we are using two radial ball bearings (aka "skate bearings"). since there is minimal thrust and side loads, these inexpensive and small-sized bearings work great.

this time we used v-blocks held to the table to act as the centering support for the leadscrew.

doing this makes it easier to work with the z-travel limited by the torsion spring. (last time we had to temporarily remove the spring).

unfortunately, as we had done last time, a smaller 1/8" bit was used and many passes were made to achieve the right size.

now onto the acetal nut, the last and final step in this upgrade.

we have come up with an anti-backlash acetal nut that should be able to fit within the existing cavity of the mill.

whereas a ballscrew and ballnut combination requires that the table/base cavity be milled out to accommodate the ballnut and flange, the acetal nut can be made into any shape.

this is the regular variety.

the anti-backlash version is being made.

this is quiet a nice upgrade. it's not only quicker to move the table since we've halved the screw pitch from 16tpi to 8tpi, it's also smoother to operate due to the bearings in the new mount and the acetal on metal contact.

even with the gibs tightened down fairly well, the handle is easy to crank.

when connected back to the cnc assembly, we can get very fast travel with no "lost steps" due to binding/torque.

we will be running some tests to show this data.

all in all, a very nice upgrade and it makes a significant difference in the manual and cnc operation of this mini mill.

the only potential downside of this upgrade would be 1. acetal nut wear (guessing here, but this is probably along the same lines as the brass nut anyway) and 2. less finite travel per degree of rotation (in actuality, a degree is a degree is a degree translated into linear travel - it's just that it takes 1/2 the previously used handle rotation to get the same amount of linear travel and this could be more difficult on the human operator to get this sort of "feedback-resolution". of course, since there's still a good amount of friction involved as compared to a ballscrew setup, the operator can still "feel" his/her way around well enough.)

this is the y-axis ballscrew.

after dulling two points of the carbide insert this is what you get.

there may be a way to upgrade without drastic modification to the table and base.

step1

the part really should be mounted better.

there's significant y-axis backlash ~0.007" (the acetal nut wasn't tapped as well as the x-axis one)

this nut took roughly 30 minutes using 5ipm feeds with .040" depth of cut at 1500 rpm with a 1/4" 2 flute round mill.

the shaft was machined and threaded at the end to accommodate these plain zinc washers.

stock acme with the brass nut

upgraded acme with acetal nut

ground ballscrew and ballnut

there are workarounds, but this will require the removal of some cast iron in the table/base.

the x-axis is next...

Display all headersTo: GrizHFMinimill@yahoogroups.com
From: figNoggle <zzz@zzznoggle.zzz>
Date: Sat, 11 Mar 2006 10:07:25 -0800 (PST)
Subject: Re: [GrizHFMinimill] Re: Mini Mill X Y backlash
Reply-To: GrizHFMinimill@yahoogroups.com

well, here's an update/"road test" of the acetal nut performance and yet
another upgrade to ballscrews... (and one very long read)

INTRO

i had some leftover ground 5tpi ballscrew and a spare non-preloaded
ballnut and decided to go all the way and test how the ground ballscrew
and
non-preloaded nut setup would compare to the rest of the setups (stock
acme and brass nut, upgraded acme and acetal nut).

ACETAL SETUP

the x-axis acetal nut has held up well since installed and used for ~100
hours off/on with docs up to .060" with a 1/2" roughing end mill in 6061
al. the primary reason here is that the acetal nut was prepared for the
tap with ~90% thread contact as spec'd by the machinery's handbook.

the y-axis acetal nut was more crudely prepared and exhibited noticable
backlash on the order .007"+. the screws themselves have held up (as you
would imagine since it's plastic/metal contact)

the good news is that the wear on the nuts has not shown as indicated by
1" travel tests with the indicator and dro (also had to make sure the dro
was reading true and it does). travel, while backlash-prone is still
consistant and repeatable during this time. thus, backlash was a result of
the preparation of the tapped holes for the acme screws.

BALLSCREWS

i may be stating the obvious here but the preliminary tests of the
ballscrew setup is quite impressive. _very_ fast rapids (probably not
recommended), repeatable within .001" (actually for the tests i ran, i'd
say it's less than that, but the dial caliper resolution is .001" and i
wouldn't want to go as far as saying that half a tick is exactly .0005"),
backlash was .002" at most.

there may be a cheap way of reducing this without going with full
preloaded nuts by increasing the size of the balls by a hair.

the downside to the ballscrew setup is that in the y-axis, your travel is
reduced by 1/4" or so due to the interference between the nut and base
based on the simple method of creating a straight flange to mount the nut
to. as such, you could grind out the base some more to create a relief and
regain your travel. alternatively, a more involved flange could be made
such that the body of the nut doesn't extend as far out.

i like the accuracy of the ballscrews, but for manual machining, i still
prefer the acetal nut the best as it offers a balance of "feedback" and
resistance while being of higher accuracy than the stock setup. the other
disadvantage of the 5tpi ballscrew is that a hair of a turn of the
handwheel yields usually a .002" jump in linear travel. it's difficult to
get the operator (me in this case) to consistently move in .001"
increments. this can be avoided by getting smaller lead screws, but the
cost then also goes up since 5/8" dia. 5tpi ballscrews are the most
inexpensive of the lot.

STOCK SETUP

as for the stock screw and brass nut, the brass nut came with considerable
backlash on the order of .015"+ even after tightening the nut down, etc.
the improvement here is to make your own or modify the existing nut.
having already modified the nuts, i can tell you that their lifespan
decreases pretty quickly and to get them to this state usually means some
serious binding on the extremes of bed travel. just my experience here,
others may have gotten theirs to work and last.

VERDICT:
($=cheapest, $$$ most expensive)

-stock screw and brass nut - keep them if no cnc and backlash is ok (btw,
i agree with ray here, get a dro and backlash at the handwheel really no
longer matters as far as table positioning is concerned). cost = $0.

-upgraded acme and acetal nut - good balance for use between manual and
cnc. downside is cost of taps to make them yourselves. affordable
in-between solution. straight-forward self-install. cost = $$

-ground/rolled ballscrews and non-preloaded/preloaded nuts - excellent for
cnc, ok for mannual machining, high cost. more work designing a good
solution that doesn't restrict travel and require machining to the mill
itself. cost = $$$

FINAL THOUGHTS

one final note (thanks to "j" who pointed this out to me), i initially
ordered a left hand screw/tap for the x and y axes thinking that the y was
the same as the x. this was an incorrect assumption (should have
checked..). the y has a LH and the x has a RH. so for any of the above
configurations your cost should at least double to get your mill to
operate as stock. with cnc, not a problem since this is configurable. with
manual, you'll have to retrain your brain to operate one of your axes in
reverse mode.