RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
If it's a stationary part resonating it may be possible to trigger it with a 246Hz sine wave (not sure if phone speaker would be enough but maybe worth a try).
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
Buzzing goes away at even the slightest uneven tuning. [...]
I think we're close to "solving" the puzzle. If we get rid of the vibrations that cause the nasty resonances at 40/80 mm/s, we might improve the VFA a lot too.
Maybe that's what Mr. Prusa referred to in yesterday's Reddit post about VFAs, when he mentioned "belt tuning"? It's not sufficient to set the right tension on each belt, but they also need to be fine-tuned relative to each other? "The well-tempered Core One" being the goal... 😊
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
I really have to unplug one motor at a time, while running my G-code...
So, even without unplugging -- do the resonances never occur during 45° movements (when only one motor is active)?
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
Hello,
I have the same problem with my fully assembled printer. The first 250 printing hours were very quiet, but suddenly this printer became very loud. In the video, it sounds much quieter. I have spent several days trying to find the cause of the issue. I disassembled and reassembled the entire gantry. I also nearly completely disassembled and reassembled the print head. Additionally, I removed and reinstalled various metal parts. No matter what I do, the noise remains.
https://drive.google.com/file/d/1DMZi49BFamTJPUfH8Qegom4EtgtUeRzI/view?usp=drive_link
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
So, even without unplugging -- do the resonances never occur during 45° movements (when only one motor is active)?
Rotated my "head movement only" project by 45°, here's the Video and frequency analysis.
It's the same 246 Hz, so no need for a screenshot.
In the video, I did the mistake of using 56.57 mm/s and 2x it. I did another test with 28.29 and it the same super annoying sound for 45° lines.
However, very nice info: Two lines of the 45° square are "fine". Not super quiet, but fine. The other two lines (right motor) is really nasty.
I guess I'd have to take out both motor, swap the pulleys and put them into each other's position as the next step.
G-Code is attached. No nozzle checks etc. being done, since it will only home, move the print bed down a bit and start moving around in thin air.
No matter what I do, the noise remains.
I have no idea what changed for your printer, but apparently it's motor related and is only really noticeable, when using the standard structural preset. Did you use another preset until that time or has the part you're printing now some smaller layers, where the printer needs to slow down for cooling and runs into the 40 mm/s or 80 mm/s?
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
Amazing work Raaz. Good idea to swap the motors and perform the same test. Hope it's not too difficult with the cables. Really curious for the results.
RE:
I had the exact same idea to print 45° prints to isolate both motors and see what happens (reducing speeds by 50% to match the gantry speed relative to X and Y).
Your test seems to confirm that it may come from a motor.
I wonder what would happen if you replace the faulty one to a new?
Many motor seems to be faulty, and some people get crazy sounds, maybe because both of their motors are bad, and others do have the resonance noise, but less pronounce because only one is bad?
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
I guess I'd have to take out both motor, swap the pulleys and put them into each other's position as the next step.
Great progress, and I agree with the follow-up. "Is the motor or the pulley to blame?" is the remaining question. Based on the frequency you measured, my bet would be on the motor. (Uneven microsteps which let the full-step frequency bleed through.)
I forget -- did you try the preliminary phase stepping calibration in firmware 6.3.3 yet?
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
j’ai une artillerie x4 et elle a des driver en 2209 bien plus silensieux un imprimante à 230 € hormis le bruit du ventilateur, elle marche bien et pas de problème de vfa avec une carte mère en 64 bits et la emmec a 10€ en 32go avec adaptateur pourquoi payer 6 fois le prix et elle ne fait pas trop de bruit avec le ventilateur de la hotend sur off.
RE:
Hope it's not too difficult with the cables. Really curious for the results.
Sadly, it's currently "impossible" for me to access the rear side, swap motors etc...
I don't really have a place to put the printer on to do it efficiently.
I'm extremely curious too, but I might not be able to do it until after a few deadlines at the end of July.
I hate it.. Since I apparently have one really smooth motor and one quite bad motor (or pulley), so I could do some really clear A/B-testing, arghs.
reducing speeds by 50% to match the gantry speed relative to X and Y
You mean reducing the speeds to sin(45°)*40 mm/s -> 28,29 mm/s?
When recording the 45° video, I used the wrong triangle in my head and calculated 2x the correct speed.
The attached video in my next post has the correct speeds and mimics 40 & 80 mm/s, but separates the motors.
Many motor seems to be faulty, and some people get crazy sounds, maybe because both of their motors are bad, and others do have the resonance noise, but less pronounce because only one is bad?
I really hope that Prusa either improves the Phase Stepping Calibration or might offer some "extra silent motors". Or maybe the community will find high quality replacement motors.
Based on the frequency you measured, my bet would be on the motor. Uneven microsteps which let the full-step frequency bleed through.
Yep, 100% agreed. It also matches my tests to dampen the resonance by touching all parts accessible with my hands during the slow 45° noise test and not being able to dampen it even the slightest amount.
The only effect I could achieve was to make the Y-bearings starting to buzz or not.
I forget -- did you try the preliminary phase stepping calibration in firmware 6.3.3 yet?
Yes, I ran both Input Shaping - & Phase Stepping Calibration.
I'm eager to test my 45° test g-code after resetting my printer and delete both IS & Phase Stepping values.
RE:
Next post with an update on my testing (thought splitting would be good to keep the overview):
Now with the better video, with the speed set to 28.29 mm/s to match the "40 mm/s" when doing 0°/90° lines but with 45° lines.
At this slower speed, truly matching the "40 mm/s", you can clearly hear the right motor causing the nasty resonances and the left one being a lot quieter.
With 2x (and 4x) the speed, like in the last video I've posted, the left motor also causes some resonances, but is still quieter.
I then printed a little, single walled, square at the same, slow speed, but couldn't see any real VFAs on any of the walls and definitely no difference between the "loud" walls and the "quiet" walls.
But when printing another, bigger square, but this time 2 walls and 97 mm/s, testing the motors independently, the Ripple Artifacts are massively different and look like this (Elegoo PLA+):
(Touching/pressing against ANY accessible parts doesn't change anything at all. I even pressed against the belts, but really, no difference at all.
Next test will be holding my mini bassshakers against different parts while sending a 246 Hz sinewave to them.)
Loud, right motor:
Quiet, left motor:
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
Why 97mm/s for the larger test print? The difference in VFA is shocking. What's a bit baffling, though, is that for the VFA it's likely the motors do not play a role but the pulleys do. But I can't imagine the pulleys having this effect on the motor resonance. Yet your print shows a clear correlation.
Would you be able to swap the pulleys and do the same sound test and print? That's a lot less work than swapping the motors and would be very interesting to see.
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
Loud, right motor:
Quiet, left motor:
Wow -- that is a very significant difference indeed. 👍
But also quite confusing for me... I had just convinced myself that VFA (belt ripple) must be caused by a mismatch of belt and pulley teeth, while the acoustic noise must be caused by uneven motor micro-stepping. And hence, that phase stepping calibration should be expected to make the printer run quieter, but that I should not get my hopes up that it would reduce VFAs.
Now what...? 🙄
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
Might be correlation but not causation. That's why I asked to swap the pulleys. If our hypothesis is correct, the VFA's should swap with the pulleys while the resonance stays with the same motor.
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
Why did you try to mimic 40mm/s but in diagonal instead of 80mm/s that seems to be the highest resonance speed?
(I may have missed something)
Thanks for the testing, that’s really interesting!
RE:
Why did you try to mimic 40mm/s but in diagonal instead of 80mm/s that seems to be the highest resonance speed?
(I may have missed something)
40 mm/s, for my printer, is resonating stronger, than 80 mm/s. In the earlier videos, I've used both speeds, but I wanted to go as slow (and loud) as possible, to be able to touch/press parts of the printer and have more time per move.
Why 97mm/s for the larger test print?
That's just one of my "quiet speeds in any direction" and a good trade-off between speed vs. structural. Nothing calculated!
Both motors aren't really loud at that speed and similar regarding vibrations. Thought that would give a fair example.
Would you be able to swap the pulleys and do the same sound test and print? That's a lot less work than swapping the motors and would be very interesting to see.
That would require loosening both tensioners to their maximum, taking the top + steel side panels off to access the screw in the bottom of the frame and taking out the motor from the top, correct?
A bit annoying with the filament sensor in the right side panel, but should be doable without moving and accessing the rear of the printer.
Or is there a way to access the motor screws from the inside, while the side panels are attached?
Now what...? 🙄
Using this quote for both of your thoughts about VFA vs vibrations:
Yep, I also don't know what's going on there.. Especially since I don't really see any ripples when using the silent/nasty 28,29 mm/s.
I really have to try to adjust my next weeks a little bit to be able to swap pulleys/motors, so we can find out if vibrations & ripples are linked to each other or I'm just having a bad pulley on a loud motor by coincidence.
In the meantime, I've researched these 3 motors that should fit without any adjustments and should be higher quality (maybe an adapter cable needed):
Moons MS17HA2P4100
Lin Engineering 4118S-62-07
LDO Premium: 42STH40-1684L100
Anyone any experience with these or more knowledge? I'd be willing to buy 1-2 motors and test them.
This is a short, interesting, but demotivating read:
https://www.reddit.com/r/3Dprinting/comments/htfi2t/moons_09_steppers_on_marlin/
RE:
Sadly, you can't edit posts here...
I've looked up the 3 motors and it's the usual AI BS stuff...
Motors that could actually be an improvement, but aren't easy to get from Germany:
https://www.linengineering.com/products/stepper-motors/hybrid-stepper-motors/4209-series/4209M-01S
https://www.moonsindustries.com/p/nema-17-high-precision-hybrid-stepper-motors/ms17ha2p4100-000004611110015917
To actually post something on-topic, here's the 45°-square at 56,57 mm/s (80 mm/s equivalent for 0°/90° lines):
Ripple:
Right Motor (bad):
Left Motor (good):
Noise:
Interestingly, at 97 mm/s, the noise difference changes to "only fan noises" vs. "I think there's a motor":
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
Someone on Reddit claims the resonance on their X motor is fixed after a replacement from Prusa:
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
Someone on Reddit claims the resonance on their X motor is fixed after a replacement from Prusa:
That's consistent with what @raaz-2 found above, I'd say: Noise frequency correlated with the motor step frequency; only one axis (i.e. one motor) affected in his printer.
I would hope that properly calibrated phase stepping can offer a firmware solution which avoids a motor replacement, but the current preliminary implementation does not seem to have worked for everyone. Maybe there are types of motor flaws/tolerances which can't be compensated by adjusting the microstep spacing?
RE: Prusa Core One not usable inhouse... Sorry, but the sound was to intensiv
Someone on Reddit claims the resonance on their X motor is fixed after a replacement from Prusa:
Yeah, but that noise is crazy bad before the replacement and my printer is nothing like that kind of bad noise wise.