Why is the smoothing value on the y axis so much larger than the smoothing value on the x axis? Are there additional tests I should do or anything I should do to my printer to allow for better results? Should I go with the reccomended values, or would something else work better? Im a bit confused why the y axis max accel is so much lower than the x axis max accel.
I would not go with any of these input shapers.
The graphs are not that good.
You want to have a single peak on each axis and not such big double peaks.
Such double peaks tell you that there is something loose.
If we see the highest peaks as the ones to look at I would say your belt tension is not high enough.
And because of those peaks being at lower frequencies the smoothing of the shapers is higher than normal.
To better check other things you can post photos of your printer especially of the tool head here.
Tension your belts to about 110 Hz over a distance of 150 mm, zip tie loose cables and fasten all screws!
Where is the accelerometer mounted to?
The resonance behavior of the printer axes are dictated by their mechanical properties: stiffness, weight, degrees of freedom etc.
On printer types where the axes have a completely different setup WRT these properties, such differences in the resonances are to be expected and normal.
The accelerometer is attached via this mount which is held in place by some bolts clamping it to the tool head/bed.
I did some mods to the tool head, specifically new fans for the heat break and for part cooling.
my printer is on the floor rn, but it came with a bit of a wobble between corners in the base which I wedged one side with a thin print, im not sure if that is sufficient for all of the wobble though.
What’s the best course of action here? just tighten every bolt and try to get everything rigid?
You don’t want any loose part on your printer especially on the moving parts that can resonate when excited.
Fasten every screw, try to tension the belts properly, check that your inductive probe does not rattle and have the printer solidly standing on a rigid ground/table.
We see from AWD CoreXY such double peaks but those should be avoidable on a bed slinger.
As those peaks are otherwise clean I think that it only results from another loose part causing that.
And the good thing on a bed slinger is that there is no dependency on the axes. So you can properly tune/setup each axis.
I’m currently redoing this stuff on a CoreXY and this is more work and pain.
So if not already done, correct the axes mapping of your ADXL for each axis and redo the shaper tests on a solid ground after processing the previously mentioned tasks.
Well, I found there was a missing rubber foot causing the wobble and fixed that, then I went around securing a few things and tightening screws and the belts and reran it, and the graph sure is different than before. (I didn’t bother remapping the axes but in both the y direction is parallel to the axis of movement)
The magnitudes on the x graph are a bit lower, but idk why the one peak split into two. (I could hear that while it was running) and on the y graph the vibrations are worse in magnitude, and im not sure about that big second peak which seems to be in the vertical, up and down direction. Im a bit confused now because the printer is much more stable with the 4th rubber foot, I didn’t expect the vibrations to get worse.
Those splits might result from slight racking/binding.
If you move the bed or print head by hand slowly (too fast might kill your board or drivers) how does it feel?
Is the accelerometer rigidly mounted to either the bed or the print head.
Every parts not rigidly mounted will resonate at its own frequency showing a peak or more in those graphs.
Even for the bed the cable might resonate/swing in Z direction or the whole portal might be excited when the bed is tested.
Although its about Vorons and CoreXY the main principles are the same so you might to check this nice video:
Reth is using Shake&Tune however and not the native Klipper feature but the graphs are equal.
