Resonance, spreadCycle and stealthChop
-
@deckingman said in Resonance, spreadCycle and stealthChop:
I'm a bit confused by that. Did you really mean CoreXYUV or something else? It's just that a CoreXYUV use two gantries, stacked one above the other. So the belt lengths are the same (but you have more of them).
each gantry has it's own belts. So instead of 2x3m I have 4x1.2m and motors don't play against each other (I do not get resonance on diagonals).
@deckingman said in Resonance, spreadCycle and stealthChop:
On the resonance thing, does it impact on print quality or are you simply trying to reduce the noise? If it's the former, there is a very simple solution.
no, I don't care about the noise, it's a performance/quality issue.
-
@matt3o I'm also confused by what it is you are moving to. Can you post a diagram of the new belt layout? I assume your comment about not having resonance on diagonals refers to the "old" corexy setup?
-
@gloomyandy said in Resonance, spreadCycle and stealthChop:
I assume your comment about not having resonance on diagonals refers to the "old" corexy setup?
yes, on a corexy I don't get resonance on diagonals. In what I'm building I have individual X and Y axis so only one motor moves to go in straight lines.
This is an old picture but should give you an idea. (yeah to be honest I don't know what is the name of this config. I believe I've seen it referenced as corexyuv once).
-
@matt3o That ain't CoreXYUV. It looks like Cartesian with two X motors and two Y motors. If you try to run CoreXYYUV kinematics on that, bad things will happen.
-
This, is CoreXYUV https://www.youtube.com/watch?v=mTPV3Ss1D-4&t=752
-
@deckingman my bad
-
@matt3o said in Resonance, spreadCycle and stealthChop:
@deckingman my bad
No worries.
BTW, the reason why you don't get resonance on a CoreXY when doing diagonals, might be speed related rather than simply the fact that just a single motor is employed. The way the belt paths are arranged on a CoreXY, means that when doing pure X or pure Y moves, the belts actually move something like 1.4X the carriage distance. So for any given carriage move at any given constant speed, the motor speed could be 1.4x different purely depending on the direction of travel. Easy to test - try doing 45 degree diagonal moves at 1.4x the speed that you get resonance on pure X or Y and see if the resonance returns. (EDIT - or it be 1.4X slower - can't remember which way round it is).
And finally, if you don't care about the noise but want to improve print quality (assuming that you really do have issues due to resonance), then simply add mass to the hot end. That probably goes against everything you've read because the 3D printing community has a fixation about reducing moving mass (despite the fact that mass is not a limiting factor of print speed). Any vibration due to the motors will be transmitted to the hot end via the belts. This vibration can set up a sympathetic resonance in the hot end if the mass is low (think tuning fork). By adding mass, you lower the resonant frequency of the hot end (think tuning fork with very thick prongs) so that it won't vibrate - or at least, it will vibrate at a much lower frequency.
It's easy to test - visit you local tyre fitting company and scrounge some stick on wheel balancing weights. Stick one on the hot end and be amazed
-
@deckingman thanks for your suggestions. I just finished connecting the new machine and I'll check if this issue is present with this configuration as well.
Have you seen my video above? It's not really just resonance, the whole printer vibrates
I believe it mostly comes from the carriage on the gantry (basically on the head).The x1.4 diagonal speed test is something I should have done... and I really don't remember if I did... anyway I'm on the new config now... lots of rails, lots of motors, I hate it already but if it works...
-
@matt3o said in Resonance, spreadCycle and stealthChop:
Have you seen my video above? .........................
Only the one you posted on 29th Jan - is there another? Difficult to to tell what's vibrating due to camera shake (or maybe an earthquake)
-
@deckingman or maybe the printer shaking!
-
@matt3o said in Resonance, spreadCycle and stealthChop:
@deckingman or maybe the printer shaking!
When the carriage is stationary?
-
@deckingman it's that bad!
-
@matt3o said in Resonance, spreadCycle and stealthChop:
@deckingman it's that bad!
Are we talking about the entire printer rocking slowly or something else? I thought your problem was high frequency resonance (as per title) rather than a very low frequency wobble. If it's the latter, surely you should be looking at fitting some levelling feet and maybe stiffening the frame, rather than messing around with motor settings.
-
@deckingman I was kidding about the shaking. Sorry if it wasn't clear. It's a high frequency resonance that starts from the motor shaft and reaches the head carriage. No amount of weight on the head makes it any better.
-
@matt3o That isn't obvious from the video.
But moving on ....... I note from your first post that you are using 0.9 degree motors. This might sound completely counter intuitive but have you tried 1.8 degree motors? If so, was it better/worse/the same?
Perhaps post you configuration files too - maybe one of us might spot something that could have a bearing on what you are experiencing (although I can't think what off hand).
In any case, let us know how the alternative kinematics works out.
-
Oookay.
I finally tested the new kinematics as per picture of few posts above. Resonance is completely gone. From 1000mm/m to 20000mm/m seems all good. I get just a tiny bit of noise around 3000mm/m but no nasty vibrations. ( @garethky just FYI).
I feel the most likely cause is the length of the belt, but I'm no expert. I don't think the problem was with my corexy setup as I tried two completely different belt routing and they both had the same effect.
Anyway the new system will be overkill but works... so I just have to design a new head now.
@deckingman said in Resonance, spreadCycle and stealthChop:
But moving on ....... I note from your first post that you are using 0.9 degree motors. This might sound completely counter intuitive but have you tried 1.8 degree motors? If so, was it better/worse/the same?
yes I tried, the resonance just moves to another speed.
-
@matt3o Well you have a solution which I guess is the main thing. But I doubt the original cause was the belt length - primarily because my CoreXYUVAB (3 gantries) has a frame size of 600mm square (it's the one I posted a video of) and none of those CoreXY kinematic gantries exhibits the behaviour you describe. But hey - you've found a fix which is the main thing so who cares.
-
@deckingman what size is the belt on your printer?
-
@matt3o said in Resonance, spreadCycle and stealthChop:
@deckingman what size is the belt on your printer?
All 6mm (and a tad less than 5 metres long IIRC).
And before you ask......Nema 17s @2 amps on XY with a moving mass of about 1.5Kgs in the Y direction. Nema 23s @2.8 Amps on UV with a moving mass of about 3Kgs in Y. Nema 17s @2 amps again on AB (can't remember the mass off hand).
Default print speed is 80 mm/sec although I have printed at 300mm/sec with a 6 input hot end feeding equal amounts of filament via 6 Bondtech BMGs. Default non-print speed is 350mm/sec. Accelerations set to 2,000 mm/sec^2 (IIRC).
Never had any resonance issue and "ringing" just doesn't happen (I put that down to the mass and rigidity of the hot end and it's mounting arrangement).
But, when one starts throwing circa 5 Kgs around at 350mm/sec, it can make the entire printer rock around (hence the load balancing AB gantry which just moves lumps of lead in the opposite directions to the XY and UV gantries).
-
@deckingman I'm clueless, I really tried everything I could and wasted dozens of spools of filament to reprint parts.... corexy is not for me. But yeah mine is (was) a lightweight compared to yours.
Do you have a duet3 or duet2? Could it be the drivers?
