Zero backlash ballscrews

I have a new set of ballscrews for my Bridgeport Series I machine that
purportedly are zero-backlash. I bought these for my CNC upgrade
project. I'm trying to verify that these are indeed zero-backlash.
This set is a direct replacement for the stock acme lead screws.
One might think the easiest thing would be to look up the part numbers,
but these are apparently new old stock from 20 years ago, made by Han
Jiang Machine Tool Works in good old China (PRC). They do look nice and
have an inspection certificate that, if true, is pretty good ground
quality. Also came with a replacement yoke to fit the ballnuts. But no
specs for backlash. Google (and Google groups on rcm etc) turns up
nothing on the part numbers (BS-001A, etc) for that manufacturer,
although this seems to be the business renamed Hanjiang Tool Co., Ltd.,
that makes various precision machines today
formatting link
. After 20 years,
the cosmoline was as hard as shellac in places, so it was quite a clean-
up job, but it did preserve the steel perfectly.
Back to the backlash issue, I thought first to install the Y screw to
just give it a try and hope for the best. My dial indicator showed
0.002" or 0.003" of backlash. Blah. It occured to me that this might
be just the wrong kind, or worn out, bearings. Do you need different
bearings to eliminate backlash on the table screws? Or perhaps a shim
washer or something? I suppose that much backlash in the bearings
wouldn't matter with acme screws, and it might just be a "feature" of
the stock bearings.
I suppose springiness in the yoke could be another source, but that much
seems excessive.
I should have tried pushing the table to isolate the backlash possibly
to the nuts themselves, but I didn't think to do that before tearing the
machine down for some way scraping (another story).
After some more musing, I thought clamp the screw in a soft-jaw vise and
measure the backlash directly with the dial indicator. I was encouraged
to see that there was no perceptible play indicated by pushing the nut
axially. There is a tiny sort-of backlash in that when you reverse the
direction of the nut (the screw being fixed in the vise) it takes a tiny
bit of rotation, perhaps 1 degree or so which translates to about
0.0005" travel, before the axial motion actually reverses. Is this
perhaps normal zero-backlash behavior, just part of the springiness of
the mechanism (again, no axial play is present as hard as I can push by
hand).
It seems to me that consistent backlash in the drive train (as I found
in the nut reversal, or would be found in timing pulleys and belts on
the screws) doesn't really matter if you have CNC compensation, what
matters is the lack of axial play from preload force, and that seems to
be below my indicator's resolution.
Another confirmation is that the nut (a single nut, with two
recirculation tubes) requires a few in-lb of torque to turn; this would
seem to be a characteristic of preloaded ballnuts? My understanding is
that non-preloaded nuts will spiral down freely just on their own weight
if held vertically.
So any expertise in proving zero-backlash, and just how close to zero
that should be in practice on the retrofitted machine, would be a big
help to me.
Richard J Kinch
Palm Beach County, Florida USA
formatting link

Reply to
Richard J Kinch
Loading thread data ...
Richard J Kinch wrote in news:Xns94CA23A6E62FDsomeconundrum@216.196.97.131:
From what you describe, these are indeed zero-backlash screws. Your freeplay noted is obviously in some other part of the mechanical drive train. You should probably have angular contact bearings where the screw mounts in the machine. These should be preloaded to factory specs for the bearings to eliminate free travel of the screw itself. How, depends on how the screw is mounted. There are several different mounting methods for ballscrews. Some run with one end completely free, some run with a roller bearing on the free end, and opposed angular contact (or tapered roller) bearings on the mounting end, and some are threaded on each end. Screws threaded on each end (where the opposed axial bearings are on each end of the screw, *Usually* require pre-loading of the screw (stretching), this is done in high-accuracy machines to reduce heat expansion.
Reply to
Anthony
Yup..and the sheer weight of the table may cause some compression effects on the bearings and ball nut. This is where "taut" ballscrews excell. The ballscrew is pulled tight between thrust bearings on both ends. Its gonna be hard to do on yours, without all new bearings and some work..and then it might bow the table a smidge.
Gunner
Some may decry the impeachment of William Jefferson Clinton as a national trauma, but not humorist P.J. O'Rourke, who thinks the proceedings are a win-win situation -- and grand entertainment to boot. Mr. O'Rourke, writing in the latest issue of the Weekly Standard, acknowledges that "some earnest souls have gone so far as to aver that impeachment has distracted President Clinton from ... raising taxes, destroying health care, appointing 1960s bakeheads to high political office, soliciting felonious campaign contributions, hanging friends out to dry for Arkansas real estate frauds, giving missile secrets to the Chinese, taking credit for the benefits of a free market about which he knows little and cares less, using U.S. military forces as fig leaves for domestic scandals and au pairs for the U.N., leading foreign policy back into the flea circus of Jimmy Carterism, having phone sex, groping patronage seekers, and snapping the elastic on the underpants of psychologically disturbed school-age White House interns entrusted with the task of delivering high-level government pizza." Ouch. Tell us what you really think, P.J. "No matter what, Bill," Mr.O'Rourke concludes, "your girlfriend's ugly, your wife hates you, and your dog can't hunt." --------------------------------------------------------------------------------
Reply to
Gunner
Take the cranks off and put an indicator on the end of the screw--then whack the table/ saddle both directions with a 2 x 4.........
If the indicator registers screw movement, your bearings or mounts are loose.
A pair of angular contact bearings set to a slight preload, and completely capured axially within the bearing housings is the usual method of eliminating axial pla of the leadscrew.
A properly adjusted cog belt drive is perfectly acceptable and contributes little if any noticable inaccuracy when performing general tolerance classes of machining.
Reply to
PrecisionMachinisT
I was curious whether the standard Bridgeport table bearings are angular contact and what their play should be. I observe that replacement parts in the High Quality Tools catalog
formatting link
for the table bearings are listed alternatively as "grease sealed ball bearings" or an optional "precision bearing set".
Thanks to all for the helpful responses.
Reply to
Richard J Kinch
Sorry, not really all that familiar with the Bridgeport, etc. retrofit arena.
Probly a match up for size in the MSC catalog.......thats *if the housings are engineered to suit bearings in pairs in the first place.
An angular contact bearing comes apart quite easily in on direction, thats why they are used in pairs if there is thrust in both directions.
Suggest do a google search for some bearing application data if you are not clear on the differance between these and standard ball and roller bearings--outside of my scope and typing skills to give much other than some basic info here.
Cheers,
Reply to
PrecisionMachinisT
I understand the variety of types. My question is, just what type of bearings did Bridgeport put on stock machines' acme leadscrews.
Reply to
Richard J Kinch
Best I could do is guess--Im not gonna go there.
LOL....
Somebody here should be able to answer though, you would think.
Reply to
PrecisionMachinisT
I can only make an educated guess (based on the part number and presence of seals) that the bearing pair at one end of each axis is a preloaded matched pair of deep groove bearings, not an angular contact pair.
As long as the pair is preloaded (no backlash) I'm not sure why the distinction would matter.
Ned Simmons
Reply to
Ned Simmons
It all has to do with contact angle.
Takes a certain amount of pressure to eject the inner race of a standard roller bearing.......In either direction......
While you can press out an angular contact bearing easily in one direction, you will meet with considerable resistance in the other.
This is why they are typicaly used in pairs, with thrusts opposing......
Reply to
PrecisionMachinisT
What I meant was that in this case there wouldn't be much difference in performance between a properly preloaded pair of deep groove bearings vs. an angular contact pair. The loads are pretty light compared to the capacity and size of the bearings.
Ned Simmons
Reply to
Ned Simmons
A follow-up to my own post some months ago:
I discovered the problem was that I had flipped one of the ball bearings around. These are labeled as Fafnir 204KT4. My decoding of that part number indicates these are 204K size (ordinary cheap radial bearings, not angular contact) but in a non-standard thickness (T4) that appears to be 15.5mm instead of the standard 14mm, which make up a matched pair of thickness 31mm. And I assume these are ground or otherwise fitted to each other to make a poor-man's thrust pair, like an angular contact pair, but with only radial bearings. One side of each bearing is shielded, and one side is open, with the open sides obviously intended to mate with each other. This is the crossfeed axis on a standard Bridgeport Series I vintage 1970s.
With the bearings installed properly the backlash is now under 0.001".
Perhaps somebody has more details on these bearing codes and specs?
I note the MSC catalog Bridgeport parts list gives a "basic machine" bearing as "H204KTT" and the "leadscrew" (which one? both?) as "RM204FT4-DB" (where the "DB" would seem to indicate angular contact type?).
Reply to
Richard J Kinch

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.