I've been working on a set of anti-backlash nuts for my RF-45 dovetail column milling machine. They are up and running now, so I thought I'd share the idea.
I started my conversion to CNC with these nuts, and I actually bought some 1/2" Nook acme screws, but thought that it doesn't make much sense for to just chuck the RF screws into the dumpster. I decided that it makes a lot of sense to keep the RF screws and make nuts for them.
I designed some nuts that are very low friction, anti-backlash, very stiff in the torsional and axial directions, and most importantly have a long life, and not allow for degradation in stiffness with wear. I also have made the preload spring loaded, so the nuts will be wear compensating. The nuts have two identical nut halves, one is fixed to the mounting block, one floats and is spring preloaded. The advantage of this vs. the split type that is preloaded radially inward by a spring is that the entire nut half must be worn away before the assembly will be subject to backlash. A split type radially preloaded nut needs only to wear away the last few threads a bit before backlash can happen.
For this I came up with a design that has a large bearing surface area to resist torsional loads. I used four shoulder bolts as linear slide rods through bored holes in the nuts. I used precision 3/8" dia. shoulder bolts from Holo-Krome to perform this function, as they are ground and are very straight, have a nice surface finish, very good concentricity with the thread, controlled shoulder diameter. Four of these bolts yields a bearing area of 1.5 sq. in. in each rotation direction.
I also decided that this much bearing area is of little use if I cantilever the nuts out on the bolts a long distance, therefore I put the moveable nut right next to the support block, the springs are the outboard element, loading the bolts in tension, and allowing me to use compression srpings, but only cantilever a short distance (depends on preload, but in the range of .2in or so).
Since the RF-45 has a lot of room under the slides (with the exception of the Y axis on the side near the column), I used 1" thick black acetal nut halves. This should give a very long wear life for these nuts. Basically it was just as easy to machine small thin nuts, as to machine large nuts that I will not have to worry about for some time.
The tricky part with the RF screw was that it is an oddball size. Mine are both 0.920 major diameter, 0.820 minor diamter, 10 tpi (of course since they're lead screws with dials 0.1"/rev). I figured these would follow the ANSI standard for thread geometry, and deviate only in the diameter. I lathe cut these internal threads with the minimum standard clearances for an ANSI acme thread, and found I had to loosen up to about the maximum of the ANSI tolerance, and they spin quite freely. They actually seem to have lees friction without grease on the threads than with, but I am still using a very light machine oil to keep the screws from rusting.
I think this design concept might be useful for many of you that are in the same boat as I am, and could be scaled up or down for other sizes (standard sizes would be a lot easier!).
I'm just learning this stuff and am curious how this is assembled. Do you loosen the four bolt/screws, thread the leadscrew into the movable nut then into the fixed one, holding them apart as the screw starts into the fixed one....then tighten the 4 bolts to get the preload? As Ricky used to say: "Splane it to me Lucy."
Also, the mounting block looks to be aluminum...is that right? Could you post a pic of the tool you cut the threads with? Is it an insert tool or did you grind one yourself?
I might try this on my X-3 which has too much backlash to suit me. It's not going to be CNC'd anytime soon but it'd be good practice to reduce the backlash and friction in the 'screws.