decelerating
#if ENABLED(S_CURVE_ACCELERATION)
uint32_t cruise_rate, // The actual cruise rate to use, between end of the acceleration phase and start of deceleration phase
acceleration_time, // Acceleration time and deceleration time in STEP timer counts
deceleration_time,
acceleration_time_inverse, // Inverse of acceleration and deceleration periods, expressed as integer. Scale depends on CPU being used
deceleration_time_inverse;
#else
uint32_t acceleration_rate; // The acceleration rate used for acceleration calculation
#endif
uint8_t direction_bits; // The direction bit set for this block (refers to *_DIRECTION_BIT in config.h)
// Advance extrusion
#if ENABLED(LIN_ADVANCE)
bool use_advance_lead;
uint16_t advance_speed, // STEP timer value for extruder speed offset ISR
max_adv_steps, // max. advance steps to get cruising speed pressure (not always nominal_speed!)
final_adv_steps; // advance steps due to exit speed
float e_D_ratio;
#endif
uint32_t nominal_rate, // The nominal step rate for this block in step_events/sec
initial_rate, // The jerk-adjusted step rate at start of block
final_rate, // The minimal rate at exit
acceleration_steps_per_s2; // acceleration steps/sec^2
#if ENABLED(DIRECT_STEPPING)
page_idx_t page_idx; // Page index used for direct stepping
#endif
#if HAS_CUTTER
cutter_power_t cutter_power; // Power level for Spindle, Laser, etc.
#endif
#if HAS_FAN
uint8_t fan_speed[FAN_COUNT];
#endif
#if ENABLED(BARICUDA)
uint8_t valve_pressure, e_to_p_pressure;
#endif
#if HAS_SPI_LCD
uint32_t segment_time_us;
#endif
#if ENABLED(POWER_LOSS_RECOVERY)
uint32_t sdpos;
#endif
#if ENABLED(LASER_POWER_INLINE)
block_laser_t laser;
#endif
} block_t;
#if ANY(LIN_ADVANCE, SCARA_FEEDRATE_SCALING, GRADIENT_MIX, LCD_SHOW_E_TOTAL)
#define HAS_POSITION_FLOAT 1
#endif
#define BLOCK_MOD(n) ((n)&(BLOCK_BUFFER_SIZE-1))
#if ENABLED(LASER_POWER_INLINE)
typedef struct {
/**
* Laser status bitmask; most bits are unused;
* 0: Planner buffer enable
* 1: Laser enable
* 2: Reserved for direction
*/
uint8_t status;
/**
* Laser power: 0 or 255 in case of PWM-less laser,
* or the OCR value;
*
* Using OCR instead of raw power,
* as it avoids floating points during move loop
*/
uint8_t power;
} laser_state_t;
#endif
typedef struct {
uint32_t max_acceleration_mm_per_s2[XYZE_N], // (mm/s^2) M201 XYZE
min_segment_time_us; // (µs) M205 B
float axis_steps_per_mm[XYZE_N]; // (steps) M92 XYZE - Steps per millimeter
feedRate_t max_feedrate_mm_s[XYZE_N]; // (mm/s) M203 XYZE - Max speeds
float acceleration, // (mm/s^2) M204 S - Normal acceleration. DEFAULT ACCELERATION for all printing moves.
retract_acceleration, // (mm/s^2) M204 R - Retract acceleration. Filament pull-back and push-forward while standing still in the other axes
travel_acceleration; // (mm/s^2) M204 T - Travel acceleration. DEFAULT ACCELERATION for all NON printing moves.
feedRate_t min_feedrate_mm_s, // (mm/s) M205 S - Minimum linear feedrate
min_travel_feedrate_mm_s; // (mm/s) M205 T - Minimum travel feedrate
} planner_settings_t;
#if DISABLED(SKEW_CORRECTION)
#define XY_SKEW_FACTOR 0
#define XZ_SKEW_FACTOR 0
#define YZ_SKEW_FACTOR 0
#endif
typedef struct {
#if ENABLED(SKEW_CORRECTION_GCODE)
float xy;
#if ENABLED(SKEW_CORRECTION_FOR_Z)
float xz, yz;
#else
const float xz = XZ_SKEW_FACTOR, yz = YZ_SKEW_FACTOR;
#endif
#else
const float xy = XY_SKEW_FACTOR,
xz = XZ_SKEW_FACTOR, yz = YZ_SKEW_FACTOR;
#endif
} skew_factor_t;
class Planner {
public:
/**
* The move buffer, calculated in stepper steps
*
* block_buffer is a ring buffer...
*
* head,tail : indexes for write,read
* head==tail : the buffer is empty
* head!=tail : blocks are in the buffer
* head==(tail-1)%size : the buffer is full
*
* Writer of head is Planner::buffer_segment().
* Reader of tail is Stepper::isr(). Always consider tail busy / read-only
*/
static block_t block_buffer[BLOCK_BUFFER_SIZE];
static volatile uint8_t block_buffer_head, // Index of the next block to be pushed
block_buffer_nonbusy, // Index of the first non busy block
block_buffer_planned, // Index of the optimally planned block
block_buffer_tail; // Index of the busy block, if any
static uint16_t cleaning_buffer_counter; // A counter to disable queuing of blocks
static uint8_t delay_before_delivering; // This counter delays delivery of blocks when queue becomes empty to allow the opportunity of merging blocks
#if ENABLED(DISTINCT_E_FACTORS)
static uint8_t last_extruder; // Respond to extruder change
#endif
#if ENABLED(DIRECT_STEPPING)
static uint32_t last_page_step_rate; // Last page step rate given
static xyze_bool_t last_page_dir; // Last page direction given
#endif
#if EXTRUDERS
static int16_t flow_percentage[EXTRUDERS]; // Extrusion factor for each extruder
static float e_factor[EXTRUDERS]; // The flow percentage and volumetric multiplier combine to scale E movement
#endif
#if DISABLED(NO_VOLUMETRICS)
static float filament_size[EXTRUDERS], // diameter of filament (in millimeters), typically around 1.75 or 2.85, 0 disables the volumetric calculations for the extruder
volumetric_area_nominal, // Nominal cross-sectional area
volumetric_multiplier[EXTRUDERS]; // Reciprocal of cross-sectional area of filament (in mm^2). Pre-calculated to reduce computation in the planner
// May be auto-adjusted by a filament width sensor
#endif
static planner_settings_t settings;
#if ENABLED(LASER_POWER_INLINE)
static laser_state_t laser;
#endif
static uint32_t max_acceleration_steps_per_s2[XYZE_N]; // (steps/s^2) Derived from mm_per_s2
static float steps_to_mm[XYZE_N]; // Millimeters per step
#if HAS_JUNCTION_DEVIATION
static float junction_deviation_mm; // (mm) M205 J
#if HAS_LINEAR_E_JERK
static float max_e_jerk[DISTINCT_E]; // Calculated from junction_deviation_mm
#endif
#endif
#if HAS_CLASSIC_JERK
// (mm/s^2) M205 XYZ(E) - The largest speed change requiring no acceleration.
static TERN(HAS_LINEAR_E_JERK, xyz_pos_t, xyze_pos_t) max_jerk;
#endif
#if HAS_LEVELING
static bool leveling_active; // Flag that bed leveling is enabled
#if ABL_PLANAR
static matrix_3x3 bed_level_matrix; // Transform to compensate for bed level
#endif
#if ENABLED(ENABLE_LEVELING_FADE_HEIGHT)
static float z_fade_height, inverse_z_fade_height;
#endif
#else
static constexpr bool leveling_active = false;
#endif
#if ENABLED(LIN_ADVANCE)
static float extruder_advance_K[EXTRUDERS];
#endif
/**
* The current position of the tool in absolute steps
* Recalculated if any axis_steps_per_mm are changed by gcode
*/
static xyze_long_t position;
#if HAS_POSITION_FLOAT
static xyze_pos_t position_float;
#endif
#if IS_KINEMATIC
static xyze_pos_t position_cart;
#endif
static skew_factor_t skew_factor;
#if ENABLED(SD_ABORT_ON_ENDSTOP_HIT)
static bool abort_on_endstop_hit;
#endif
#ifdef XY_FREQUENCY_LIMIT
static int8_t xy_freq_limit_hz; // Minimum XY frequency setting
static float xy_freq_min_speed_factor; // Minimum speed factor setting
static int32_t xy_freq_min_interval_us; // Minimum segment time based on xy_freq_limit_hz
static inline void refresh_frequency_limit() {
//xy_freq_min_interval_us = xy_freq_limit_hz ?: LROUND(1000000.0f / xy_freq_limit_hz);
if (xy_freq_limit_hz)
xy_freq_min_interval_us = LROUND(1000000.0f / xy_freq_limit_hz);
}
static inline void set_min_speed_factor_u8(const uint8_t v255) {
xy_freq_min_speed_factor = float(ui8_to_percent(v255)) / 100;
}
static inline void set_frequency_limit(const uint8_t hz) {
xy_freq_limit_hz = constrain(hz, 0, 100);
refresh_frequency_limit();
}
#endif
private:
/**
* Speed of previous path line segment
*/
static xyze_float_t previous_speed;
/**
* Nominal speed of previous path line segment (mm/s)^2
*/
static float previous_nominal_speed_sqr;
/**
* Limit where 64bit math is necessary for acceleration calculation
*/
static uint32_t cutoff_long;
#if ENABLED(ENABLE_LEVELING_FADE_HEIGHT)
static float last_fade_z;
#endif
#if ENABLED(DISABLE_INACTIVE_EXTRUDER)
// Counters to manage disabling inactive extruders
static uint8_t g_uc_extruder_last_move[EXTRUDERS];
#endif
#if HAS_SPI_LCD
volatile static uint32_t block_buffer_runtime_us; // Theoretical block buffer runtime in µs
#endif
public:
/**
* Instance Methods
*/
Planner();
void init();
/**
* Static (class) Methods
*/
static void reset_acceleration_rates();
static void refresh_positioning();
static void set_max_acceleration(const uint8_t axis, float targetValue);
static void set_max_feedrate(const uint8_t axis, float targetValue);
static void set_max_jerk(const AxisEnum axis, float targetValue);
#if EXTRUDERS
FORCE_INLINE static void refresh_e_factor(const uint8_t e) {
e_factor[e] = flow_percentage[e] * 0.01f * TERN(NO_VOLUMETRICS, 1.0f, volumetric_multiplier[e]);
}
static inline void set_flow(const uint8_t e, const int16_t flow) {
flow_percentage[e] = flow;
refresh_e_factor(e);
}
#endif
// Manage fans, paste pressure, etc.
static void check_axes_activity();
// Update multipliers based on new diameter measurements
static void calculate_volumetric_multipliers();
#if ENABLED(FILAMENT_WIDTH_SENSOR)
void apply_filament_width_sensor(const int8_t encoded_ratio);
static inline float volumetric_percent(const bool vol) {
return 100.0f * (vol
? volumetric_area_nominal / volumetric_multiplier[FILAMENT_SENSOR_EXTRUDER_NUM]
: volumetric_multiplier[FILAMENT_SENSOR_EXTRUDER_NUM]
);
}
#endif
#if DISABLED(NO_VOLUMETRICS)
FORCE_INLINE static void set_filament_size(const uint8_t e, const float &v) {
filament_size[e] = v;
// make sure all extruders have some sane value for the filament size
LOOP_L_N(i, COUNT(filament_size))
if (!filament_size[i]) filament_size[i] = DEFAULT_NOMINAL_FILAMENT_DIA;
}
#endif
#if ENABLED(ENABLE_LEVELING_FADE_HEIGHT)
/**
* Get the Z leveling fade factor based on the given Z height,
* re-calculating only when needed.
*
* Returns 1.0 if planner.z_fade_height is 0.0.
* Returns 0.0 if Z is past the specified 'Fade Height'.
*/
static inline float fade_scaling_factor_for_z(const float &rz) {
static float z_fade_factor = 1;
if (!z_fade_height) return 1;
if (rz >= z_fade_height) return 0;
if (last_fade_z != rz) {
last_fade_z = rz;
z_fade_factor = 1 - rz * inverse_z_fade_height;
}
return z_fade_factor;
}
FORCE_INLINE static void force_fade_recalc() { last_fade_z = -999.999f; }
FORCE_INLINE static void set_z_fade_height(const float &zfh) {
z_fade_height = zfh > 0 ? zfh : 0;
inverse_z_fade_height = RECIPROCAL(z_fade_height);
force_fade_recalc();
}
FORCE_INLINE static bool leveling_active_at_z(const float &rz) {
return !z_fade_height || rz < z_fade_height;
}
#else
FORCE_INLINE static float fade_scaling_factor_for_z(const float&) { return 1; }
FORCE_INLINE static bool leveling_active_at_z(const float&) { return true; }
#endif
#if ENABLED(SKEW_CORRECTION)
FORCE_INLINE static void skew(float &cx, float &cy, const float &cz) {
if (WITHIN(cx, X_MIN_POS + 1, X_MAX_POS) && WITHIN(cy, Y_MIN_POS + 1, Y_MAX_POS)) {
const float sx = cx - cy * skew_factor.xy - cz * (skew_factor.xz - (skew_factor.xy * skew_factor.yz)),
sy = cy - cz * skew_factor.yz;
if (WITHIN(sx, X_MIN_POS, X_MAX_POS) && WITHIN(sy, Y_MIN_POS, Y_MAX_POS)) {
cx = sx; cy = sy;
}
}
}
FORCE_INLINE static void skew(xyz_pos_t &raw) { skew(raw.x, raw.y, raw.z); }
FORCE_INLINE static void unskew(float &cx, float &cy, const float &cz) {
if (WITHIN(cx, X_MIN_POS, X_MAX_POS) && WITHIN(cy, Y_MIN_POS, Y_MAX_POS)) {
const float sx = cx + cy * skew_factor.xy + cz * skew_factor.xz,
sy = cy + cz * skew_factor.yz;
if (WITHIN(sx, X_MIN_POS, X_MAX_POS) && WITHIN(sy, Y_MIN_POS, Y_MAX_POS)) {
cx = sx; cy = sy;
}
}
}
FORCE_INLINE static void unskew(xyz_pos_t &raw) { unskew(raw.x, raw.y, raw.z); }
#endif // SKEW_CORRECTION
#if HAS_LEVELING
/**
* Apply leveling to transform a cartesian position
* as it will be given to the planner and steppers.
*/
static void apply_leveling(xyz_pos_t &raw);
static void unapply_leveling(xyz_pos_t &raw);
FORCE_INLINE static void force_unapply_leveling(xyz_pos_t &raw) {
leveling_active = true;
unapply_leveling(raw);
leveling_active = false;
}
#else
FORCE_INLINE static void apply_leveling(xyz_pos_t&) {}
FORCE_INLINE static void unapply_leveling(xyz_pos_t&) {}
#endif
#if ENABLED(FWRETRACT)
static void apply_retract(float &rz, float &e);
FORCE_INLINE static void apply_retract(xyze_pos_t &raw) { apply_retract(raw.z, raw.e); }
static void unapply_retract(float &rz, float &e);
FORCE_INLINE static void unapply_retract(xyze_pos_t &raw) { unapply_retract(raw.z, raw.e); }
#endif
#if HAS_POSITION_MODIFIERS
FORCE_INLINE static void apply_modifiers(xyze_pos_t &pos, bool leveling=ENABLED(PLANNER_LEVELING)) {
TERN_(SKEW_CORRECTION, skew(pos));
if (leveling) apply_leveling(pos);
TERN_(FWRETRACT, apply_retract(pos));
}
FORCE_INLINE static void unapply_modifiers(xyze_pos_t &pos, bool leveling=ENABLED(PLANNER_LEVELING)) {
TERN_(FWRETRACT, unapply_retract(pos));
if (leveling) unapply_leveling(pos);
TERN_(SKEW_CORRECTION, unskew(pos));
}
#endif // HAS_POSITION_MODIFIERS
// Number of moves currently in the planner including the busy block, if any
FORCE_INLINE static uint8_t movesplanned() { return BLOCK_MOD(block_buffer_head - block_buffer_tail); }
// Number of nonbusy moves currently in the planner
FORCE_INLINE static uint8_t nonbusy_movesplanned() { return BLOCK_MOD(block_buffer_head - block_buffer_nonbusy); }
// Remove all blocks from the buffer
FORCE_INLINE static void clear_block_buffer() { block_buffer_nonbusy = block_buffer_planned = block_buffer_head = block_buffer_tail = 0; }
// Check if movement queue is full
FORCE_INLINE static bool is_full() { return block_buffer_tail == next_block_index(block_buffer_head); }
// Get count of movement slots free
FORCE_INLINE static uint8_t moves_free() { return BLOCK_BUFFER_SIZE - 1 - movesplanned(); }
/**
* Planner::get_next_free_block
*
* - Get the next head indices (passed by reference)
* - Wait for the number of spaces to open up in the planner
* - Return the first head block
*/
FORCE_INLINE static block_t* get_next_free_block(uint8_t &next_buffer_head, const uint8_t count=1) {
// Wait until there are enough slots free
while (moves_free() < count) { idle(); }
// Return the first available block
next_buffer_head = next_block_index(block_buffer_head);
return &block_buffer[block_buffer_head];
}
/**
* Planner::_buffer_steps
*
* Add a new linear movement to the buffer (in terms of steps).
*
* target - target position in steps units
* fr_mm_s - (target) speed of the move
* extruder - target extruder
* millimeters - the length of the movement, if known
*
* Returns true if movement was buffered, false otherwise
*/
static bool _buffer_steps(const xyze_long_t &target
#if HAS_POSITION_FLOAT
, const xyze_pos_t &target_float
#endif
#if HAS_DIST_MM_ARG
, const xyze_float_t &cart_dist_mm
#endif
, feedRate_t fr_mm_s, const uint8_t extruder, const float &millimeters=0.0
);
/**
* Planner::_populate_block
*
* Fills a new linear movement in the block (in terms of steps).
*
* target - target position in steps units
* fr_mm_s - (target) speed of the move
* extruder - target extruder
* millimeters - the length of the movement, if known
*
* Returns true is movement is acceptable, false otherwise
*/
static bool _populate_block(block_t * const block, bool split_move,
const xyze_long_t &target
#if HAS_POSITION_FLOAT
, const xyze_pos_t &target_float
#endif
#if HAS_DIST_MM_ARG
, const xyze_float_t &cart_dist_mm
#endif
, feedRate_t fr_mm_s, const uint8_t extruder, const float &millimeters=0.0
);
/**
* Planner::buffer_sync_block
* Add a block to the buffer that just updates the position
*/
static void buffer_sync_block();
#if IS_KINEMATIC
private:
// Allow do_homing_move to access internal functions, such as buffer_segment.
friend void do_homing_move(const AxisEnum, const float, const feedRate_t);
#endif
/**
* Planner::buffer_segment
*
* Add a new linear movement to the buffer in axis units.
*
* Leveling and kinematics should be applied ahead of calling this.
*
* a,b,c,e - target positions in mm and/or degrees
* fr_mm_s - (target) speed of the move
* extruder - target extruder
* millimeters - the length of the movement, if known
*/
static bool buffer_segment(const float &a, const float &b, const float &c, const float &e
#if HAS_DIST_MM_ARG
, const xyze_float_t &cart_dist_mm
#endif
, const feedRate_t &fr_mm_s, const uint8_t extruder, const float &millimeters=0.0
);
FORCE_INLINE static bool buffer_segment(abce_pos_t &abce
#if HAS_DIST_MM_ARG
, const xyze_float_t &cart_dist_mm
#endif
, const feedRate_t &fr_mm_s, const uint8_t extruder, const float &millimeters=0.0
) {
return buffer_segment(abce.a, abce.b, abce.c, abce.e
#if HAS_DIST_MM_ARG
, cart_dist_mm
#endif
, fr_mm_s, extruder, millimeters);
}
public:
/**
* Add a new linear movement to the buffer.
* The target is cartesian. It's translated to
* delta/scara if needed.
*
* rx,ry,rz,e - target position in mm or degrees
* fr_mm_s - (target) speed of the move (mm/s)
* extruder - target extruder
* millimeters - the length of the movement, if known
* inv_duration - the reciprocal if the duration of the movement, if known (kinematic only if feeedrate scaling is enabled)
*/
static bool buffer_line(const float &rx, const float &ry, const float &rz, const float &e, const feedRate_t &fr_mm_s, const uint8_t extruder, const float millimeters=0.0
#if ENABLED(SCARA_FEEDRATE_SCALING)
, const float &inv_duration=0.0
#endif
);
FORCE_INLINE static bool buffer_line(const xyze_pos_t &cart, const feedRate_t &fr_mm_s, const uint8_t extruder, const float millimeters=0.0
#if ENABLED(SCARA_FEEDRATE_SCALING)
, const float &inv_duration=0.0
#endif
) {
return buffer_line(cart.x, cart.y, cart.z, cart.e, fr_mm_s, extruder, millimeters
#if ENABLED(SCARA_FEEDRATE_SCALING)
, inv_duration
#endif
);
}
#if ENABLED(DIRECT_STEPPING)
static void buffer_page(const page_idx_t page_idx, const uint8_t extruder, const uint16_t num_steps);
#endif
/**
* Set the planner.position and individual stepper positions.
* Used by G92, G28, G29, and other procedures.
*
* The supplied position is in the cartesian coordinate space and is
* translated in to machine space as needed. Modifiers such as leveling
* and skew are also applied.
*
* Multiplies by axis_steps_per_mm[] and does necessary conversion
* for COREXY / COREXZ / COREYZ to set the corresponding stepper positions.
*
* Clears previous speed values.
*/
static void set_position_mm(const float &rx, const float &ry, const float &rz, const float &e);
FORCE_INLINE static void set_position_mm(const xyze_pos_t &cart) { set_position_mm(cart.x, cart.y, cart.z, cart.e); }
static void set_e_position_mm(const float &e);
/**
* Set the planner.position and individual stepper positions.
*
* The supplied position is in machine space, and no additional
* conversions are applied.
*/
static void set_machine_position_mm(const float &a, const float &b, const float &c, const float &e);
FORCE_INLINE static void set_machine_position_mm(const abce_pos_t &abce) { set_machine_position_mm(abce.a, abce.b, abce.c, abce.e); }
/**
* Get an axis position according to stepper position(s)
* For CORE machines apply translation from ABC to XYZ.
*/
static float get_axis_position_mm(const AxisEnum axis);
static inline abce_pos_t get_axis_positions_mm() {
const abce_pos_t out = {
get_axis_position_mm(A_AXIS),
get_axis_position_mm(B_AXIS),
get_axis_position_mm(C_AXIS),
get_axis_position_mm(E_AXIS)
};
return out;
}
// SCARA AB axes are in degrees, not mm
#if IS_SCARA
FORCE_INLINE static float get_axis_position_degrees(const AxisEnum axis) { return get_axis_position_mm(axis); }
#endif
// Called to force a quick stop of the machine (for example, when
// a Full Shutdown is required, or when endstops are hit)
static void quick_stop();
// Called when an endstop is triggered. Causes the machine to stop inmediately
static void endstop_triggered(const AxisEnum axis);
// Triggered position of an axis in mm (not core-savvy)
static float triggered_position_mm(const AxisEnum axis);
// Block until all buffered steps are executed / cleaned
static void synchronize();
// Wait for moves to finish and disable all steppers
static void finish_and_disable();
// Periodic tick to handle cleaning timeouts
// Called from the Temperature ISR at ~1kHz
static void tick() {
if (cleaning_buffer_counter) --cleaning_buffer_counter;
}
/**
* Does the buffer have any blocks queued?
*/
FORCE_INLINE static bool has_blocks_queued() { return (block_buffer_head != block_buffer_tail); }
/**
* Get the current block for processing
* and mark the block as busy.
* Return nullptr if the buffer is empty
* or if there is a first-block delay.
*
* WARNING: Called from Stepper ISR context!
*/
static block_t* get_current_block();
/**
* "Release" the current block so its slot can be reused.
* Called when the current block is no longer needed.
*/
FORCE_INLINE static void release_current_block() {
if (has_blocks_queued())
block_buffer_tail = next_block_index(block_buffer_tail);
}
#if HAS_SPI_LCD
static uint16_t block_buffer_runtime();
static void clear_block_buffer_runtime();
#endif
#if ENABLED(AUTOTEMP)
static float autotemp_min, autotemp_max, autotemp_factor;
static bool autotemp_enabled;
static void getHighESpeed();
static void autotemp_M104_M109();
static void autotemp_update();
#endif
#if HAS_LINEAR_E_JERK
FORCE_INLINE static void recalculate_max_e_jerk() {
const float prop = junction_deviation_mm * SQRT(0.5) / (1.0f - SQRT(0.5));
LOOP_L_N(i, EXTRUDERS)
max_e_jerk[E_INDEX_N(i)] = SQRT(prop * settings.max_acceleration_mm_per_s2[E_INDEX_N(i)]);
}
#endif
private:
/**
* Get the index of the next / previous block in the ring buffer
*/
static constexpr uint8_t next_block_index(const uint8_t block_index) { return BLOCK_MOD(block_index + 1); }
static constexpr uint8_t prev_block_index(const uint8_t block_index) { return BLOCK_MOD(block_index - 1); }
/**
* Calculate the distance (not time) it takes to accelerate
* from initial_rate to target_rate using the given acceleration:
*/
static float estimate_acceleration_distance(const float &initial_rate, const float &target_rate, const float &accel) {
if (accel == 0) return 0; // accel was 0, set acceleration distance to 0
return (sq(target_rate) - sq(initial_rate)) / (accel * 2);
}
/**
* Return the point at which you must start braking (at the rate of -'accel') if
* you start at 'initial_rate', accelerate (until reaching the point), and want to end at
* 'final_rate' after traveling 'distance'.
*
* This is used to compute the intersection point between acceleration and deceleration
* in cases where the "trapezoid" has no plateau (i.e., never reaches maximum speed)
*/
static float intersection_distance(const float &initial_rate, const float &final_rate, const float &accel, const float &distance) {
if (accel == 0) return 0; // accel was 0, set intersection distance to 0
return (accel * 2 * distance - sq(initial_rate) + sq(final_rate)) / (accel * 4);
}
/**
* Calculate the maximum allowable speed squared at this point, in order
* to reach 'target_velocity_sqr' using 'acceleration' within a given
* 'distance'.
*/
static float max_allowable_speed_sqr(const float &accel, const float &target_velocity_sqr, const float &distance) {
return target_velocity_sqr - 2 * accel * distance;
}
#if ENABLED(S_CURVE_ACCELERATION)
/**
* Calculate the speed reached given initial speed, acceleration and distance
*/
static float final_speed(const float &initial_velocity, const float &accel, const float &distance) {
return SQRT(sq(initial_velocity) + 2 * accel * distance);
}
#endif
static void calculate_trapezoid_for_block(block_t* const block, const float &entry_factor, const float &exit_factor);
static void reverse_pass_kernel(block_t* const current, const block_t * const next);
static void forward_pass_kernel(const block_t * const previous, block_t* const current, uint8_t block_index);
static void reverse_pass();
static void forward_pass();
static void recalculate_trapezoids();
static void recalculate();
#if HAS_JUNCTION_DEVIATION
FORCE_INLINE static void normalize_junction_vector(xyze_float_t &vector) {
float magnitude_sq = 0;
LOOP_XYZE(idx) if (vector[idx]) magnitude_sq += sq(vector[idx]);
vector *= RSQRT(magnitude_sq);
}
FORCE_INLINE static float limit_value_by_axis_maximum(const float &max_value, xyze_float_t &unit_vec) {
float limit_value = max_value;
LOOP_XYZE(idx) {
if (unit_vec[idx]) {
if (limit_value * ABS(unit_vec[idx]) > settings.max_acceleration_mm_per_s2[idx])
limit_value = ABS(settings.max_acceleration_mm_per_s2[idx] / unit_vec[idx]);
}
}
return limit_value;
}
#endif // !CLASSIC_JERK
};
#define PLANNER_XY_FEEDRATE() (_MIN(planner.settings.max_feedrate_mm_s[X_AXIS], planner.settings.max_feedrate_mm_s[Y_AXIS]))
extern Planner planner;