What's new
VORON Design

Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members!

Jon's endless battle against VFA's of unknown origin


New member
Staff member
Voron Owner
Printer Model
Extruder Type
Clockwork 2
Cooling Type
Hello people of the internets. I'm creating this thread both to track my progress, and to document this age old battle. This way others can read along and learn. Or perhaps somebody notices something we've missed, or has some smart ideas.

What i have tried so far is ALOT. Hopefully it will be summarized better later on. But my Trident and V2.4 produce more or less exactly the same artifacts. It's quite fascinating and mind numbening.
Since this is the same between two printers i started comparing similarities and differences:

  • Rain&dew bearing stacks
  • Titanium gantry backers
  • Runice idlers and pulleys
  • Ldo frame
  • Both using stock omron probe to avoid additional issues
  • Stock endstop pod
  • Igus cable chain and no pcb
  • Stealtburner toolhead with delta 5015 fan
  • 3x4 reverse bowden
  • 24v everything
  • TMC2209 AND 2240 (tested both)
  • Belts: vt epdm v2 regular ones
  • Size: vt 250 v2 300
  • Linear rails: vt cpc v2 genuine hiwin
  • Motors: vt ldo something v2 bom omc
  • Printed parts: vt fiberlogy abs v2 flashforge abs main and fillamentum asa accent
  • Model: vt rigid gantry v2 flying
  • Front belt tensioners: vt stock v2 beefy
  • Extruder: vt cw2 v2 g2 (think the motor also differs)
  • Hotend: vt rapido v2 dragon hf
  • Hotend fan: vt some random trianglelab fan v2 creality fan
To start with the issue at hand:


Testing part cooling fanspeed to figure out what it affects:

It has AN effect, just not the one I'm looking to fix.

Comparing printed parts between the printers again:

Swapping belts to ll2gt rf bought from mellow on aliexpress:

bottom: old epdm belt
top: new ll2gt rf from mellow

Swapped motors from ldo-42sth48-2504ac to ldo-42sth48-2004ac:
With TMC Autotune enabled:

With stock config (No TMC Autotune):

Is it belts? Nope.

Trying something new with the vibrations measurements for Frix_x:

if running at 70mm/s, a pattern of 1.5-1.6mm is necessary done at 43,75-46.66Hz
  1. Comment this line in the vibration calibration macro (L179 at the end): RUN_SHELL_COMMAND CMD=plot_graph PARAMS="VIBRATIONS {direction}". This will avoid getting it started automatically to allow you to do it manually with the parameters you want
  2. Restart Klipper to take the modification and start the macro. This will produce a lot of adxl345-sp*.csv files in the /tmp folder of your Pi
  3. Using SSH, run ~/printer_data/config/scripts/graph_vibrations.py /tmp/adxl345-sp*.csv -o /tmp/vibration_test.png -a XY --max_freq 150 (check that the filename in /tmp is correct, I'm not sure about the name and can't check curerntly since I'm not near a printer right now)


Well. That adds up. There's noise in that frequency range. Now what?!
And it's a constant that's there regardless of run_current.

Testing again with 200mm moves, and 1000accel/decel:

Which is where i'm currently. To be continued.
Last edited: