Linear movement analysis, vertical pattern
Can someone help me with this graphs? I can't find any information about how to "read" them.
I have a vertical pattern on my prints, more visible under 160mm/s.
11 Replies
printer size? (effects vibe freq range go/no-go)
magic-beige•13mo ago
300 , sorry
decoupled steppers, gates pulleys, gates toothed idlers front side of xy joiners, gates belt, 3.1 fitted on the other side of xy joiners, running cooled steppers, orbiter, dragon st flow, ebb36 ( separate adxl)
top right graph (above) most peaks should all be below the 150-200 hz range. That double peak in the top left graph tells me you've got freeplay in your carriage to linear rail or something in the carriage is loose enough to vibrate itself. Belt tensions for a 300 size should be roughly in the 80-100hz range (the smaller the object the higher the resonance freq)
my own similar graphs but on a VC500
magic-beige•13mo ago
But excepting that vibration at the carriage you don't see something else being wrong? I had problems with binding belts in the past. Thanks!
your multi-spike graph.... you should be able to spot linear rays of spikes starting at zero and radiating out across all the speed lines at various frequencies IF you've properly eliminated all vibrations see the examples at https://github.com/worksasintended/klipper_linear_movement_analysis
GitHub
GitHub - worksasintended/klipper_linear_movement_analysis
Contribute to worksasintended/klipper_linear_movement_analysis development by creating an account on GitHub.
(last pic I drew on)
??? binding belts? how are the belts binding?
magic-beige•13mo ago
before they were binding and I was able to see only one peak when I measured the belt tension, now is fine. I was wondering if you can tell something from those graphs if the belts are fine ( not binding )
I will reprint all the toolhead parts first, thanks a lot @ptegler !
Before a reprint.... consider remounting your ADXL on a different axis and or location and rerun the vibe tests. I've heard of the wires people are using to the ADXL itself was even causing the bulk of their analysis readings
magic-beige•13mo ago
thanks for the tip!
magic-beige•13mo ago
Sorted it. This is how the graphs looks now, 130-160-180 speed for quality.
