I've been thinking about the difference in backlash between the different ballnut options. For example, when asked what's the standard backlash you get with different ballnut options, people typically say about 0.004 with standard ballnut, 0.002 with preload, and 0.001 or less with double ballnut. Can there really be such large differences between the different nuts? And are those differences due to the ballnuts themselves? For example, for preloaded nuts, you insert slightly oversize (by less than 0.0005) balls than standard ballscrews. Wouldn't this mean that preloaded nuts would only yield a decrease of 0.0005 or less of backlash than standard nuts? I think most of the backlash in different ballnut setups come from other factors like bearing , pulley, and motor coupler play.
For me, it doesn't make sense to spend $1000 on double-preloaded ballnuts for a hobby machine, and then spending only $10 on cheap couplers with lots of play. The "weakest link" in backlash is usually not the ballscrew. As an example, Project5K who bought the standard CNCFusion ballscrew kit, only got 0.0005 backlash on all axis, which is as good as anyone could want for hobby applications. Where dneisler was getting 0.030 backlash from his CNCFusion DELUXE ballscrew kit because of a cheap coupler.
I think more attention should be paid to eliminating backlash by preloading your bearings, getting better motor couplers and timing belts, than spending hundreds of dollars opting for a better ballnut option that only marginally decreases your backlash.
What do you think?