best rigidity mods for SO2?

[Hans]

I don't see any reason to change out our bearings. Internal clearance does nothing for us because we preload our bearings. Runout in cheap ABEC-1 bearings this size is worst-case around .0003", certainly way better tolerance than the V-wheels are made to and also better than our rails' tolerance. Plus, all we could hope to improve is the positioning accuracy -- RIC and tolerancing don't affect the rigidity of the bearing. The bearings are one of the stiffest parts of the machine already.

[WillAdams]

Thanks! I added a brief note on that, following your text and a footnote link back to your post.

ISTR thinking on this a bit more, a previous mention that the bearings were amazingly good for what one pays for them and quite well-suited to their usage in the machine.

[chomdh]

When I suggested lower RIC on the bearings I was thinking that would help reduce the twisting of the z axis plate about the acme rod and z makerslide. If different bearings aren't the answer are there any other suggestions? I do have 2 more v wheel kits on order. I'll install those see what happens. Just trying to beef up what I can before installing my DNP611.

[cvoinescu]

The bearings I see around aren't really that good (I'm looking at early Inventables, more or less current Inventables, year-old OpenBuilds, and the ones I have -- they're all essentially the same, by the way). As far as I know, the ABEC class doesn't do much for the worst problem, which is the fact that the races can twist with respect to each other. Coupled with the elastic Delrin V-wheel they're embedded in, this allows the wheels to twist slightly when subjected to off-axis loads. This is why one-piece dual-row V-groove bearings are so much better: for those to twist, there would have to be radial play.

[Hans]

That's a good point, I wasn't thinking about side loads and twisting. Plain ball bearings can twist about 1 degree without much resistance, but when you mount two of them and hold them in a housing like our V-wheel shell I doubt there's much angular play. If it was just one bearing, we could see .008" axial motion, but at that short distance I imagine even the Delrin is pretty stiff holding the two bearings aligned to each other. Axial play in the bearings is probably enough to pay attention to despite preload, and C2 RIC might work to reduce that, though only testing could tell for sure. I imagine the simplest improvement is to increase preload on the Delrin wheels to help keep the balls centered in their grooves. The steel V-wheels have integrated angular-contact bearings, which are really ideal for our purposes. Has anyone figured out how long plain makerslide will stand up to the steel wheels?

[cvoinescu]

Has anyone figured out how long plain makerslide will stand up to the steel wheels?

The anodization goes in very short order, but the wheels continue to run fairly smoothly while digging into the aluminium. At some point (after a few hundred hours of use, if I recall that conversation correctly), the rail becomes unusable, because the center section is worn more than the sides, and any adjustment is too tight for the ends and too loose for the middle. I'm sure precision degrades way before that.

[AlBorland]

I did some quick measurements on one wheel of my z axis.
Axial clearance (dial test indicator mounted on X cariage, probe on outer race of bearing) measured to ~0.06mm with fairly low force, but did not increase any more with higher force. That tells me that the radial preload does not do much for axial clearance.
Deflection of delrin wheel and makerslide contact (indicator mounted to z makerslide, probe on outer race of bearing) increased fairly linearly up to 0.12mm with medium force applied.
Depending on the ratio of lever arms, the result at the tip of the cutting bit can be better or worse.

The measured axial clearance falls pretty nicely within the theoretical range: Clearance class MC3 for miniature deep groove bearings specifies 5..10µm radial clearance, axial clearance is not specified but is usually 8.5 .. 10 times higher than the radial clearance, giving 42..100µm.
Note: ABEC/ISO only specify manufacturing tolerances and precision, but not the clearances designed into the bearing.

Just exchanging the bearings would not do much, especially since better solutions exist (steel rails + V bearings like dualvee).

[Xaracen]

"And while you're at it, put the front and rear X V-wheels on single, long bolts, with metal spacers, going through both plates (although it's a bit harder to match the distance between the wheels to that between the Vs of the MakerSlide accurately)."

I did that on my v1 shapeoko, and it certainly helped, but I got an even bigger benefit beyond that when I replaced the z-axis v-wheel bolts on the aluminium z-plate with longer bolts that extended through the x-axis motor plate in the same fashion as the x-axis v-wheel mod, though I had to drill receiving holes in the motor plate (and move one of the idler wheels a little to make room). That extra mod made a huge difference in the rigidity of the carriage at just the places where the torsion forces from the z-axis actually apply themselves to the z-plate.

Before I did that but after I had extended the x v-wheel bolts, I could actually see the z-plate flex as I pushed on the z-slide. Once I extended the z-wheel bolts as well, that flexing is now invisibly small, and the swing at the toolbit tip is now the tiniest fraction of what it once was.

Z-wheel bolts extended to x-motor plate.

z-wheel mod 1.jpg (74.22 KiB) Viewed 2078 times

Moved the idler wheel to the left to make room for the new bolt.

z-wheel mod 2.jpg (69.05 KiB) Viewed 2078 times

[WillAdams]

That's brilliant!

Next mod for my SO1.

[ThatOtherGuy435]

Damn, I've been messing with redesigns of my S1 Z axis for months now trying to resolve that flexing plate. I was almost ready to completely toss it out and rebuild to the S2 standard, even though I had concerns about the flex in that (smaller, but steel) plate.

Now it looks like I'm just going to order some long bolts. Brilliant!