I’ll post photos of the Orca VFA test after I run the 2nd one.
- Input shaping should not be dependent on the print surface.
- More information would be quite handy.
I don’t think that @Ltek is literally meaning “surfaces” but instead means “tables” or “benches” that is supporting the printer. I would think how the printer is supported would significantly affect how the resonance testing/compensation would present itself/work.
Definitely more information is required.
2 Likes
Yes, I corrected the title… mounting methods
Creality K1max with printer’s stock very small rubber feet
100lb MDF wood filing cabinet : solid, cabinets 4 walls go all the way down to the padded carpet floor / no ‘feet’
… first chart : printer on a 32lb concrete paver, on cabinet top
… second chart : printer directly on cabinet top
… thirds chart (new attached here) : printer directly on cabinet top + larger stock rubber feet added to printer.
Correct to assume the lower PSD # for the peak the better? If yes, looks like the more I decouple from the cabinet the better?
Cabinet was used two years with my Artillery X2, very solid only vibrates when the K1 is printing at very high speed, never moved with the X2 at all.
Someone in FB forum said this is Ringing, not VFA, that is caused by acceleration being high… my acceleration is high. It just seemed odd to me that the waves are farther apart the faster the print got (standard Orca VFA 40mms bottom to 200mms top)
thanks for the help guys… I spent 2 days trying/testing different things (belt tension tests / frequency, tighten all screws, nothing wiggles/loose, etc) and wasting a lot of filament! I printed a large print that is not usable too, which is how this was found. I’ve been printing small / low wall items til now so never saw much of it.
1 Like
Could I ask that you do one more test and that is the printer on the floor, not on the cabinet? Ideally, the floor would be on grade (ie the ground) so that the printer’s vibrations are transmitted directly to the Earth, which is the most massive thing we have to work with.
In terms of vibration, the printer is a motion source and the cabinet is essentially a spring-damper combination. By putting the printer on the ground, you can get an idea of what the cabinet’s effect is on the printer’s vibration.
With that, you should have measurements on:
- Grade: An immobile surface
- Cabinet: ls the support you’re expecting to use
- Cabinet with rubber feet: I’m expecting that this is basically identical to 2 as you’re not really changing anything (other than add a small spring/damper between the printer and the cabinet which is a larger spring/damper)
- Cabinet with paving block: This will be interesting to see as you’re placing a an additional mass on top of the cabinet spring/damper
I might try some similar experiments here. I’m curious to see how important a relatively rigid platform is compared to a totally rigid one is with regard to vibration and, ultimately, print quality.
Will do test on grade tomorrow and …
Printer (small rubber feet) → Concrete Block → 1" foam → cabinet
… curious if the concrete + foam adds mass and decoupling from the cabinet.
with Printer → Large Rubber Feet → Cabinet; VFA tower looks like this…
Which vibration tower are you using? I was expecting something like the standard one from the Klipper documentation (found here: Resonance Compensation - Klipper documentation).
When you say “decouple” I think you mean that you’re trying to isolate the printer from the cabinet and end up with a surface that doesn’t move.
That won’t happen, what you will be doing is creating a more complex network of masses, springs and dampers which may reduce the vibration experienced by the printer or, more likely, change the natural vibration frequency of the printer with minimal change in amplitude.
I’m interested in seeing your results.
Just took my own advice and did TEST_RESONANCES AXIS=X/Y on one of my custom CoreXY printers on the bench and on the floor. I combined the X & Y results so that they will hopefully appear on the same page display for easy comparison:
I was surprised that there wasn’t more difference in the primary frequencies but expected the difference in the power spectral density.
Vertical Fine Artifacts probably is one of the bigger mysteries of 3D printing and pinning down its root cause seems very elusive.
Known or suspected reasons are:
- Resonances from the steppers
- Belt and pulley pitch
- “Wobble” from the extruder’s gears
Even identifying, which is causing what is highly debated. TBH, I have no clue where to link the input shaper to the FVA phenomenon.
The above pattern can be caused by motor resonance as @Sineos mentioned.
I see and try to fight that on my printers as well. There is a speed band where this pattern is quite dominant and when printing faster its nearly gone.
What I recently read about is that with lower belt tension the belt teeth might become more visible and with higher tension the motor resonance artifacts become more visible.
However depending on the used stepper drivers there are some more influences to ripple/artifacts.
For example the TMC5160 and obviously other as well only store a quarter of a sine wave in their MSLUT causing (rounding) errors/artifacts during reconstruction of a full sine wave/circle.
And the better/more precise a printer is the more it becomes visible.
If you fight this then most likely feeder artifacts (diagonal artifacts) are the next enemy…
point of decoupling is two things…
- reduce vibration transfer between objects.
- decoupling can, potential, allow the mass to move as a “system”, with the vibration instead of fight against it (which can internalize vibrations into the components - extruder, etc)
Good videos on this…
Simple 3D printer noise reduction with a concrete paver — CNC Kitchen
https://www.youtube.com/watch?v=EWxpN_Sw5Pg
People have a misperception that they can reduce vibrations by using foam, rubber feet, etc.
On my old Ender 5 (my do stupid things machine) I did similar tests as well. See [FR] Add accelerometers support to Klipper for input shapers tuning · Issue #3027 · Klipper3d/klipper · GitHub and following posts
Overall very mixed results.
Fully agree with this. Exactly for this reason I belive that bolting down a printer on a concrete slab is a bad idea since all resonance energy must be “drained” in the frame instead of being transformed into non harmful movement of the entire system.
1 Like
I’ve been experimenting with the concrete paver theory… does it reduce the ‘effect’ of the vibrations (on a print) by adding mass to the ‘system’ – then decouple with foam, large rubber feet, etc; from its mounting platform (table, cabinet, etc).
Long ago, I designed speakers and would experiment by placing speakers on an ‘open’ sand box (box without a top) directly. Or make the base fillable with sand. Technical measurements showed it worked to some extent.
For PSD higher vs lower… your article from a few months ago and seems to indicate higher = better?
I guess this is a bit of simplification. As far as my understanding goes:
- Higher PSD = stiffer system
- Higher PSD = more residual resonances even after the shaper
- Exceptionally low PSD = potential problems in the motion system
- One distinct resonance frequency = possible to use a very “sharp” shaper with low smoothing and low impact on acceleration
- Broad resonance spectrum = need to use broader shapers that are causing more smoothing or limitation of acceleration
I did a ton of tests with different configs… interested in what you guys think is the ‘best’ one. Photo gallery of all results Album — Postimages
Generally, all your graphs are looking pretty good, IMO.
What you should aim for:
- One single sharp spike
- The spike at higher frequencies is better, so measures that move the spike to the right are good
- Minimum of additional directions, i.e. for X at best no contribution of Y and Z
LMK what one or two are ‘best’ and I’ll tell you what config they are counter then…
small stock feet w/o spool and holder on back
small stock feet w/ spool and holder on back
large stock feet
18x18 32lb paver → small stock feet
small stock feet … paver on counter next to it
large stock feet
4lb 1.5" polyethylene foam → paver → small stock feet
1/2" grey EVA (ethylene vinyl acetate) foam → paver → small stock feet
Thomas Sandlanderer (hope I got the spelling) did a test with a printer suspended by bungie cords in air to test the theory. I believe he is the one who also did some tests against springy feet, or possibly CNC Kitchen? (I think the former was inconclusive, i.e. didn’t matter much. the later was worse)
Concrete slabs were the best I think (from memory), but what wasn’t considered in the tests I watched was general noise amplification. Sometimes a nice thick solid piece of wood is perfect for allowing your printer to sing its loving harmonies throughout not just A room, but multiple rooms within your household 
I use felt feet under my printer, if the prints suffer slightly, the wrath of imperfections will certainly be magnitudes less than potential alternatives from other human beings within the home.