Possible threading issue in Klippy?

General Discussion
21

I think not, atleast I didn’t change anything, I use relative extrusion but not relative movement.
This is the start of the gcode i printed:

M107
;TYPE:Custom
M104 S0
M140 S0
START_PRINT EXTRUDER_TEMP=243 BED_TEMP=70 
G21 ; set units to millimeters
G90 ; use absolute coordinates
M83 ; use relative distances for extrusion
; Filament gcode
M107
;LAYER_CHANGE
;Z:0.25
;HEIGHT:0.25
;BEFORE_LAYER_CHANGE
G92 E0
;0.25


G1 Z.25 F13500
;AFTER_LAYER_CHANGE
;0.25
G1 E-5 F7200
G1 X47.935 Y47.026 F13500
G1 E5 F7200
;TYPE:Skirt/Brim
;WIDTH:0.44
G1 F2880
G1 X49.317 Y45.867 E.07245
...
22

Your slicer is issuing the G90 after the START_PRINT, so if the printer is already in relative mode you will receive this error. Add G90 to the beginning of this macro and the error will disappear.

23

I’ve been using klipper now since two weeks, but I think I never pressed pause before. So the pause macro left it in relative state. Yeah that looks like it was the culprit.

24

My colleague, who has also seen this problem, runs on a Pi3, so I agree that processing power is not the main problem for the out of range commanded movement. But in my experience, thread issues tend to surface more often on weaker computers as the probability of two threads running concurrently increases.

When it comes to the CR10 Smart (Pro)/Creality WiFi Box issue, I watched the printer as it died on the chimney on the Benchy an hour ago, it stuttered a couple of times before it shut down, so the CPU is genuinely running out of steam. I can make it print by slowing down the accelerations for outer perimeter from 2000mm/s^2 to 1000, that gives it just enough breathing room to survive. But it is marginal. From the deck up the Benchy was running at a load >1 with just a single web client. Decrease in free memory seen in the graph earlier in the thread is all cache and buffers, available does not change much.

25

I have never pressed pause, so that is not the issue for me.

26

Without a log file and the accompanying gcode, it’s not possible to say what caused this error.

That said, out of range errors are almost always caused by incorrect bed mesh configuration, or a mismatch between absolute and relative positioning.

27

It was indeed the switch to relative in the pause macro causing this, I’ve now added G90 and M83 to the START_PRINT macro and the error is gone. Thanks for your help.

28

Revisiting my START_PRINT macro I see there is no explicit G90 before the bed leveling, so that would cause the issue if something has set G91 for some reason. Will update the macro.