RE: Bad Vibrations
Hi, I am not very familiar with phase stepping. But do I understand correctly that phase stepping is related to input shaping and accelerometer calibration will solve issue? I understood previously maybe incorrectly that input shaping is just acceleration control correction and phase stepping is more of way of controlling the motor itself? Or is prusa accelerometer calibration do both ?
Input shaping and phase stepping are two different things, but they are both calibrated using the accelerometer.
Input shaping will indeed modify the acceleration profiles. It reduces frequency components where the printer has mechanical resonances (and will hence amplify these components anyway). The resonances stem from the printer setup as a whole, where elastic parts (belts, frame parts) and suspended masses form oscillators.
Prusa has not really discussed what "phase stepping" does under the hood, to my knowledge, and phase stepping seems to be their proprietary name for it. From a superficial look at the source code, it seems to modify the timing of the individual motor (micro-)steps, to compensate for manufacturing tolerances in the motor itself.
The goal of input shaping is to avoid overshoots caused by accelerations while allowing the printer to move and accelerate fast. The goal of phase stepping is to avoid resonances (which would cause acoustic noise and/or ripple in the print) when the print head moves at specific constant speeds.
Thank you for detailed answer! Much more clear now. So to be sure, prusa accelerometer calibration will also adjust phase stepping? Or just input shaping?
RE: Bad Vibrations
At the moment, the procedure for automated phase stepping calibration is not enabled in the Core One firmware yet. Phase stepping is already active in the firmware, but it is not using any printer-specific corrections for motor tolerances.
Apparently the Prusa developers or tech support have stated (in Github discussions?) that they are working on enabling the calibration. I have not seen that exchange myself; maybe others can chime in with first-hand updates.
RE:
Here the ink to the Github: https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/4461
Quote from Danopernis:
Correction: phase stepping calibration for CORE One is planned, but we are still tweaking the procedure for 400 step motors and attachable accelerometer (XL has 200 steps motors and fixed accelerometer), therefore we have hidden the calibration from the menu. Toggle will not be present in the menu at all. CORE One is always using phase stepping, without calibration it is just not applying corrections to motor currents.
Reopening the issue for you to keep track, but rest assured, we are working on this even without the github issue being open 🙂
RE: Bad Vibrations
My door rattles, working on a solution
RE: Bad Vibrations
Hi,
I got my Core One on Tuesday. Yesterday I noticed that it causes very strange vibrations at certain speeds, which sounds really annoying. Do any of you have the same problems?I will link 2 videos here and one from a user on Reddit. I asked him for permission to link his video here. Thanks for that.
Video1(Reddit User): https://drive.google.com/file/d/1zIRUpOxvKbxX1axxFDzwJXd2llBJ_nNP/view
Video2: https://drive.google.com/file/d/1xSH24GtOmZrpb0oVeTMGv-muq2-_DV8a/view?usp=sharing
Video 3: https://drive.google.com/file/d/1vn1MmRWzNYe2cmueXFN_ytKZySeAFWbw/view?usp=sharing
Those vibrations are clearly caused by resonance in the sheet metal frame. This is why you need an accelerometer in the print head. One size fits all input shaping is not going to be great for this Core One printer with a sheet metal frame.
RE: Bad Vibrations
Those vibrations are clearly caused by resonance in the sheet metal frame. This is why you need an accelerometer in the print head. One size fits all input shaping is not going to be great for this Core One printer with a sheet metal frame.
I agree with your assumption that the audible noise will stem largely from enclosure resonances. But I am not so sure that input shaping can significantly reduce these.
The accelerometer will measure resonances in the CoreXY mechanism, i.e. the system mainly comprised of the elastic belts and the suspended mass of the Nextruder head. And the input shaper will optimize the input signal to avoid overshoots in this system. Given that the CoreXY unit is only coupled indirectly to the enclosure (via the back and base plate), I don't think the accelerometer would pick up much of the enclosure vibrations, and I don't think input shaping can do that much to avoid them. If it tried to do so, that might even make the print quality worse, since the input shaper would dampen certain drive frequencies which are not resonances of the CoreXY at all, and would hence be under-represented in the print.
Passive dampening of enclosure resonances seems like the more promising approach to me. It would probably be effective to attach dampening mats to the panels (and it would provide better thermal insulation as a side benefit), but there is probably not enough room on the inside, and it would look ugly if attached on the outside. Introducing strips of dampening material between the frame parts and the panels (before they are screwed/riveted on) might be a practical option.
RE:
Those vibrations are clearly caused by resonance in the sheet metal frame. This is why you need an accelerometer in the print head. One size fits all input shaping is not going to be great for this Core One printer with a sheet metal frame.
I agree with your assumption that the audible noise will stem largely from enclosure resonances. But I am not so sure that input shaping can significantly reduce these.
The accelerometer will measure resonances in the CoreXY mechanism, i.e. the system mainly comprised of the elastic belts and the suspended mass of the Nextruder head. And the input shaper will optimize the input signal to avoid overshoots in this system. Given that the CoreXY unit is only coupled indirectly to the enclosure (via the back and base plate), I don't think the accelerometer would pick up much of the enclosure vibrations, and I don't think input shaping can do that much to avoid them. If it tried to do so, that might even make the print quality worse, since the input shaper would dampen certain drive frequencies which are not resonances of the CoreXY at all, and would hence be under-represented in the print.
Passive dampening of enclosure resonances seems like the more promising approach to me. It would probably be effective to attach dampening mats to the panels (and it would provide better thermal insulation as a side benefit), but there is probably not enough room on the inside, and it would look ugly if attached on the outside. Introducing strips of dampening material between the frame parts and the panels (before they are screwed/riveted on) might be a practical option.
On the XL many people made mods to dampen the panels when it was making lots of noise before phase stepping. I did nothing to my machine and when they implemented the phase stepping all of the noise went away.
I guess my point is that you could dampen the panels, but it may be all for nothing if they address the vibration with improved phase stepping. They did say they were working on it. If it were me I'd wait.
RE: Bad Vibrations
I guess my point is that you could dampen the panels, but it may be all for nothing if they address the vibration with improved phase stepping. They did say they were working on it. If it were me I'd wait.
Agree, it probably makes sense to wait for the next firmware release. If Prusa don't get phase stepping calibration released with that, one can still start to play with dampening. (Or if phase stepping does not properly fix the acoustic noise for some units. The open XL does not have those large enclosure panels, and in the Core One I would expect some unit-to-unit variances regarding how well they are clamped and silenced via the rivets and screws.)
RE:
I have the same issue, with strange vibrations at certain speeds, which sounds really annoying. For me it seems to be in the Y axis.
RE: Bad Vibrations
I'm having the same issues. Really bad sounding resonance. I found that rotating my model 45 degrees completely removes the sound (as well as printing at faster speeds). I'm hoping this is fixed in a firmware update.
I also have bad VFA's too.
RE: Bad Vibrations
I have waited since start of december, hope they all dont arrive like this.
RE: Bad Vibrations
Welcome to the harmonics of the hardware you are using, this affects every device, and when one of the parts start to move in certain way it will trigger resonation of other elements, such as side panels, frames, whatever.
For me this happens with Prusa Mini+ when doing initial calibration or some other moves.
The best option to mitigate it would be to avoid certain frequencies if possible in the firmware, or attaching some additional weight to certain elements to lower resonation frequencies of those parts. Both can introduce other artifacts, though.
See my GitHub and printables.com for some 3d stuff that you may like.
RE: Bad Vibrations
This does concern me TBH
I guess we'll see. I did notice that the CNC kitchen review did mention that the printer was louder than previous Prusa printers.
--> MK4 - MK4S - MINI+ - MMU3 - Accelerometer Guide - BambuLab A1 Combo <--
RE: Bad Vibrations
Attaching and using the accelerometer is a simple enough process (link to video guide in my sig), and while it would be lovely to have this built into the printer it's a sensor you do not really use very often. Just on a one-off basis really. Saying that, I'm glad I kept hold of mine to reuse for the Core One. It is handy to have around I guess.
--> MK4 - MK4S - MINI+ - MMU3 - Accelerometer Guide - BambuLab A1 Combo <--
RE: Bad Vibrations
Attaching and using the accelerometer is a simple enough process (link to video guide in my sig), and while it would be lovely to have this built into the printer it's a sensor you do not really use very often. Just on a one-off basis really. Saying that, I'm glad I kept hold of mine to reuse for the Core One. It is handy to have around I guess.
I am not so much concerned about the small effort to attach the sensor vs. having it built in permanently. But am disappointed that Prusa are skimping on the cost and do not include the accelerometer with every Core One.
At the same time they instruct users to set a non-standard belt tension to compensate for the skewed gantry. And they announce that they are working on phase stepping calibration to compensate for motor tolerances and make the Core One run more smoothly. If a customer finds out after the fact that he needs the accelerometer, that's not only extra money for the part and shipping, but also an annoying wait -- in order to fix a weakness which one would not have expected to find in the design in the first place. Not a great user experience, and avoidable if Prusa could swallow just a few Euros of extra cost.
RE: Bad Vibrations
Is there a Prusa guide for the Accelerometer? I have an Accelerometer.
I have really bad noises / vibrations. It is so bad I cannot use it overnight. The printer is on a paving slab on Sorbothane feet, this made my MK4s silent. Unfortunately, the Core One can be heard throughout the house and next door when it's on the loud / vibrating travel speeds. There is a brick cavity wall between the two houses.
I put a decibel meter (phone app) infront of the Core One and it was 53dB when printing, peaking at 63dB when it was making the loud travel movements along with the bad VFA's.
I've attached a video of the Core One sound, you can hear the MK4s in the background that is next to it. I apologise for the potato quality video!
RE: Bad Vibrations
Is there a Prusa guide for the Accelerometer? I have an Accelerometer.
I have really bad noises / vibrations.
There is hope that the acoustic noise (and VFAs) which some units exhibit can be reduced once Prusa enables calibration of the "phase stepping" control. At the moment, phase stepping is already enabled in the firmware, but it cannot be calibrated to compensate for unit-specific motor manufacturing tolerances. Hopefully this will be supported in an upcoming firmware -- and at that point you can use your accelerometer to calibrate and reduce noise.
RE: Bad Vibrations
Link in sig for Accelerometer guide - although its for the MK4s it should be a very similar process for the Core One...
Is there a Prusa guide for the Accelerometer? I have an Accelerometer.
I have really bad noises / vibrations. It is so bad I cannot use it overnight. The printer is on a paving slab on Sorbothane feet, this made my MK4s silent. Unfortunately, the Core One can be heard throughout the house and next door when it's on the loud / vibrating travel speeds. There is a brick cavity wall between the two houses.
I put a decibel meter (phone app) infront of the Core One and it was 53dB when printing, peaking at 63dB when it was making the loud travel movements along with the bad VFA's.
I've attached a video of the Core One sound, you can hear the MK4s in the background that is next to it. I apologise for the potato quality video!
--> MK4 - MK4S - MINI+ - MMU3 - Accelerometer Guide - BambuLab A1 Combo <--
RE: Bad Vibrations
Link in sig for Accelerometer guide - although its for the MK4s it should be a very similar process for the Core One...
But that's for tuning the input shaper, right? I don't think that would make a major difference with the LFAs and noise @milkie has described, since these occur while moving at a constant speed. We will have to wait for Prusa to implement the unit-specific calibration for phase-stepping in the firmware to get these improved, in my understanding.
RE: Bad Vibrations
I correct myself it is on the X axis.
You can also se the vibrations in the print.
Its fine and all with the accl board, but the point is that I should not have to do that. Its something for prusa to fix, no something I have to fix.
I have the same issue, with strange vibrations at certain speeds, which sounds really annoying. For me it seems to be in the Y axis.