Commit Graph

334 Commits

Author SHA1 Message Date
Sonny J
110faae986 More '%' modulo opertor removals and some housecleaning.
- Serial functions contained quite a few modulo operations that would
be executed with high frequency when streaming. AVR processors are very
slow when operating these. In one test on the Arduino forums, it showed
about a 15x slow down compared to a simple if-then statement. -
Clarified some variable names and types and comments.
2011-09-15 20:32:15 -06:00
Sonny J
4d03c4febc Further planner improvements and misc minor bug fixes. Memory savings and increased buffer size.
- Update grbl version and settings version to automatically reset
eeprom. FYI, this will reset your grbl settings. - Saved
3*BLOCK_BUFFER_SIZE doubles in static memory by removing obsolete
variables: speed_x, speed_y, and speed_z. - Increased buffer size
conservatively to 18 from 16. (Probably can do 20). - Removed expensive!
modulo operator from block indexing function. Reduces significant
computational overhead. - Re-organized some sqrt() calls to be more
efficient during time critical planning cases, rather than non-time
critical. - Minor bug fix in planner max junction velocity logic. -
Simplified arc logic and removed need to multiply for CW or CCW
direction.
2011-09-13 21:57:16 -06:00
Sonny J
ffcc3470a3 Optimized planner re-write. Significantly faster. Full arc support enabled by rotation matrix approach.
- Significant improvements in the planner. Removed or reordered
repetitive and expensive calculations by order of importance:
recalculating unchanged blocks, trig functions [sin(), cos(), tan()],
sqrt(), divides, and multiplications. Blocks long enough for nominal
speed to be guaranteed to be reached ignored by planner. Done by
introducing two uint8_t flags per block. Reduced computational overhead
by an order of magnitude.   - Arc motion generation completely
re-written and optimized. Now runs with acceleration planner. Removed
all but one trig function (atan2) from initialization. Streamlined
computations. Segment target locations generated by vector
transformation and small angle approximation. Arc path correction
implemented for accumulated error of approximation and single precision
calculation of Arduino. Bug fix in message passing.
2011-09-06 19:39:14 -06:00
Sonny J
d75ad82e49 Minor update for memory savings in ring buffer and fleshed out commenting.
No changes in functionality. Path vectors moved from ring buffer to
local planner static variables to save 3*(BUFFER_SIZE - 1) doubles in
memory. Detailed comments. Really need to stop micro-updating. Should be
the last until a planner optimization (ala Jens Geisler) has been
completed.
2011-09-04 18:53:25 -06:00
Sonny J
f1e5ff35ec More minor bug fixes in planner.
Reverse planner was over-writing the initial/buffer tail entry speed,
which reset the forward planner and caused it to lose track of its
speed. Should now accelerate into short linear segments much nicer now.
2011-09-04 11:19:08 -06:00
Sonny J
5e2e935bda Minor bug fixes in planner. 2011-09-03 23:22:27 -06:00
Sonny J
75bd4c5ac3 Add G02/03 arc conversion/pre-processor script and example streaming script
Beta pre-processor script used to clean and streamline g-code for
streaming and converts G02/03 arcs into linear segments. Allows for full
acceleration support if the pre-processed g-code is then streamed to
grill, sans G02/03 arcs. Added a simple example streaming script for
Python users.
2011-09-03 16:08:42 -06:00
Sonny J
5c2150daa9 Significantly improved junction control and fixed computation bugs in planner
- Junction jerk now re-defined as junction_deviation. The distance from
the junction to the edge of a circle tangent to both previous and
current path lines. The circle radii is used to compute the maximum
junction velocity by centripetal acceleration. More robust and
simplified way to compute jerk.   - Fixed bugs related to entry and exit
factors. They were computed based on the current nominal speeds but not
when computing exit factors for neighboring blocks. Removed factors and
replaced with entry speeds only. Factors now only computed for stepper
trapezoid rate conversions.  - Misc: Added min(), next_block_index,
prev_block_index functions for clarity.
2011-09-03 15:31:48 -06:00
Sonny J
badb638df9 Moved comment and block delete handling into protocol.c from gcode.c. Fixes bug when comment and block delete are not isolated. Blank lines ignored.
Comments, block delete characters, and blank lines are no longer passed
to the gcode parser and should free up some memory by ignoring these
characters. Gcode parser now expects clean gcode only. There was a bug
if there were block deletes or comments not in the first character (i.e.
spindle on/off for proofing geode without turning it on, or a NXX
followed by a comment). This should fix it by bypassing the problem.
Left a commented line for easily turning on and off block deletes for a
later feature, if desired.
2011-08-15 19:39:44 -06:00
Sonny J
ed5e5d1181 Revert ea5b8942db2616e93fc0478922010c3bab7c0481^..HEAD 2011-08-15 19:37:22 -06:00
Sonny J
896a6b9199 Moved comment and block delete handling into protocol.c from gcode.c. Fixes bug when comment and block delete are not isolated. Blank lines ignored.
Comments, block delete characters, and blank lines are no longer passed
to the gcode parser and should free up some memory by ignoring these
characters. Gcode parser now expects clean gcode only. There was a bug
if there were block deletes or comments not in the first character (i.e.
spindle on/off for proofing geode without turning it on, or a NXX
followed by a comment). This should fix it by bypassing the problem.
Left a commented line for easily turning on and off block deletes for a
later feature, if desired.
2011-08-15 19:28:14 -06:00
Sonny J
971e50aa9a Revert 517a0f659a06182c89cafe27ee371edccad777a4^..HEAD 2011-08-15 19:15:43 -06:00
Sonny J
a2837943c0 Revert "Moved comment and block delete handling to be done in protocol.c rather than gcode.c. Prevents these from being held in memory. Also, fixes bug when comments and block delete character are mixed with g-code."
This reverts commit ea5b8942db.
2011-08-15 19:14:39 -06:00
Sonny J
8b0556bcfd Revert fdc90f1821f1f5edb7756fcddce75b4b4fbf6bbf^..HEAD 2011-08-15 19:14:25 -06:00
chamnit
fdc90f1821 Removed comment and block delete handling from gcode.c. Parser expects clean gcode. 2011-08-15 17:10:08 -06:00
chamnit
ea5b8942db Moved comment and block delete handling to be done in protocol.c rather than gcode.c. Prevents these from being held in memory. Also, fixes bug when comments and block delete character are mixed with g-code. 2011-08-15 17:06:50 -06:00
Simen Svale Skogsrud
517a0f659a reordered a couple of functions for consistency 2011-08-02 22:44:32 +02:00
Simen Svale Skogsrud
5f09dba95d Merge pull request #34 from beardicus/pulse_error_msg
Step Pulse Error Message thanks to beardicus.
2011-07-19 04:04:55 -07:00
Brian Boucheron
d3f1f82e5f added error message for step pulses < 3 microseconds 2011-07-17 13:35:44 -04:00
Simen Svale Skogsrud
5466bc0c33 fixed a nil dereferencing bug in planner_forward_pass_kernel thanks to jv4779 2011-07-10 21:54:03 +02:00
Simen Svale Skogsrud
804837a111 corrected an outdated comment 2011-07-09 22:35:25 +02:00
Simen Svale Skogsrud
a9aa7d4d39 need to bump the version number because of the serial refactoring 2011-06-06 20:00:57 +02:00
Simen Svale Skogsrud
391efe37ed avoid line buffer overflow 2011-06-06 08:38:38 +02:00
Simen Svale Skogsrud
c3fda5ac21 fixed a bug where the default step mask would not actually reflect the real pin out 2011-06-05 21:34:53 +02:00
Simen Svale Skogsrud
9e09a502e9 updated readme to reflect real state of 168-compatibility 2011-06-05 21:12:56 +02:00
Simen Svale Skogsrud
bd336867a6 changed default fp decimal places to 3 2011-06-05 20:58:21 +02:00
Simen Svale Skogsrud
553c44a93f Fixed serial.c after tests on real hardware 2011-06-05 20:54:23 +02:00
Simen Svale Skogsrud
74db3e6c5b undid changes to printIntegerInBase that resulted in the integers being printed in reverse... 2011-06-04 22:17:51 +02:00
Simen Svale Skogsrud
903303579f NOW printFloat probably works 2011-06-03 21:56:55 +02:00
Simen Svale Skogsrud
a7c13c497a a new printFloat that probably works 2011-06-03 21:56:18 +02:00
Simen Svale Skogsrud
f0843db46e refactored printIntegerInBase to work without a buffer + minor cleanup 2011-06-03 21:50:02 +02:00
Simen Svale Skogsrud
69be1240be refactored a complex comment 2011-06-03 15:43:14 +02:00
Simen Svale Skogsrud
bfe7a2ad84 flossed the serial module 2011-06-03 15:42:28 +02:00
Simen Svale Skogsrud
9488cb329c cleaned up function definitions for serial and print modules and added a comment about the deeply flawed printFloat method 2011-06-03 15:36:14 +02:00
Simen Svale Skogsrud
5eea9f4c7c pruned extraneous comment 2011-06-03 15:33:56 +02:00
Simen Svale Skogsrud
ab8bae74c1 pruned some undefined declarations from the print module 2011-06-03 15:33:12 +02:00
Simen Svale Skogsrud
9a5fed4fda changed file headers to reflect the totally refactored state of the serial and print modules 2011-06-03 15:31:59 +02:00
Simen Svale Skogsrud
c0b4b8309a cleaned up serial completing support for non blocking tx and refactoring formatting functions into a new module 'print' 2011-06-03 15:28:14 +02:00
Simen Svale Skogsrud
8793b555e0 cleaned up data types in serial module, all data now uint8_t 2011-06-01 09:45:15 +02:00
Simen Svale Skogsrud
ee3139d283 changed serialAvailable to serialAnyAvailable which does not calculate the number of bytes, only if there are any at all 2011-06-01 09:36:15 +02:00
Simen Svale Skogsrud
defabc80ed renamed wiring_serial to serial to reflect its complete remake 2011-05-31 22:45:38 +02:00
Simen Svale Skogsrud
33a940cdf8 added asynchronous write as provided by Arthur Wolf 2011-05-31 22:29:50 +02:00
Simen Svale Skogsrud
f4c1cbac3c added a fix for division by zero errors that would occur on very low feed rates (thanks to Arthur Wolf for hardcore investigation) 2011-05-31 18:37:37 +02:00
Simen Svale Skogsrud
a25812f36b made inverted stepper driving the default as per the request of the Grbl Shield team and frankly myself. Who doesn't need their stepper signals inverted anyway? 2011-05-31 13:13:45 +02:00
Simen Svale Skogsrud
c2aec12004 converted the STEPPER_ENABLE_PIN to a STEPPER_DISABLE_PIN as per the request of Alden Hart of Grbl Shield fame. 2011-05-31 13:08:42 +02:00
Simen Svale Skogsrud
74dcf58c5c simplification in gcode.c also thanks to etienne 2011-02-25 15:02:55 +01:00
Simen Svale Skogsrud
d914089e91 spindle could not be stopped. Thanks again to Etienne for reporting 2011-02-25 15:00:47 +01:00
Simen Svale Skogsrud
cd074bc590 foolproofed steps/mm setting (must be > 0) 2011-02-25 13:40:48 +01:00
Simen Svale Skogsrud
4cac11ec86 fixed synchronization problem with spindle_control 2011-02-25 13:30:03 +01:00
Simen Svale Skogsrud
a18a89c779 fixed linebreaks, was LFCR, is CRLF 2011-02-24 16:08:06 +01:00