- guide-stepper:
- add queue to send commands to stepper-ctl task
- add function that tells stepper task to move to zero
- increase travel length full axis length 110
- control: move guide to zero at reset
- Cmake: enable colored output
- prevent hard stops (no decel ramp) when target changes while
decelerating already (overshoot and move back)
- optimize/adjust stepper-test mode, add macro for test with single switch
Note: tested the current state, works quite well. very rarely crashes
probably dude to race conditions in isr?
- Add waitForStop function to custom stepper driver
- use that function in guide-stepper to ensure stepper actually was at
max/min position before dir change
- move stepper config e.g. speed accel steps... from stepper.cpp to
config.h
- uncomment and rework code in guide-stepper.cpp (for traveling the
stepper based on encoder) to work with new custom driver instead of
previous library
Note currently in STEPPER_SIMULATE_ECNODER mode (use button as encoder),
see macro
- add functions
setSpeed (mm/s)
setTargetPosMm
- add macros and function to define and convert variables to actual units
speed in mm/s instead of unknown
abs pos in mm instead of steps
- simplyfy isr code
new driver works well while testing with one button with debug output
- small changes to isr function that made it fully functional!
- add stepper debug task
- minor optimizations necessary
Partially implemented stepper driver as planned in diagram
works partially (reset button toggles distances)
major issues though - wip
needs debugging and testing
basic task for controlling the stepper
(no accel / decel etc)
works at current speed settings but long moves trigger watchdog
so dropped this idea and using timer as before
Some adjustments to make it work
- fix direction change (temporary)
- fix isr initial start
See notes in stepper.cpp
Todo:
- fix crash after accelerating
- add initial speed (takes long to accelerate from 0
-
- Add stepper driver code from scratch
- Remove / comment out code that used DendoStepper library
- Test custom driver with buttons (guide-stepper.cpp)
untested limited features
- decrease home speed
- invert direction in stepper test and home function due to changed
wiring
- stretch travel steps calculation to multiple steps and variables for
better overview and debugging
- prevent partial steps from being lost -> fixes rounding issue
TESTED: works in general, but steper movement is too jerky. Also there
is a Hardware issue: When VFD/motor is on the steper does not work
properly -> 50ohm pullup near stepper driver fixed this
- update config and connection plan
- swap B+ and B- of encoder to invert the rotational direction,
so the direction in code make more sense
- swap VFD and stepper pin. For some reason stepper did not work with
dir pin on mos2. Maybe pulse affected dir pin, try again with denoise
pcb
- optimize some variable and macro names
- change/add function that travels guide certain steps
- stepper task: - calculate steps depending on encoder steps
- move steps if at least 1 step possible
- prepare poti modifier
-> untested
- Add datasheets of VFD and stepper driver
- Change pins in config for stepper driver
- DISABLED LAMP (pin now used for stepper)
- update connection plan with stepper section and wiring
- add static type to all local variables in every source file
so the scope of each variable is limited to that file.
This prevents conflicting variable names with other source file
(e.g. 'state' variable in vfd.cpp)
- simplify encoder concept
wrap all used functions from rotary_encoder.h library in
custom encoder.hpp file
-> only one file has to be included where encoder is used
-> global variable 'encoder' is not necessary anymore
-> reduces duplicate code for length calculation
- update all files where encoder functions where used accordingly
To be able to use the encoder in future stepper task as well
- make certain variables global
- initialize encoder in main
- outsorce encoder functions in encoder.cpp / h
- adjust config
While testing the setup
Optimizations of problems noticed while testing the setup:
- mount too weak
- only very slow speeds relevant
- less distance required
including the files in source file instead of header file fixed
'conflicting type' errors in idf macros when using newer version
ESP-IDF v4.4.4-148-g4c2afac355
- add macro variable STEPPER_TEST (config.hpp)
if set: only `task_stepper-test` is started (main.cpp)
- add file for stepper/guide related code (guide-stepper.cpp/hpp)
currently defines a task that simply moves the axis left to right
repeatedly with a speed defined by a potentiometer connected to ADC
GPIO34
Add feature and config option #define TARGET_LENGTH_OFFSET
that makes it possible to set a distance that gets added to the target
length, thus preventing falling short on the actual length when spools
rotates back slightly after stop.
Also move reached tolerance to config.
- Pending auto cut (countdown) can be canceled using any button on the
panel
- Reset button does not reset the length when countdown cancledo
- Fix bug where display did not initialize correctly after start some
times
-> simply init again after welcome msg
- reduce countdown time by 500ms
- Basic lamp functionality: simply turn on when not in COUNTING or TARGET_REACHED
state
- Remove unnecessary comments, fix code formatting in several files
Fix bug where display is stuck when countdown stopped via reset button
Add functionality to stop countdown via turning auto-cut switch off
Fix display priority issues with using 2 boolean variables and a delay
before counting
While testing the machine with current firmware, a few parameters got
slightly adjusted:
- dynamic spped lvl later slow
- longer slow speed at start
- less threshold for stying in target reached
- decreased countdown beep interval at auto cut
- adjusted display interval in manual mode (faster)
control.cpp:
- Add new systemState CUTTING
- change systemState_t from enum to enum class (conflicting with
cutter enum)
- Add CUTTING functionality to CUT switch and auto cut after target
reached if enabled
cutter.cpp:
- Add function cutter_isRunning()
config:
- Add GPIO_LAMP macro to config
=> tested system with new stripboard and actual hardware: system
works as expected with new all combinations of the 4 switches
connected to 1 adc
- Adjust lookup voltage for switch combinations, after connecting the
stripboard to actualy used pcb (apparently there was a >100
difference using an other adc channel + pcb)
- define loglevel for analogswitches to WARN
Rework config.cpp and config.hpp:
- different naming convention for evaluated switch objects
- thus updated macros for switches
- add new evalswitch instances for 4 switches on 1 adc
- update switch assignment, use the 4 new inputs
Update connection plan with new assignment, add stripboard details
switchesAnalog:
- Add functions for obtaining states from each switch
- Remove public handle function, (now run locally at each state request of
a pin)
- adjust main.cpp to work with new functions
- create new files with code for demultiplexing the 4 switches from the
one analog signal
- define lookup table with measurements done with the build pcb, test
switches and esp32 breakout board
- add note do config that gpio used for digital in is now used for
this multiplexer
- main.cpp: disable control task, run new handle function in a loop for
testing
- use evaluatedSwitch class for position switch instead of reading gpio
directly because it triggered incorrectly too often
- cutter_stop() only switch to CANCELED when not in IDLE already
Add encoder test:
- used for calibrating the length measurement by counting the steps
per meter
- enabled with #define ENCODER_TEST in config.hpp
- display1: enocder steps
- display2: converted length in meter
Distance conversion:
- defining STEPS_PER_METER instead of MEASURING_ROLL_DIAMETER in
config.hpp
- defining the steps per meter by issuing 3 test measurements with
4mm2 solar cable in the new encoder test mode
