Re-factored system states and alarm management. Serial baud support greater than 57600.

- Refactored system states to be more clear and concise. Alarm locks
processes when position is unknown to indicate to user something has
gone wrong.

- Changed mc_alarm to mc_reset, which now manages the system reset
function. Centralizes it.

- Renamed '$X' kill homing lock to kill alarm lock.

- Created an alarm error reporting method to clear up what is an alarm:
message vs a status error: message. For GUIs mainly. Alarm codes are
negative. Status codes are positive.

- Serial baud support upto 115200. Previous baudrate calc was unstable
for 57600 and above.

- Alarm state locks out all g-code blocks, including startup scripts,
but allows user to access settings and internal commands. For example,
to disable hard limits, if they are problematic.

- Hard limits do not respond in an alarm state.

- Fixed a problem with the hard limit interrupt during the homing
cycle. The interrupt register is still active during the homing cycle
and still signal the interrupt to trigger when re-enabled. Instead,
just disabled the register.

- Homing rate adjusted. All axes move at homing seek rate, regardless
of how many axes move at the same time. This is unlike how the stepper
module does it as a point to point rate.

- New config.h settings to disable the homing rate adjustment and the
force homing upon powerup.

- Reduced the number of startup lines back down to 2 from 3. This
discourages users from placing motion block in there, which can be very
dangerous.

- Startup blocks now run only after an alarm-free reset or after a
homing cycle. Does not run when $X kill is called. For satefy reasons
This commit is contained in:
Sonny Jeon
2012-11-14 17:36:29 -07:00
parent e6ad15b548
commit 559feb97e2
14 changed files with 466 additions and 420 deletions

View File

@ -41,30 +41,41 @@ void limits_init()
LIMIT_PORT |= (LIMIT_MASK); // Enable internal pull-up resistors. Normal high operation.
if (bit_istrue(settings.flags,BITFLAG_HARD_LIMIT_ENABLE)) {
LIMIT_PCMSK |= LIMIT_MASK; // Enable specific pins of the Pin Change Interrupt
PCICR |= (1 << LIMIT_INT); // Enable Pin Change Interrupt
LIMIT_PCMSK |= LIMIT_MASK; // Enable specific pins of the Pin Change Interrupt
PCICR |= (1 << LIMIT_INT); // Enable Pin Change Interrupt
} else {
LIMIT_PCMSK &= ~LIMIT_MASK; // Disable
PCICR &= ~(1 << LIMIT_INT);
}
}
// This is the Limit Pin Change Interrupt, which handles the hard limit feature.
// This is the Limit Pin Change Interrupt, which handles the hard limit feature. A bouncing
// limit switch can cause a lot of problems, like false readings and multiple interrupt calls.
// If a switch is triggered at all, something bad has happened and treat it as such, regardless
// if a limit switch is being disengaged. It's impossible to reliably tell the state of a
// bouncing pin without a debouncing method.
// NOTE: Do not attach an e-stop to the limit pins, because this interrupt is disabled during
// homing cycles and will not respond correctly. Upon user request or need, there may be a
// special pinout for an e-stop, but it is generally recommended to just directly connect
// your e-stop switch to the Arduino reset pin, since it is the most correct way to do this.
// TODO: This interrupt may be used to manage the homing cycle directly with the main stepper
// interrupt without adding too much to it. All it would need is some way to stop one axis
// when its limit is triggered and continue the others. This may reduce some of the code, but
// would make Grbl a little harder to read and understand down road. Holding off on this until
// we move on to new hardware or flash space becomes an issue. If it ain't broke, don't fix it.
ISR(LIMIT_INT_vect)
{
// Only enter if the system alarm is not active.
if (bit_isfalse(sys.execute,EXEC_ALARM)) {
// Kill all processes upon hard limit event.
if ((LIMIT_PIN & LIMIT_MASK) ^ LIMIT_MASK) {
mc_alarm(); // Initiate system kill.
sys.state = STATE_LIMIT; // Set system state to indicate event.
}
// TODO: This interrupt may be used to manage the homing cycle directly with the main stepper
// interrupt without adding too much to it. All it would need is some way to stop one axis
// when its limit is triggered and continue the others. This may reduce some of the code, but
// would make Grbl a little harder to read and understand down road. Holding off on this until
// we move on to new hardware or flash space becomes an issue. If it ain't broke, don't fix it.
// Ignore limit switches if already in an alarm state or in-process of executing an alarm.
// When in the alarm state, Grbl should have been reset or will force a reset, so any pending
// moves in the planner and serial buffers are all cleared and newly sent blocks will be
// locked out until a homing cycle or a kill lock command. Allows the user to disable the hard
// limit setting if their limits are constantly triggering after a reset and move their axes.
if (sys.state != STATE_ALARM) {
if (bit_isfalse(sys.execute,EXEC_ALARM)) {
mc_reset(); // Initiate system kill.
sys.execute |= EXEC_CRIT_EVENT; // Indicate hard limit critical event
}
}
}
@ -104,6 +115,11 @@ static void homing_cycle(bool x_axis, bool y_axis, bool z_axis, int8_t pos_dir,
// Compute the adjusted step rate change with each acceleration tick. (in step/min/acceleration_tick)
uint32_t delta_rate = ceil( ds*settings.acceleration/(60*ACCELERATION_TICKS_PER_SECOND));
#ifdef HOMING_RATE_ADJUST
// Adjust homing rate so a multiple axes moves all at the homing rate independently.
homing_rate *= sqrt(x_axis+y_axis+z_axis);
#endif
// Nominal and initial time increment per step. Nominal should always be greater then 3
// usec, since they are based on the same parameters as the main stepper routine. Initial
// is based on the MINIMUM_STEPS_PER_MINUTE config. Since homing feed can be very slow,
@ -191,12 +207,13 @@ static void homing_cycle(bool x_axis, bool y_axis, bool z_axis, int8_t pos_dir,
void limits_go_home()
{
// Enable only the steppers, not the cycle. Cycle should be complete.
// Enable only the steppers, not the cycle. Cycle should be inactive/complete.
st_wake_up();
// Jog all axes toward home to engage their limit switches at faster homing seek rate.
homing_cycle(false, false, true, true, false, settings.homing_seek_rate); // First jog the z axis
homing_cycle(true, true, false, true, false, settings.homing_seek_rate); // Then jog the x and y axis
// First jog z-axis to clear workspace, then jog the x and y axis.
homing_cycle(false, false, true, true, false, settings.homing_seek_rate); // z-axis
homing_cycle(true, true, false, true, false, settings.homing_seek_rate); // xy-axes
delay_ms(settings.homing_debounce_delay); // Delay to debounce signal
// Now in proximity of all limits. Carefully leave and approach switches in multiple cycles