Consecutive processing of output_pins in macro (SET_PIN) on rpi pico

Buzz · February 25, 2022, 11:05am

I’m currently wondering if the output_pins in an macro are / should be processed in a consecutive manner.

I got following constellation here:

RPI Pico (MCU) connected to an RPI3 which runs MainsailOS.

What me got to the problem:
I tried to run a HD44780 Display connected at the RPI Pico, but that has not worked properly and now I’m trying to initialize the Display manually by settings output_pins.

I’m sure that the display is connected correctly, because I programmed the pi pico in micropython and display worked here flawlessly.

MCU und Output PIN Configuration:

[mcu rp_pico]
serial: /dev/serial/by-id/usb-Klipper_rp2040_E6611C08CB077322-if00
baud: 250000
restart_method: rpi_usb

[output_pin pico_led]
pin: rp_pico:gpio25

[output_pin lcd_rs]
pin: rp_pico:gpio0

[output_pin lcd_e]
pin: rp_pico:gpio1
value: 1

[output_pin lcd_d4]
pin: rp_pico:gpio5

[output_pin lcd_d5]
pin: rp_pico:gpio4

[output_pin lcd_d6]
pin: rp_pico:gpio3

[output_pin lcd_d7]
pin: rp_pico:gpio2

Now to my actual problem:

When I run the lcd_test macro

[gcode_macro all_data_low]
gcode:
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=0
  SET_PIN PIN=lcd_d4 VALUE=0

[gcode_macro pulse]
gcode:
  SET_PIN PIN=lcd_e VALUE=0
  SET_PIN PIN=lcd_e VALUE=1

[gcode_macro lcd_init]
gcode:
  SET_PIN PIN=lcd_d4 VALUE=1
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d7 VALUE=0
  pulse

[gcode_macro lcd_test]
gcode:

  #Default States for GPIOs
  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=1
  all_data_low
  
  #LCD Initializatation (https://web.alfredstate.edu/faculty/weimandn/lcd/lcd_initialization/lcd_initialization_index.html - "4-Bit Interface, Initialization by Instruction")
  lcd_init
  lcd_init
  lcd_init

I can see the outputs behaving correctly with my logic analyzer.

(Picture must be unfortunately external because i’m not allowed to upload more then one picture in my post.)

RS=lcd_rs
EN=lcd_en
Bus:3 = lcd_d4
Bus:2 = lcd_d5
Bus:1 = lcd_d6
Bus:2 = lcd_d7

But when I just make small changes to the macro, it seams that the output_pins would no be handled consecutively.

When I run the extended version of lcd_test macro

[gcode_macro all_data_low]
gcode:
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=0
  SET_PIN PIN=lcd_d4 VALUE=0

[gcode_macro pulse]
gcode:
  SET_PIN PIN=lcd_e VALUE=0
  SET_PIN PIN=lcd_e VALUE=1

[gcode_macro lcd_init]
gcode:
  SET_PIN PIN=lcd_d4 VALUE=1
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d7 VALUE=0
  pulse

[gcode_macro lcd_test]
gcode:

  #Default States for GPIOs
  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=1
  all_data_low
  
  #LCD Initializatation (https://web.alfredstate.edu/faculty/weimandn/lcd/lcd_initialization/lcd_initialization_index.html - "4-Bit Interface, Initialization by Instruction")
  lcd_init
  lcd_init
  lcd_init

  #Initial Function Set
  all_data_low
  SET_PIN PIN=lcd_d5 VALUE=1
  pulse

The my logic anlyzer shows the following output:

RS=lcd_rs
EN=lcd_en
Bus:3 = lcd_d4
Bus:2 = lcd_d5
Bus:1 = lcd_d6
Bus:2 = lcd_d7

From macro defintion I would have expected that the Bus:2 Signal should have been put zero much later.

Maybe I also get the wrong idea of how do this. I’m pretty new to the klipper world.

Can somebody please explain:

Is this behavior correct ?
Are the output_pins processed in a consecutive way ?

Kind regards,
Sebastian

Sineos · February 25, 2022, 11:29am

Macro evaluation in Klipper is kind of black magic. Ref to:
https://www.klipper3d.org/Command_Templates.html#the-printer-variable

And

github.com/Klipper3d/klipper

GCODE Variables don't actually update until a second calling of macro

opened 06:19PM - 22 Feb 22 UTC

closed 08:16PM - 22 Feb 22 UTC

adawalli

I believe I have uncovered an issue in klipper, where the set variable command d…oesn't "take" until a second pass of the macro is executed. ``` [gcode_macro CHECK_RAND] variable_bucket_pos: 4 gcode: SET_GCODE_VARIABLE MACRO=CHECK_RAND VARIABLE=bucket_pos VALUE={(range(2) | random)} { action_respond_info("location is %s" % (bucket_pos)) } ``` The above code will demonstrate the issue. An "invalid" value of 4 is the initial value for `bucket_pos`. If you look at the set logic, we actually expect a value of either `0` or `1`. However, after every `firmware_restart`, you will get a 4! [klippy.log.zip](https://github.com/Klipper3d/klipper/files/8119184/klippy.log.zip)

Buzz · February 25, 2022, 11:46am

Thanks for the reply.

So according to this information:
When I try to achieve a sequence of multiple output pins to be written in an consecutive manner, i need to do a SET_PIN for all outputs, also if I just want to change a single output_pin?

Buzz · February 25, 2022, 12:02pm

So, my Idea from the post before isn’t working also.

I got the same outputs on the logic analyzer if I run this macro. (See first post)

[gcode_macro lcd_test2]
gcode:
  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=1
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=0
  SET_PIN PIN=lcd_d4 VALUE=0

  # Stage 1 INIT
  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=1
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d4 VALUE=1

  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=0
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d4 VALUE=1

  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=1
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d4 VALUE=1

  #stage 2 init
  SET_PIN PIN=lcd_rs VALUE=0 
  SET_PIN PIN=lcd_e VALUE=0
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d4 VALUE=1

  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=1
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d4 VALUE=1

  #stage 3 init
  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=0
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d4 VALUE=1

  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=1
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d4 VALUE=1

  #Initial Function Set
  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=1
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d4 VALUE=0

  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=0
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d4 VALUE=0

  SET_PIN PIN=lcd_rs VALUE=0
  SET_PIN PIN=lcd_e VALUE=1
  SET_PIN PIN=lcd_d7 VALUE=0
  SET_PIN PIN=lcd_d6 VALUE=0
  SET_PIN PIN=lcd_d5 VALUE=1
  SET_PIN PIN=lcd_d4 VALUE=0

Can this somehow be done, what i’m trying to achieve?

Sineos · February 25, 2022, 1:17pm

I’m by no means a Klipper macro expert, so beware my advice:
You could try to do something like:

[gcode_macro MAIN]
MY_SETUP_MACRO
MY_STAGE1_MACRO
MY_STAGE2_MACRO

or if this does not help, then following chain:

[gcode_macro MAIN]
MY_SETUP_MACRO

[gcode_macro MY_SETUP_MACRO]
...
...
MY_STAGE1_MACRO

[gcode_macro MY_STAGE1_MACRO]
...
...
MY_STAGE2_MACRO

This should ensure that the MY_* macros are executed consecutively when calling MAIN.

Buzz · February 25, 2022, 1:42pm

Thanks, but both of your suggestions weren’t working for what I’m trying to achieve.

Neither the main calls child macros, nor the macro chain way.

The error for me is still the same, the output BUS:3 is set to zero before it is called by the macro to do so.

I drawed a red line to indicate, what I would expect the signals to be. And what is programmed in the macro.

But at arround -250ms the BUS:3 pins get set to zero, but from the way the SET_PIN commands were defined in the macro, it should stay at high, and go to low “much” later. Like indicated with the red line.

It seams to me that macro is parsed before, and because there is no change between on the output_pin, the command setting it zero is executed earlier then defined in the macro sequence.

Buzz · February 26, 2022, 9:13pm

I gave up using this approach. I found a way to get my HD44780 clone display working.

Thanks @Sineos for your tips.

koconnor · March 1, 2022, 5:32am

FYI, the set_output_pin updates are all ordered by time. However, your macros above would all be scheduled to be run at the same time, and nothing guarantees ordering between them. (That is, we only strive for set_output_pin to be ordered relative to other events that alter timing - not relative to other set_output_pins which do not nominally consume time.)

To answer the higher level question though (as you likely found) macros are not intended to drive display devices. This is something one would typically implement a klipper host python modules to achieve.

-Kevin

Buzz · March 1, 2022, 6:46am

@koconnor

Thank you for your explanation.

At that point I just had no further idea how to continue on my display problem. And my idea was to do the initialization by toggling GPIOs and see where the difference are on my logic analyzer.

But I managed to get it working. It has just needed a few tweaks in the delays on the utilization of the parallel display interface.

See here:

Topic		Replies	Views
HD44780 connected over RP2040 (RPi Pico) General Discussion	3	1404	March 1, 2022
Run macro from macro? General Discussion	2	795	February 29, 2024
Malformed command 'SET_PIN PIN= General Discussion	5	821	March 15, 2024
Probe triggered prior to movement General Discussion	2	392	May 29, 2024
MCU Config: possible to have 'optional pin' in config message to mcu Developers	1	50	September 1, 2024

Consecutive processing of output_pins in macro (SET_PIN) on rpi pico

Related topics