@ Ahren, the zip ties are used to get more clamping force on the threads, With the O-rings I was measuring .005" on the X and Y after a few weeks of use, and .010" of backlash on the Z, I tried beefier O-rings Which helped, except they started cracking in just a few days.

Not wanting to go through the trouble of removing the leadscrews again I used Zip ties, which take all of the lash out of the threads. I have no issues with lost steps with the extra drag, even at 500 IPM rapids. I check for lost steps using a G-27 command, which homes to Optical home switches, and tells me how far home was from the last 0 point. My home switches have many times the accuracy than the steps per inch on the machine. The zip ties are loose enough you can work them off of the nuts with the screws in place. The backlash almost disappears when I wrap tape around the helical couplers. (but the tape doesnt last very long) Before I could rack the axis' on the nuts by hand and see the lash in the indicator. you can see and hear .010" on the Z. There is no play in the leadscrews, they are under tension between AC bearings at both ends, except on the Z where both AC bearings are pre-loaded at the top of the screw.

I check backlash with a Mititoyo .001" dial indicator in a Noga magnetic base.

Your's and dumpsters/Kerks nuts use the same principle, but the main difference is the dumpster and Kerk nuts have much more thread area clamping the screws. I think that having a longer "collet" section and double or triple O-rings, as well as using silicone or similiar O-rings may help.

I'm not so much looking for 0 backlash, as this is an impossible feat, I want the same backlash regardless of wear on the nuts, I ran Dumpster's on our last machine, and after 5 years of 4+ hour a day cutting the backlash was the same as day 1~small enough to not need any compensation in software. We used modified (for 0 lash) lovejoy couplers.

When I have backlash comp set right, I get very good results, a 3" wide piece of .375" 6061 cut with a .25" 2 flute carbide, 40IPM and .02" Z steps, leaving .005" for a full deptch finish pass at 8 IPM leaves me with a part that measures 3.0005" measured with a mititoyo dial caliper. Circles however have spots on the quadrant points that software compensation just can't deal with.

@ Arbo, the caulk backing rod is in with the weatherstripping.