Topic: Bushings or bearings? Doesn't matter.
I've read through bunch of articles on the internet about users of 3d printers replacing their 3d printed/plastic or bronze/brass bushings with linear/ball bearings and vice versa.
I've come to conclusion that unless you have step skipping from friction (and I'm not 100% sure on this either) bushings give no difference from bearings in terms of print quality.
Owners of 3d printers using linear bearings mod their 3d printers by switching to bushings and vice versa.
There's few reasons people switch but I honestly think none is worth the time.
Linear bearing printer owners mod their printers to use bushings instead to reduce the noise, but I don't get how that bothers them when there's more noise from fans and motors.
And then some swap their stock bushings with bearings to get rid of slack, but in reality its too small to make any noticeable difference on a modern FDM printer (3rd link below). Not to mention cheap bearings aren't perfect when it comes to tolerance either.
Here are few of the many links of people switching from bushings to bearings and vice versa, comparisons and good points:
http://www.reddit.com/r/Reprap/comments … s_vs_igus/
https://groups.google.com/forum/#!topic … 60SiLLrnf4
http://forums.reprap.org/read.php?1,221763
https://www.youtube.com/watch?v=tCUEYB1U6UY
http://www.thingiverse.com/thing:526660
REAL LM8UU bearings (like $6 each from Misumi, not the $0.80/ea Chinese crap) are also super quiet.
Can be a source of backlash [the play in a bushing], yeah. Although gravity tends to keep them pretty well centered. The carriage's weight is trying to accelerate it downwards at 9800 mm/s2, and your bot acceleration is probably under 3000mm/s2, so gravity is definitely winning the fight on where the bushing sits on the rod.
If your bot is still, the bushings will contact the rod at 0 degrees (top of rod). Let's say you crank up your bot to 9800mm/s. The bushings will sit at 45 degrees now. If you have 0.1mm of play between the bushing and the rod, the net horizontal deflection due to shifting from a 0 degree contact to a 90 degree contact would be 0.05mm. Then at 45 degrees, it's sqrt(2)/2*0.05 for about 0.035mm error. Double that to find the backlash, 0.07mm. That's pretty small -- rod flex at that insane acceleration is way larger.
And my fav from all the links:
Honestly, it's kind of a "grass is greener" thing. A lot of FlashForge people upgrade their carriages from linear bearings to bushings, and a lot of Makerbot people upgrade their bushings to linear bearings.
I hope these links helped clear some things up.
If you're a Solidoodle owner and your carriage is damaged, then go for bearings, because the designs of the stock carriages are not open source and printable and the ones on sale are too expensive and the only alternative carriage we have on thingiverse uses LM8UU bearings.
If, however, you have no problem, I wouldn't advice replacing bushings with bearings just for the sake of improving your machine's print quality, because it likely won't.
It was already pretty evident from the direct drive mod results by several people that any backlash issue is caused by the short looped MXL motor belt (read this: http://www.soliforum.com/post/81693/#p81693), not bushings, bearings, shafts, motors or VREFs. But here's some more evidence.