The temperature and fsensor ISR re-enable interrupts while executing.
However, we still need to protect the epilogue of the ISR so that
the saved return address is not altered while returning.
We hoist the body of the function out of the isr in both cases for
clarity (and to avoid a stray return bypassing the lock/cli), so that
the re-entrant portion is clearly indicated.
This should fix the "STATIC MEMORY OVERWRITTEN" error messages randomly
happening when stepping at high frequency (where either isr is
preempted more frequently).
Factor-out MIN/MAXTEMP [BED/AMB] out of the error message, which is now
built at runtime instead.
Introduce two missing ultralcd functions lcd_setalertstatus and
lcd_updatestatus to handle regular strings.
246272 -> 246084 = 188 bytes saved
Partially revert 285b505c73a54e9af01816e3a614de73ad181851 so that
we ensure heaters are disabled ASAP in case of potential bugs
in the max_*_error functions.
Take advantage of the NTC thermistor found on the Einsy as an additional
safety measure, following the steps of the other MIN/MAXTEMP errors.
Introduce two configurable params AMBIENT_MINTEMP and AMBIENT_MAXTEMP
in the variant defines and set them for the MK3/MK3S to -30/+100
respectively.
AMBIENT_MINTEMP is primarily intended to catch a defective board
thermistor (to ensure MAXTEMP would be properly triggered) and thus the
trigger temperature is set just above the sensing limit and well below
the operating range.
AMBIENT_MAXTEMP is set at 100C, which is instead 20C above the maximum
recommended operating temperature of the Einsy. The NTC thermistor is
located just above the main power connector on the bottom of the board,
and could also help in detecting a faulty connection which can result in
rapid overheating of the contacts.
As for MAXTEMP, we cut power to the heaters, print fan and motors to
reduce power draw. Resume is not possible except by resetting the
printer, since the user is highly advised to inspect the board for
problems before attempting to continue.
In max/min_temp handlers remove the redundant disable_heater() call.
Handlers already need to call Stop(), which will disable all heaters
as the first step.
Fix comments in order to mention that all heaters get disabled.
Use "MAX/MINTEMP BED" correctly in both the LCD and serial.
Remove most of the original complexity from advance_spread.
Instead of accumulating time to be scheduled, plan ahead of time each
eISR tick using the next main interval + an accumulator (eISR_Err),
which keeps everything much simpler.
The distribution of the advance ticks is now using the real LA
frequency, which leaves a bit more time between the last LA tick and
the main stepper isr.
We take advantage of the accumulator to force a LA tick right after the
first main tick, which removes a +/- 1 scheduling error at higher step
rates.
When decompressing, we force 2 steps instead, so that the direction
reversal happens immediately (first tick zeros esteps, second inverts
the sign), removing another +/- 1 error at higher step rates.
If you're using flow to correct for an incorrect source diameter, which
is probably the main usage when using the LCD, then LA shouldn't be
adjusted.
It's still unclear what the effect of M221 in gcode should be regarding
overall extrusion width. If M221 means "thicker lines", then LA should
also be adjusted accordingly.
This stems from the fact that the source diameter/length needs to be
known in order to determine a compression factor which is independent of
the extrusion width, but the FW only ever sees one value currently (the
extrusion length) which combines both.
This makes it impossible for the FW to adjust for one OR the other
scenario, depending on what you expect for M221 to mean.
The e/D ratio should be calculated using the extrusion length.
As such, purify the e_D_ratio from the current extruder multiplier in
order to account correctly for flow adjustments.
Turns out for high-res curved models the numerical error and the
SLOWDOWN handling in the planner can cause enough variance in the
calculated pressure to trigger LA to continuosly, making matters worse.
Clamp LA again, but only during extrusion, so that the runaway error is
limited by the current segment length.
Before PR #2591 LA was automatically capped during cruising or
deceleration. However we now rely on reaching the current pressure state
exactly to stop. When dual/quad stepping inside the eISR we might incur
in oscillating behavior if we do not handle it correctly.
This might be the cause behind #2757
This now changes e_step_loops to be a phase-local variable, so we now
reset it each phase too (instead of per-segment).
When switching to a new trapezoid step with the right pressure, cancel
any pending eISR right away.
Similarly do not schedule another eISR if the pressure will be reached
by the end of the eISR.
This was done in the past to preserve the current LA_phase. This is not
needed anymore, since it will be reset at each trapezoid step when LA
is re-initialized.
There used to be a single stage where an extruder reversal could occur,
but since PR #2591 reversals can happen up to two times per trapezoid.
Reset LA_phase when ADV_INIT is set, since it is re-inizialized only
when needed a few lines afterward. This improves performance by avoiding
to check the phase continuosly to the end of the trapezoid.
Likewise, always set ADV_INIT during the first cruising step, also to
force a LA_phase reset.
PR #2591 made LA compression always account for retractions instead of
discarding the current compression steps. While this fixed overextrusion
in short segments followed by wipes, it uncovered another issue in how
the compression steps are spread during the trapezoid calculations
leading to gaps in segments followed by retractions (as highlighted by
/some/ prints in #2693).
LA1.5 always computes the required target compression steps for a
segment at nominal speed. Because of how the extra steps are allocated
using multiples of the accelerating frequency, if the segment is
truncated before cruising is reached, an additional cycle of steps can
be inserted before deceleration starts. Deceleration is also not
guaranteed to be symmetric where up to _two_ cycles can be skipped
depending on the stepping cycle, leading to a situation where a
symmetric acceleration/deceleration block will lead up to a cycle of
accumulated compression.
While forcing an the extra step during deceleration is possible by
tweaking the error term (eISR_Err), this doesn't guarantee balance in
all cases. The underlying issue is that the function is aiming a
compression which cannot be reached (nominal speed), and not at the
effective max speed reached in the trapezoid, thus moving the average
result higher over time.
We fix this by calculating the effective maximum speed (and compression)
reached during the trapezoid, which stops compression on the required
cycle irregardless of the error term, balancing the result.
This is the first unoptimized POC: this is not for production: a lot of
calculations are redundand and could work directly in steps/s^2.
This saves nearly all of additional FLASH usage of previous commit. It is only 70B worse than if no skipping is done in temperature compensation gcode G76.
