Raw resonance data needed


I am working on some resonance tuning improvements, and I’d need some raw resonance testing data. I’m basically interested in the data for all kinds of printers - with low, medium, and high resonance frequencies. I cannot promise any concrete improvements, but your help would still be very much appreciated! Basically, one could run


and provide the generated csv files with the raw data (probably compressed). And you could also post the input shaper parameters that you are currently using (if any).

Voron-24-raw_data.7z.txt (5.5 MB)
shaper_freq_x: 91.4
shaper_type_x: 3hump_ei
shaper_freq_y: 39
shaper_type_y: ei

Tiny-L-raw_data.7z.txt (5.7 MB)
shaper_type_x = mzv
shaper_freq_x = 70.2
shaper_type_y = mzv
shaper_freq_y = 52.6

Thanks for the data, much appreciated!

Voron2.4_raw_data_y_20230403_091938.zip (7.6 MB)
shaper_freq_x: 71.2
shaper_type_x: mzv
shaper_freq_y: 38.4
shaper_type_y: mzv

Thanks for the data! Just a small question, I see that the resonance data suggest different optimal values (around 45 mzv on X, Y axis is much closer, around 37 with the current algorithm). Did you somehow arrive at different optimal value for X, or is it maybe just because you didn’t run the resonance tuning for some time and the stiffness drifted with time?

Yes I have not run the test in a very long time.
I haven’t got around to changing as yet the parameters, but that is a good point might need to check the X carriage for tightness of bolts etc.

raw_data_x_20230405_174851.zip (7.6 MB)
raw_data_x_20230405_172801.zip (8.0 MB)
a couple of tests 100mm either side of bed center on the Y axis

raw_data_y_20230407_155236.zip (7.2 MB)
raw_data_x_20230407_155000.zip (7.5 MB)

shaper_freq_x: 63
shaper_type_x: zv
shaper_freq_y: 44.8
shaper_type_y: mzv

Cartesian Ender 5 Style, since we already have Vorons

I know this is “yet another Voron”, but here is data from my 2.4 that’s 300 mm spec in XY and 350 spec in Z (giving me an exact 300 mm cubed print volume with kinematic bed mounts). I had to re-run the resonance sweep anyway because I just replaced CW1 with Galileo. The switch to Galileo and about 500 hours of print time since the previous calibration made no appreciable difference to the recommended input shaper.

I previously characterized the printer on my print desk, on a carpeted concrete floor and on a solid concrete floor, all at Z heights of 20 mm, 150 mm and 280 mm, and all of these combinations at room temperature and then at hot chamber temperatures (that’s right, a total of 18 resonance measurements). Based on all this data, I previously selected the following input shaper parameters using the hot chamber temperatures only:

shaper_freq_x: 62.8      # enclosure at ~50°C, print accelerations <= 11600
shaper_type_x: mzv
shaper_freq_y: 44.2      # enclosure at ~50°C, print accelerations <= 5800
shaper_type_y: mzv

I am keeping these unchanged based on today’s re-run.

raw_data_x_20230409_173541.zip (3.6 MB)
raw_data_y_20230409_173319.zip (3.6 MB)

If I get a bit more time I will also get data from my heavily modified CR-10S Pro. It’s a bit of a circus with a bed slinger…

Thank you for the data! Yes, a large bed slinger data would also be very awesome to have.

No problem!

For the sake of science I collected three data sets for you, all from the Y axis only:

  1. Stock bed with WhamBam magnet only, no print surface (REFERENCE)
  2. REFERENCE with Creality borosilicate glass print surface (+953 grams)
  3. REFERENCE with Mandala Rose 300mm Voron 2.4 aluminum bed (+1589 grams)

Pertinent printer configuration includes 0.9 degree 2A Moons’ stepper on the Y axis, Gates belting, Mellow pulleys, OPENBUILDS POM V-wheels that I adjusted to near perfection (zero perceived play with virtually no preload - this is equivalent to about 2 hours of work on the CR-10S Pro since the Y axis design is completely over-constrained in the XY plane). The MCU is a Duet3D Mini 5+ running latest Klipper with 128 microsteps on XY steppers.

For amusement, this is what the test setup looked like with the MRW bed on top:

Note that I used my Voron ADXL mount with double-sided adhesive tape between the mount and the bed, additionally clamped to the bed with clips, as shown on the photograph. The ADXL wire is also strain-relieved by a clip attachment to the reverse Bowden tube, to avoid whipping. I know that results would be quite different if the accelerometer was placed in the center of the bed, but I do not think it matters for your data collection purposes).

REF.zip (3.7 MB)
REF+953.zip (3.5 MB)
REF+1589.zip (3.4 MB)


1 Like

Heavily modified bed slinger Kingroon kp3s

shaper_freq_x: 85.8
shaper_type_x: ei
shaper_freq_y: 29.0
shaper_type_y: ei
raw_data_kp3s.zip (4.0 MB)