So I have been trying to learn more and figure out why these graphs show a Z-axis peak on my Y resonances.
A little info:
The printer is a RatRig v-core3 300 (CoreXY).
ADXL is mounted “on its back”, correct orientation for X/Y/Z directions.
I tried multiple sensors.
I re-assembled the entire gantry and toolhead assembly.
Loosened and tightened belts to compare how it effects resonances
Attached are the latest measurements taken for X and Y axis.
I’ve been searching and asking for more information, but mostly I hear that its “oke”, or “expected”, or that the orientation is wrong. Occasionally I hear an honest “don’t know” :).
In my reasoning the orientation alone doesn’t explain both a Y and a Z peak in the same graph, when movement is only on X or Y.
Anyway, the real question as stated above is that I would really like to understand why that Z bump is in my Y graph. From what I gather on several Discord servers is that its quite common, but no one really seems to know why. Some graphs though don’t show the Z bump which makes me question and curious about my own measurements.
Thanks!
Maybe a bit of definition will help: What is a mechanic resonance:
- You apply a periodic force on a system, the excitation frequency
- You measure the “answer” of your system with the accelerometer
- Depending on the excitation frequency the system will only answer weakly (low amplitude / low Power Spectral Density) or stronger (high amplitude).
- A local maximum is called resonant frequency (the peaks in you graph)
- Every single part in your system may react differently to the given excitation frequency. This means every part has its own “natural frequency” at which it answers with a high amplitude
- What you measure is the complex answer of each single aspect of your system
Generally the system’s answer is not bound to a certain direction. This means an excitation in X or Y direction can lead to a resonance in Z.
You will barely find an answer, as to why you have a certain resonance in one or another direction. This depends on so many factors:
- Excited mass, e.g. typically Y axes are heavier than X axes
- Stiffness differences in the different directions
- Loose / moving parts that have a certain degree of freedom in on or the other direction
- etc.
If you really want to find the answer you most likely will have to bisect your system, for example:
- Remove hot-end, then measure
- Change belt tensions, then measure
- …
In the end even this might be futile if your entire frame is giving this particular answer.
1 Like
Thank you, that is very helpful indeed.
Does the graph and the scale allow for converting to a travel distance on the Z axis?
In other words, would a certain amplitude suggest loose parts or play on the gantry or toolhead?
Or perhaps the tolerances on the rails/cartridges (0.04mm I believe) are enough to cause that peak? Because if that is the case, it might suggest binding of the cartridges on the rails in systems with no peak on Z …
Lastly, would frequent and repeated resonance tests cause more stress on components than regular use (I’d think so but I have no idea), and perhaps cause early fatique on items like printed parts?
I do not believe that such conclusion are possible. Neither the spectrum nor the amplitude do tell you where to look, IMHO.
That being said, I have seen graphs that looked perfect, i.e. no resonance at all and low power density. Turned out that the linear rails where binding so hard that no vibrations could be generated.
Of course. The amplitudes / strength of the resonance is much higher than the exciting forces. It is to be expected that such testing puts strain on bearing, joints etc. Also screws that are not secured by some thread locker might eventually come loose.
Some general remarks, also WRT your graph:
- Low frequency resonance are more evil than high frequency ones. They cause more ringing. → Your Y frequency is pretty low
- More aggressive shapers cause more smoothing and reduce your maximum recommended acceleration → The recommended shaper for Y is the nearly the most aggressive one, thus giving you a max accel recommendation of below 3600
- Again the Y graph: The additional peaks on X and especially Z contribute to this recommendation as the algorithm “sees” a very broad spectrum
I would rather choose the MZV shaper and try to figure out, what is causing these excessive Z- spikes. Maybe remove the entire toolhead and mount the ADXL there instead. If you get the same graph, then at least you know it not caused by the toolhead.
Then go on to work on the frame / linear guides. Make sure everything is square, Y linear rails are parallel and perpendicular to X, no loose screws etc. Maybe add some frame stiffening measures.
You could also try to remove the build platform or move it fully down and see if this contributes.
1 Like
Thanks for taking the time to answer, I appreciate it.
I’ll look into it some more and take your advise at heart.
So I’ve been having a go at this almost continuously. I did sleep … I think 
.
How does this look?
Lets compare old vs. new
X:
- Power Spectral Density: 46000 vs 13000
- Peak: 60 Hz vs 40 Hz
While the intensity got lower, the peak moved to the left (lower frequency) → I would rather consider this worse than before
Y:
- Power Spectral Density (main peak): 17500 vs 72000
- Main Peak: 30 Hz vs 25 Hz
- Z Peak Power Spectral Density: 12000 vs 16000
- Z Peak: 80 Hz vs 75 Hz
Intensity dramatically increased and also shifted a bit to the left → Definitively worse than before
Interesting. What did you do to create these changes?
Disassembled the gantry and reassembled it.
Its depressing. Initially, I thought those looked pretty good but as I thought about it I came to the conclusion that it wasn’t good at all and I really don’t know what I am doing 
.
The belts were too loose on those new graphs. As I started tightening them and testing each few turns the initial graphs started coming back. Nothing has changed after dis- and reassembling the gantry.
One good thing came out of it, it turned out the gantry wasn’t perpendicular to the frame extrusions. It wasn’t off by much but at least that is corrected. The carriages on all rails run very smoothly and there is no binding at all on those.
Maybe I should just accept it as it is and actually print something instead of fiddling around with things I don’t know much about .
Here are the latest graphs. I think I am done tinkering unless you (someone) really believes theres something wrong and can suggest where to look.
Thanks for the help, its appreciated.
1 Like
