Changed webserver and network code to support multiple concurrent
connections
Removed support for differential-mode ultrasonic z-sensor
Reduced X and Y instantDv from 15 to 10 mm/sec to allow outlines of
holes to e drawn at lower speed
Changed temperature reading code to give thermistor disconnected reading
of -273C when there is a negative high ADC offset correction
When processing an M201 command, leave the acceleration of any axis that
is not specified as the existing value instead of setting it to a silly
value.
Corrected Platform.cpp to work with long filenames - note that this fix
requires a commented-out section of code in file ff.c in the SD_HSMCI
library to be re-instated.
Fixed bug introduced at version 057o-dc42 that broke the G32 bed
compensation. Added the Z heights of the bed compensation points to the
information displayed by M111 S2. Simplified the code that does
Z-probing in fast and slow stages. Added function sncatf and used it to
simplify the code in several places.
Changed the step interrupt service routine to use the average step size
when calculating the step interval, instead of the actual step size
which depend son whether X and Y are moving simultaneously. This should
give smoother movement, because from an inertial point of view the axes
are independent. It also fixes a bug whereby the accelerations were
incorrectly calculated, because DDA::AccelerationCalculation assumes a
uniform step size directly from start point to end point, whereas the
ISR was assuming a zigzag path. Also fixed bug with resetting the
minimum z-probe value seen when using an ultrasonic transducer in
differential mode.
1. Fixed bug whereby the first XY move after a Z move would often be at
a speed much higher than requested.
2. Added support for direct-mode ultrasonic sensor using command M558
P3. Command M558 P4 selects differential-mode ultrasonic sensor.
3. Changed temperature parameter in G31 command from T to S.
Added L and H parameters to the M301 and M304 commands to allow SAM3X
ADC offsets to be corrected for. Also attempted a fix for absolute
extruder movement. - but this will still be affected by rounding error.
Removed some old binaries.
1. Fixed bug introduced at version 57r whereby axis travel limits were
not applied properly.
2. Increased maximum travel on 2nd phase of Z-probing from 1mm to 10mm
3. Added parameters R (thermistor 25C resistance) and B (thermistor
beta) to M301 command
4. Added support for M304 command with same parameters as M301 command
5. PID and thermistor parameters are now saved to flash memory
Tidied up axis homing tracking.
When checking endstops, wait for move to complete before accepting
further moves, otherwise subsequent moves use the wrong coordinates.
Temperatures and Z probes are now monitored continuously using a tick
interrupt to kick off ADC conversions. ADC is now run in 12-bit mode.
Thermistor readings are passed through averaging filters. Thermistors
are monitored for overheat conditions and bad readings in the tick ISR
and the appropriate heater is turned off (useful because the main loop
sometimes gets suspended while trying to do USB communication).
Use watchdog timer to monitor the tick interrupt - needs patch to
Arduino Due core library. Add facility to test watchdog timer (M111
S1001). Added an error status word to record that errors have occurred
(e.g. over-temperature).
M111 command changes so that S0 turns debug off, S1 turns debug on, S2
reports free memory - also now reports the type of the last restart and
the error status word. Fixed problem whereby M111 debug reports were not
sent to the web interface.
Implemented M999 command, which resets the Duet.
Removed an unused variable.
Changed some more "char*" to "const char*".
Changed extruder PID parameters and added more explanation for them.
Fixed a random hang after z-homing in the previous release. Also make Z
probe temperature coefficient default to zero always, and Z probe
calibration temperature default to current bed temperature always.
1. Z-probing is now done in two stages: a fast stage at the configured
home feed rate until within 10% of the target value, then a slow stage
at 20% of that feed rate.
2. Provisional support for ultrasonic Z-probe.
3. Added calibration temperature and height temperature coefficient to Z
probe parameters.
4. Z-probe parameters and Z-probe type are now saved to flash memory so
that they survive power-off and reset cycles. Separate parameters are
retained for IR and ultrasonic probes in case both are fitted.
5. Fixed issue with doing slow Z-moves immediately after Z-homing or
probing.
Implemented M301 command to get/set PID parameters. Improved the PID
parameters and made minor changes to the PID algorithm to improve PID
performance with Ormerod hot end.
Increase TCP window size from 1 to 2 packets and increase maximum amount
of web data sent to 1460. This decreases the page load time fro a
Windows client fro 6 seconds to less than 1 second. It also fixes a
problem whereby when using a Ubuntu client, file upload and direct print
didn't work unless the reprap.js file was edited to reduce the maximum
amount of data sent per request. Also we now initialize unused memory
with known data so that in response to the M111 S1 command we can better
estimate the amount of free memory.
Reverted STRING_LENGTH to 1029 to avoid random print hangs that occurred
using 0.57f firmware. Reduced max reported buffer size to fit within
this. Report additional memory usage detail in response to M111 S0.
Increase incoming web gcode buffer size to 1200 chars to allow more
gcodes to be sent in each message
Include free buffer size in poll response
Include response sequence number in poll response
Implement M30 (delete file) via USB interface
Support M503 and M111 commands via web interface
Added new fields "zprobe" and "resp" to webserver poll response
Immutable string parameters are passed as const char* not char*
Increase max number of files displayed to 42
Range-check heater parameters in incoming commands to avoid buffer
overflow if heater number is out of range
In a previous commit I changed the temperature calculation to give more
accurate readings at extreme values. This change meant that a
disconnected thermistor no longer shows as absolute zero. Coupled with
the change in thermistor beta value, this meant that a disconnected
extruder thermistor would show as -28.2C, which is above the usual error
threshold of -30C. This commit introduces a special case for a
disconnected thermistor and restores the reading to absolute zero for
that case.
Add code to support sensors that allow the IR output to be modulated, to
reduce the sensitivity to ambient IR. Use M558 P2 command to enable
sensor modulation. Also changed the sensor reading averaging code to
give more consistent z-height seeking.
This change approximately doubles the speed of SD-card uploading via USB
to 2.2Mbyte/sec on the test system, provided that Pronterface has been
patched to not wait for an acknowledgement after sending each line of
g-code.
Fixed printing over USB and via HTTP so that it uses the same lookahead
mechanism as for printing from SD provided the data arrives fast enough.
Added an incoming buffer for data arriving from USB. NOTE: Pronterface
needs to be patched to send the data fast enough for this to work well.
Also changed signatures of some member functions to be const-correct
Improved network error recovery so that it it possible to reconnect
after some types of network error have occurred
Corrected calculation of temperatures to give more accurate results
towards the limits of the ADC range
file names into a structure created in RepRap that was too short for them. Result - buffer overflow misery. For the moment I've commented out the bit of code in ff.c that handles long names (look for the string ******), but this needs to be tracked down and dealt with properly.
1. Z home now moves home, rather than just setting the coordinate.
2. Wait for temps implemented.
Also the JSON requests for the g code filelist has been temporarily commented
out in reprap.htm. It was causing the firmware to hang sometimes. To be investigated.
Machine now prints nicely...
track up the low-precision slope of the IR probe before measuring down
the high-precision slope near 0 mm. But this means that the probe has to
be set only a fraction of a mm higher than the nozzle itself. Also (at
the moment for safety) the code thinks Z = 0 occurs when it is actually
0.2mm. This is easy to fix when the rest is reliable.
