New super generic firmware

Hi guys, after reading lots of code (repetier, teacup, not-grbl) and try simulate motion controller on freebasic, i develop a new firmware based on my motion controller.

I code it as generic as possible so it can be compiled at many cpu. so far i have tested work on Nano V3 and using my own 3d board NanoNano.
It use a motion loop with micros() to get timing (just like not-grbl)

But it compile success on megacore, esp8266, just not have been tested.

also if you compile using PC Quincy C++ you can see the motion controller simulation graph. Or you can see the old simulation code in freebasic.

[github.com]

Currently implement this:

-Per motor step using float speed ramping acceleration (smooth ) *
-per axis acceleration *
-per axis max feedrate *
-per axis jerk *
-path planner (forward) *
-step per mm for each axis *
-implement motor stepper hardware layer *
-implement GCODE parser and processing *
-implement flow control between motion gcode and non motion gcode *
-4 axis X Y Z E *
-implement endstop reading (xmin xmax , etc)*
-eeprom configuration (step/mm, accel, travel accel, jerk, max axis) *
-heater with PID E0 *
-config files to set pins and parameter *
-G0 and G1 can have different acceleration (travel vs feed/extrude) *

Working on:
backlash for all motor

Maybe not:
backward planner
interrupt timing

[www.youtube.com]

Thank you for sharing your project. It takes a great deal of work to create something functional and then share it openly.

I can appreciate that.

If you can, could you explain what the idea behind implementing "flow control between motion gcode and non motion gcode". I'm guessing that you are doing this because code is not using interrupts - is that right?

Edited 1 time(s). Last edit at 01/07/2018 11:06AM by newbob.

yes its not use interrupt.

also its generic and slow, i am still optimizing the motion loop, i measure it need more than 200us for each loop. so it cant make motor with high step to move fast.
my printer is 175step/mm and i think it only max at 20-30mm/s

I was printing and see there is problem with hole, its not round then i make simulation in PC


Big bug in i dont know where, perhaps the bresenham ? perhaps the gcode parsing ?

i compare preview from repetier and karyacontrollerin PC winBGI graphics.



time to hunt bugs.

FOUND the bugs in parsing value



Edited 1 time(s). Last edit at 01/09/2018 07:21AM by ryannining.

2hours printing xiro landing on karyacontroller

I have just uploaded these images:

Nice. How do you compile it for ESP8266?

Well i am familiar with esp, since i have project some times ago making camera slider and also i have use not-grbl in the past (yes this was super generic cnc firmware too, with no path planner).

I can compile and upload to nodemcu/wemos d1 and then test the gcode , first enable output_enable in config_pins.h to enable debugging text, just like this. Its show the speed ramp up and down, when i send command G0 X5, and also return temp from analog pin (still wrong), still focus on atmega, since i already have the board. For esp, i dont have pcb board, so need to make pcb first.

Also by enable #timing in INO file, i check that esp8266 can process the core motionloop in less than 5microseconds = 200k steps/sec, thats amazing because nano v3 is need 100us after i optimize the code.
Thats why i add new code, if the gcode processed was G0 (travel) then the delay is not calculated on each step, but after some microseconds (adjusted in motion.h, UPDATE_F_EVERY) so it can move the motor fast. Division in nano v3 is slow, although we just need 1 division on update the delay for each step.

Edited 1 time(s). Last edit at 01/10/2018 12:38PM by ryannining.

Thanks. I'll try to install it on ESP8266.

EDIT: it compiles just fine on Witty ESP8266

Edited 1 time(s). Last edit at 01/10/2018 11:00PM by newbob.

Connect the output pin to some A4988 driver / other motor driver

it should work. The endstop must be normally closed.

for esp i think should use shift register to get more output pins

Next, I'll try compiling it for ESP32.

Update

config_pins.h have this setting now

-COREXY and XZ
-backlash for all motor
-invert motor direction

need to backup eeprom before upload new because the eeprom layout change. Backup normally using repetier eeprom setting->Export.

STM32F103 work in progress, make the board and will test.

Implement SDCARD, optimize ram usage, and able to use SDCARD using SDFAT library, and configured using 8 movebuffer.

On NANO V3 328p:

Sketch uses 28372 bytes (92%) of program storage space. Maximum is 30720 bytes.
Global variables use 1725 bytes (84%) of dynamic memory, leaving 323 bytes for local variables. Maximum is 2048 bytes.


Still got low mmeory warning from compile, but i have test insert sdcard with a file print.gcode, which automatically loaded and run.

Print using micro sdcard and controlbox.

Its been long time

our firmware now tested well on STM32, the PCB board for Blue pill also included in the github.

For WemosD1 mini also tested, and pcb included.

We use it for our laser cutting, cnc and mini 3d printer.

now for AVR, STM, ESP8266 are all using TIMER interrupt.

For WEMOS D1, the firmware act also as a web server and websocket server, so can receive gcode from websocket.

Check [github.com] and [github.com]

Super cute smallest 32bit 3d printer board using wemos d1 mcu, on my corexy machine


Connect wirelessly using TCP/IP on repetier host,


All update can be checked here.
https://github.com/ryannining/karyacontroller

New feature : upload Bitmap data to machine and use the bitmap data as Pulse On and Off on laser engraving process.

Now its able to store GCODE into internal wemos flash, and start print without pc (by tap limit switch 3 times), the screenshot show simple html javascript app to compress and upload gcodes using wifi.