Ringing Part 9: Vibrations, bushings, and 3-point mounting

So this is how the build platform vibrates when it is being moved along the Y axis (and no other parts of the printer are moving):

PLA LM8UU-style bushings is this style of object.
Small-contact-area (SCA) PLA bushings is the design I showed you on the previous page.
LM8UU steel bearings are the bearings that came with the printer.

One way to interpret the graphs is to look for the bearing/bushing style that reduces build platform vibration to the minimum on all non-moving axes (X and Z). In that case, the steel bearings are the worst and the LM8UU-style bushings are just a bit better than the SCA bushings.

But the more important thing to look for is how the Y axis moves with those bearings/bushings installed. Ideally we would want to see a really gradual and smooth curve, which would mean that the bearings offer little resistance to the stepper motor and can accelerate or decelerate smoothly. The LM8UU-style bushing has a very high acceleration peak and the curve of those peaks is really steep, like the motor had to go to a higher speed to get it moving in the first place, and as soon as that force overcame the stiction, it kind of flew out.

It generally looks like the steel bearings came out on the bottom here, but there’s something to be said for the rigidity of the bearings. The graphs don’t seem to show it because it wasn’t a violent motion, but when I was using the SCA bearings I could see and feel the bed tilting along the X axis as it was being dragged back and forth. Certainly not ideal.

3-point carriage suspension

The i3’s plate comes with five mounting options for bearings even though there are only 4 bearing blocks provided. There’s the typical one-in-each-corner configuration that the printer uses by default, and the fifth mount is located halfway along the right side, and that’s for 3-point suspension. Since it takes 3 points to define a plane and 4 points to over-constrain it, I compared vibration between 4-point and 3-point suspensions in case this was creating a lot of friction.

Again, this is the vibration of the bed while it is being moved along Y only, with no other parts of the printer moving. The SCA results didn’t seem too different between the suspensions (I think they were too broken-in already), but the steel bearings were interesting.

Vibration along X increased, but not much. I call that no real change.
Also no real change for the steepness of the curves or the peak acceleration on Y.
But far less vibration on Z, in line with the smooth SCA bushings! That single peak at the start was when the carriage collided with the end-stop because I had pushed the bed too far back before running the gcode.

Now here’s the real question: Did the Z vibration drop because the carriage was no longer over-constrained, or because I could choose not to use the roughest bearing?

The Y carriage is too flexible

A common complaint about the cheaper i3 builds, including the Wanhao machines, is that the Y carriage (the metal plate where the bearing blocks and the heated bed are attached) is far too thin and flexible. This leads to people being unable to properly level their beds because the carriage keeps flexing underneath and pulling the corners out of alignment. Indeed I’ve also had this problem, which I guess is why Wanhao instructs users to first screw down all of the springs fully and then raise them all by a uniform amount. That also explains why there are a lot of thumbscrew designs on Thingiverse with gradations marked onto them; they’re there so that people can rotate each thumbscrew by a known amount. I haven’t found a design that I like and I think that 3D printing doesn’t have the resolution for really nice gradations anyway, so I’ll probably just print or draw some stickers to apply to the OEM thumbscrews.

More to the point, when I was having those problems with levelling the bed because of the flexible carriage, I also noticed that the bearings ran with a lot more friction than if the spring tension was uniformly distributed so that the carriage was not warped (as much). So much tension that it felt like I was using newly-printed PLA bushings. When I pulled the SCA bushings out of the printer, I could also see that three of them were mostly worn along the top faces only, but the fourth bushing was evenly worn. So it seems like this one bushing was properly aligned, and the bed was pushing all of the others down hard.

In my mind, this flexibility can only hurt my attempts to remove the ringing from my prints, plus it’s a total pain in the ass. My seeming inability to level the bed using a dial indicator makes total sense now. So I’ve ordered a 6 mm composite carriage from eBay and it should be here within two weeks. It’s a 3-point suspension together with a setup for 3-point bed levelling, and it looks like the only thing I have to do to set that up is to drill a hole into the heated bed where there is no circuitry.