I want to be able to run my Chiron during night-time, but my neighbours tries to sleep at that time of day. I have put a lot of effort into making the printer silent. The controller-fan and power-supply-fan now run slower and only when needed, I have put the printer on those damper feet found on Thingiverse, I have added rods to stabilize the gantry and I have switched to TMC2209 for X, Y and Z. I have made sure the TMCs are tuned for 24V and are using 'stealth chop' below 225 rpm (120 mm/s). The result is impressive compared to how noisy the printer was when I bought it.
One problem was not solved, when Y-axis moves at between 35 and 45 mm/s, the printer starts to loudly vibrate at 220 to 280 Hz. To mitigate that, I first tried to slow down printing, but then print quality was hurt by oozing. Now I have reverted back to my desired print speed, but limited Y-axis feed-rate in Marlin to 32 mm/s using
M203 Y 32. It is quiet enough, but many movements are slowed down, affecting quality.
What can I do to stop or mute these vibrations?
ADDED 2021-03-30 I have no proper accelerometer, so I used a MPU6050 accelerometer and some code in an ESP8266. It samples roughly 400 times per second and gives a vibration-value that can be used to estimate the level of vibrations. It is not frequency-compensated, but gives a hint of what is happening. And it can be used to compare settings.
I also made a program based on an example in TMCStepper so I could run the stepper and see/change all settings. This program also relays the measurements from my accelerometer so I can make graphs.
To do this I removed the stepper from the chassis, as I need to be somewhat quiet when doing these tests. It looks like this:
This graph sums up my findings pretty well: X-axis shows motor-speed in full steps per second, fsps. fsps corresponds to the frequency of the noises I want to conceal. 200 fsps = 1 revolution per second = 60 rpm. On Chiron Y-axis 60 rpm = 32 mm/s = F1920.
Y-axis is shown with the scale log10(accelerometer values).
What I read from this chart (and my many measurements): TMC2209 is much quieter below 120 fsps. StealthChop is fantastic at low speeds. StealthChop shall ONLY be used at really low speeds in a 3D printer. If at all, since switching mode also is noisy. SpreadCycle use 256 usteps/fstep. A4988 use 16 usteps/fstep. I think that is the reason for TMC to be more quiet at low speeds.
What I do not understand: Why is TMC2209 seems to be more loud at higher speeds. It might make different vibrations that my makeshift accelerometer rates worse, of course. But can I have a defective TMC2209? Or are the BigTreeTech boards bad? Or have I got fake TMC2209 with worse performance?