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
before adding the travelMm function the algorithm for moving the
axis between left and right was designed and evaluated via commandline
first.
-> Add this test-code in this commit
When the VFD was running the stepper driver did not work correctly.
The especially the DIR signal got unusable and triggered randomly.
Add custom stripboard that reduces the noise on DIR and PULSE data
lines.
The circuit simply increases the current in the data lines significantly
and crops the voltage to 0-5V using a zener diode.
While testing with an oscilloscope the circuit seems effective.
- add kicad schematic
- add spreadsheet with stripboard layout
- update connection: plan with stripboard, update changed 5V connection
- 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
This file currently contains
- crude model of linear-axis
- angled bracket for mounting the axis to item
- 2 versions of interface between rolls and axis
- rolls
- 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.
Add add cad files of all designed and 3d-printed parts.
Mostly Freecad and a few older SolidEdge files (_SE)
Add wire-labels.fods containing the printed labels for wires in
control-box and switching cabinet
Add docs folder with relevant images and datasheets of the used components
- 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
Update connection plan to match the actual current wiring
Add page to connection plan which describes which cables are connected
between what components
- 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
evaluatedSwitch component:
- remove duplicate code for inverted and non-inverted mode
- add new constructor for instance with a function to obtain current
input state instead of gpio pin (e.g. needed for 4switches on analog input)
- 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
Add details and calculations for a stripboard with a resistor network that makes it
possible to connect 4 switches (to gnd) to one adc pin of the controller
- 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
