Basic Information:
Printer Model: Custom build based on RF2000 with Voron Stealthburner and EBB36
MCU / Printerboard: ATMega2560 based
Host / SBC: Raspberry Pi 3
Baudrate: 250000
I’m using klipper already for some years and it is working great with max_velocity 350mm/s. When trying to push the settings to the physical limits by increasing max velocity/accel I got some unreliable results. So today I tried it more systematically and measured the G1 move speed (via mainsail console) first on X-axis alone, then XY combined, then XYZ combined.
X-axis: no problem to move with 450mm/s.
XY combined: I get this “reschedule timer in past” error. I could make it working with 375mm/s.
XYZ combined again results in a failure with 375mm/s.
- I would like to find out the bottleneck. Is it on the raspi generating the ramps (or is this done on the mcu?), or is it the bandwidth via serial line, or is the MCU the bottleneck.
- How to find out the theoretical max values? How does max_velocity and max_accel affect these values?
- Does XYZ combined really matter? I believe that slicers normally only move on XY until the layer is finished. Then they move in Z, but not XYZ combined.
- Extruder is another load. So probably XYE combined is an additional load for the Raspi, but not for the MCU, because the extruder uses a different board. That’s why finding the right bottleneck again is very important.
I hope someone with more knowledge about the klipper code can give me some insight here.
Thanks a lot.