- Overhauled the stepper algorithm and planner again. This time
concentrating on the decoupling of the stepper ISR completely. It is
now dumb, relying on the segment generator to provide the number of
steps to execute and how fast it needs to go. This freed up lots of
memory as well because it made a lot tracked variables obsolete.
- The segment generator now computes the velocity profile of the
executing planner block on the fly in floating point math, instead of
allowing the stepper algorithm to govern accelerations in the previous
code. What this accomplishes is the ability and framework to (somewhat)
easily install a different physics model for generating a velocity
profile, i.e. s-curves.
- Made some more planner enhancements and increased efficiency a bit.
- The changes also did not increase the compiled size of Grbl, but
decreased it slightly as well.
- Cleaned up a lot of the commenting.
- Still much to do, but this push works and still is missing feedholds
(coming next.)
- Returned the max step rate to 30kHz. The new arc algorithm works uses
so much less CPU overhead, because the segments are longer, that the
planner has no problem computing through them.
- Fixed an issue with the acceleration independence scaling. Should now
work with accelerations above 400mm/sec^2 or so.
- Updated README
- Arc mm_per_segment parameter was removed and replaced with an
arc_tolerance parameter, which scales all arc segments automatically to
radius, such that the line segment error doesn't exceed the tolerance.
Significantly improves arc performance through larger radius arc,
because the segments are much longer and the planner buffer has more to
work with.
- Moved n_arc correction from the settings to config.h. Mathematically
this doesn't need to be a setting anymore, as the default config value
will work for all known CNC applications. The error does not accumulate
as much anymore, since the small angle approximation used by the arc
generation has been updated to a third-order approximation and how the
line segment length scale with radius and tolerance now. Left in
config.h for extraneous circumstances.
- Corrected the st.ramp_count variable (acceleration tick counter) to a
8-bit vs. 32-bit variable. Should make the stepper algorithm just a
touch faster overall.
- Maximum velocity for each axis is now configurable in settings. All
rapids/seek move at these maximums. All feed rates(including rapids)
may be limited and scaled down so that no axis does not exceed their
limits.
- Moved around auto-cycle start. May change later, but mainly to ensure
the planner buffer is completely full before cycle starting a streaming
program. Otherwise it should auto-start when there is a break in the
serial stream.
- Reverted old block->max_entry_speed_sqr calculations. Feedrate
overrides not close to ready at all.
- Fixed intermittent slow trailing steps for some triangle velocity
profile moves. The acceleration tick counter updating was corrected to
be exact for that particular transition. Should be ok for normal
trapezoidal profiles.
- Fixed the Timer0 disable after a step pulse falling edge. Thanks
@blinkenlight!
NOTE: This push is a work-in-progress and there are known bugs that
need to be fixed, like homing acceleration being incompatible. Released
for testing. Settings will definitely be overwritten, as new settings
were needed.
- Acceleration independence installed in planner. Each axis can now
have different accelerations and Grbl will maximize the accelerations
depending on the direction its moving. Very useful for users like on
the ShapeOko with vastly different Z-axis properties.
- More planner optimizations and re-factoring. Slightly improved some
of the older calculations, but new acceleration calculations offset
these improvements. Overall no change in processing speed.
- Removed planner nominal length checks. It was arguable whether or not
this improved planner efficiency, especially in the worst case scenario
of arcs.
- Updated readme and changed to markdown format.
- Improved planner execution speed by 5% or more. Re-factored most of
the calculations in terms of the square of velocity. This removed a lot
of sqrt() calculations for every planner_recalculate.
- Oops again. Thought the new planner changes made things much better,
but there was a bug. Improvements we on the order of 20% execution time
reduction, rather than half. The increase to 30kHz Ranade timer
frequency also increased the overall overhead, so the total planner
change? Zero. But, it's still better.
- Planner execute speed has been more than halved from 4ms to 1.9ms
when computing a plan for a single line segment during arc generation.
This means that Grbl can now run through an arc (or complex curve)
twice as fast as before without starving the buffer. For 0.1mm arc
segments, this means about the theoretical feed rate limit is about
3000mm/min for arcs now.
- Increased the Ranade timer frequency to 30kHz, as there doesn't seem
to be any problems with increasing the frequency. This means that the
maximum step frequency is now back at 30kHz.
- Added Zen Toolworks 7x7 defaults.
- Brand-new stepper algorithm. Based on the Pramod Ranade inverse time
algorithm, but modified to ensure step events are exact. Currently
limited to about 15kHz step rates, much more to be done to enable 30kHz
again.
- Removed Timer1. Stepper algorithm now uses Timer0 and Timer2.
- Much improved step generation during accelerations. Smoother. Allows
much higher accelerations (and speeds) than before on the same machine.
- Cleaner algorithm that is more easily portable to other CPU types.
- Streamlined planner calculations. Removed accelerate_until and
final_rate variables from block buffer since the new stepper algorithm
is that much more accurate.
- Improved planner efficiency by about 15-20% during worst case
scenarios (arcs).
- New config.h options to tune new stepper algorithm.
- Updated interface protocol to play nicer with interface programs. All
Grbl responses beginning with '$' signifies a setting. Bracketed '[]'
responses are feedback messages containing either state, parameter, or
general messages. Chevron '<>' response are from the real-time status
messages, i.e. position.
- M2 Program end command was causing a system alarm. Fixed. Thanks
@blinkenlight !
- Added a purge buffer (and lock) command. This is an advanced option
to clear any queued blocks in the buffer in the event of system
position being lost or homed. These queued blocks will likely not move
correctly if not purged. In typical use, the purging command releases
the homing axes lock in case a user need to move the axes off their
hard limit switches, but position is not guaranteed. Homing is advised
immediately after.
- Created a system-wide sync current position function. Cleans up some
of the repetitive tasks in various places in the code that do the same
thing.
- Removed the clear all switches command '$S'. Not really needed and
helped clean up a sync call.
- Other minor tweaks. Readme updated slightly..
(All v0.8 features installed. Still likely buggy, but now thourough
testing will need to start to squash them all. As soon as we're done,
this will be pushed to master and v0.9 development will be started.
Please report ANY issues to us so we can get this rolled out ASAP.)
- User startup script! A user can now save one (up to 5 as compile-time
option) block of g-code in EEPROM memory. This will be run everytime
Grbl resets. Mainly to be used as a way to set your preferences, like
G21, G54, etc.
- New dry run and check g-code switches. Dry run moves ALL motions at
rapids rate ignoring spindle, coolant, and dwell commands. For rapid
physical proofing of your code. The check g-code switch ignores all
motion and provides the user a way to check if there are any errors in
their program that Grbl may not like.
- Program restart! (sort of). Program restart is typically an advanced
feature that allows users to restart a program mid-stream. The check
g-code switch can perform this feature by enabling the switch at the
start of the program, and disabling it at the desired point with some
minimal changes.
- New system state variable. This state variable tracks all of the
different state processes that Grbl performs, i.e. cycle start, feed
hold, homing, etc. This is mainly for making managing of these task
easier and more clear.
- Position lost state variable. Only when homing is enabled, Grbl will
refuse to move until homing is completed and position is known. This is
mainly for safety. Otherwise, it will let users fend for themselves.
- Moved the default settings defines into config.h. The plan is to
eventually create a set of config.h's for particular as-built machines
to help users from doing it themselves.
- Moved around misc defines into .h files. And lots of other little
things.
- Allowed status_message function to be called by others. This is to
centralize all feedback into protocol.c.
- Fixed a bug where line number words 'N' were causing the parser to
error out.
- Allowed homing routine feed rates to move slower than the
MINIMUM_STEP_RATE parameter in config.h.
- Homing performs idle lock at the end of the routine.
- Stepper idle lock time will now not disable the steppers when the
value is set at 255. This is accomodate users who prefer to keep their
axes enabled at all times.
- Moved some defines around to where they need to be.
- Limit pin internal pull-resistors now enabled. Normal high operation.
This will be the standard going forward.
- Updated all of the 'double' variable types to 'float' to reflect what
happens when compiled for the Arduino. Also done for compatibility
reasons to @jgeisler0303 's Grbl simulator code.
- G-code parser will now ignore 'E' exponent values, since they are
reserved g-code characters for some machines. Thanks @csdexter!
- The read_double() function was re-written and optimized for use in
Grbl. The strtod() avr lib was removed.
- Added acceleration to the homing routine.
- Homing now accounts for different step rates when moving multiple
axes without exceeding acceleration limits.
- Homing now updates all internal positioning variables to machine zero
after completion.
- "Poor-man's" debounce delay added.
- Updated the delay_us() function to perform faster and more accurate
microsecond delays. Previously, the single increments would add
noticeable time drift for larger delays.
- Fix a bug in the stepper.c prescalar calculations that was changed in
the last commit.
- Other minor fixes.
- G54 work coordinate system support. Up to 6 work coordinate systems
(G54-G59) available as a compile-time option.
- G10 command added to set work coordinate offsets from machine
position.
- G92/G92.1 position offsets and cancellation support. Properly follows
NIST standard rules with other systems.
- G53 absolute override now works correctly with new coordinate systems.
- Revamped g-code parser with robust error checking. Providing user
feedback with bad commands. Follows NIST standards.
- Planner module slightly changed to only expected position movements
in terms of machine coordinates only. This was to simplify coordinate
system handling, which is done solely by the g-code parser.
- Upon grbl system abort, machine position and work positions are
retained, while G92 offsets are reset per NIST standards.
- Compiler compatibility update for _delay_us().
- Updated README.
- Program stop support (M0,M1*,M2,M30*). *Optional stop to be done.
*Pallet shuttle not supported.
- Work position is set equal to machine position upon reset, as
according to NIST RS274-NGC guidelines. G92 is disabled.
- Renamed mc_set_current_position() to mc_set_coordinate_offset().
- Fixed bug in plan_synchronize(). Would exit right before last step is
finished and caused issues with program stops. Now fixed.
- Spindle now stops upon a run-time abort command.
- Updated readme and misc upkeeping.
- Fixed a premature step end bug dating back to Simen's 0.7b edge
version is fixed, from which this code is forked from. Caused by Timer2
constantly overflowing calling the Step Reset Interrupt every 128usec.
Now Timer2 is always disabled after a step end and should free up some
cycles for the main program. Could be more than one way to fix this
problem. I'm open to suggestions.
- _delay_ms() refactored to accept only constants to comply with
current compilers. square() removed since not available with some
compilers.
- Grbl now tracks both home and work (G92) coordinate systems and does
live updates when G92 is called.
- Rudimentary home and work position status reporting. Works but still
under major construction.
- Updated the main streaming script. Has a disabled periodic timer for
querying status reports, disabled only because the Python timer doesn't
consistently restart after the script exits. Add here only for user
testing.
- Fixed a bug to prevent an endless serial_write loop during status
reports.
- Refactored the planner variables to make it more clear what they are
and make it easier for clear them.
- Added machine position reporting to status queries. This will be
further developed with part positioning/offsets and maintaining
location upon reset.
- System variables refactored into a global struct for better
readability.
- Removed old obsolete Ruby streaming scripts. These were no longer
compatible. Updated Python streaming scripts.
- Fixed printFloat() and other printing functions.
- Decreased planner buffer back to 18 blocks and increased TX serial
buffer to 64 bytes. Need the memory space for future developments.
- Begun adding run-time modes to grbl, where block delete toggle, mm/in
reporting modes, jog modes, etc can be set during runtime. Will be
fleshed out and placed into EEPROM when everything is added.
- ALPHA status. - Multitasking ability with run-time command executions
for real-time control and feedback. - Decelerating feed hold and resume
during operation. - System abort/reset, which immediately kills all
movement and re-initializes grbl. - Re-structured grbl to easily allow
for new features: Status reporting, jogging, backlash compensation. (To
be completed in the following releases.) - Resized TX/RX serial buffers
(32/128 bytes) - Increased planner buffer size to 20 blocks. - Updated
documentation.
- Fleshed out the original idea to completely remove the long slope at
the end of deceleration issue. This third time should absolutely
eliminate it.
- Changed the acceleration setting to kept as mm/min^2 internally,
since this was creating unneccessary additional computation in the
planner. Human readable value kept at mm/sec^2.
- Updated grbl version 0.7d and settings version to 4. NOTE: Please
check settings after update. These may have changed, but shouldn't.
- Before updating the new features (pause, e-stop, federate override,
etc), the edge branch will soon be merged with the master, barring any
immediate issues that people may have, and the edge branch will be the
testing ground for the new grbl version 0.8.
- The long standing issue of a long slope at deceleration is likely
fixed. The stepper program was not tracking and timing the end of
acceleration and start of deceleration exactly and now is fixed to
start and stop on time. Also, to ensure a better acceleration curve fit
used by the planner, the stepper program delays the start of the
accelerations by a half trapezoid tick to employ the midpoint rule. -
Settings version 3 migration (not fully tested, but should work) -
Added a MINIMUM_PLANNER_SPEED user-defined parameter to planner to let
a user change this if problems arise for some reason. - Moved all
user-definable #define parameters into config.h with clear comments on
what they do and recommendations of how to change them. - Minor
housekeeping.
- Fixed the planner TODO regarding minimum nominal speeds. Re-arranged
calculations to be both more efficient and guaranteed to be greater
than zero. - Missed a parenthesis location on the rate_delta
calculation. Should fix a nearly in-perceptible issue with incorrect
acceleration ramping in diagonal directions. - Increased maximum dwell
time from 6.5sec to an 18hour max. A crazy amount more, but that's how
the math works out. - Converted the internal feedrate values to mm/min
only, as it was switching between mm/min to mm/sec and back to mm/min.
Also added a feedrate > 0 check in gcode.c. - Identified the long slope
at the end of rapid de/ac-celerations noted by stephanix. Problem with
the numerical integration truncation error between the exact solution
of estimate_acceleration_distance and how grbl actually performs the
acceleration ramps discretely. Increasing the
ACCELERATION_TICKS_PER_SECOND in config.h helps fix this problem.
Investigating further.
- 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.
- 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.
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.
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.
- 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.
