Lathe - Backlash

Backlash Fixes (Oct 9, 2001)	Backlash adjustment (Sep 14, 2003)
More backlash (Oct 11, 2001)	Backlash Fixes (Nov 12, 2003)
Backlash fixes etc. (Oct 11, 2001)	Backlash (Sep 22, 2004)
0.100 backlash (Mar 15, 2002)	Backlash and movement of carriage (Dec 12, 2004)
Variable backlash (May 28, 2003)	Backlash and accrue (Dec 26, 2004)

Backlash Fixes
I'm trying to find a way to get zero backlash on a 9" south bend lathe, over a cross-slide travel of about half an inch, in order to do some high precision taper work with a sort of single motor CNC taper attachment setup. Zero backlash is necessary because the stepper motor will be retracting the tool in the direction of cutting force. I should also mention that this is not a production setting - i.e., it only needs to work for about an hour at a time, and some rebuilding between uses is acceptable. There is no way a ballnut will fit, and I don't want to mill out the channel in the cross-slide like I did on my bigger lathe. So far I have two ideas. 1) Install a spring that will push the tool away from the work. Using a fairly long one, the variation in force over half an inch of travel won't be large. The total force can be quite large, as the stepper actually only has to retract the tool, never advance it. This seems like a simple fix (which could be suited to longer travel with some sort of constant force device) but I never hear of it being used. OR 2) Buy some new screw stock, obtain a loose fitting nut, and inject moglice (the Teflon-matrix-epoxy bearing rebuilding material). I'd use new screw material to get consistent unworn threads, sacrificial material for any needed lapping, to be able to simply throw it away if I don't get the release agent right, and to maybe choose an ideal threadform. Anyone done this successfully? What would the ideal threadform for moglice be? Chris (1765)
Your ideas on zero backlash are interesting but first I'm curious as to why the cut has to be in the retracting mode. Can the cut be made from left to right instead of right to left (or vice versa)? Anthony (1780)
You don't need zero back lash -- the force of the cut will take out the back lash and anyway there is no such thing as zero back lash so you might as well accept the back lash you have and allow for it. Yasmiin (1781)
Actually, I do, as the tool has to retract in the direction of the cutting force a controlled amount, while the cutting force varies with the changing depth of cut. And no there is not zero backlash, but there is backlash measured in tenths, and I need it. The tapers in question are 36-40" long, so the compound slide is out. And they are not linear, so a taper attachment won't work (yes, a cam follower would, but CNC is simpler and more flexible) Chris (1782)
I'm sorry, but this simply isn't true. Cutting forces are not always sufficient to overcome table friction. When the feed opposes the cutting force, things work well. When the feed is with the cutting force, you can't count on the force being enough to push the tool back as the screw groove that is limiting it retracts. The reason for having an antibacklash screw and a stiff setup is to make finish diameter a function of tool position, relatively immune to cut depth and tool advance/retract. Chris (1784)
Damn and I have been doing this for years without getting dimension errors -- I sure must be stupid -- just over looked all those errors -- sure do wonder why all those pieces fitted together though. Perhaps you might loosen you gib screws a bit and save your self a lot of trouble -- but then what do I know -- I am just a stupid machinist. Guess I will just have to go out and replace all my feed screws on my 4 Monarch 10EE's -- what you don't learn on the internet. Yasmiin (1785)
Seems to me there is a reason 10EE's cost more than 9" South Bends. In a deep cut (.125 off radius), the force vector is almost along the spindle, leaving very little perpendicular component to push the cross slide around on an old, worn dovetail. Loosen the gibs and it will probably chatter. There is a reason CNC machines use ballscrews, and it's not just for fast rapids. Chris (1786)
Well now, wasn't that terribly helpful and courteous! Thank you so much, Ma'am. Nowhere in this exchange have I seen anyone trying to imply you were stupid, except you. From my interpretation of Chris' original message, I would have to agree that your comments were off base. When making a profiling cut, where the tool may move back and forth on the x axis, he would definitely need a zero backlash screw, or something close. Using CNC, the backlash can be programmed in, so that when the screw reverses, it reverses an additional amount. In fact, that is how at least some CNC's work. Even with ballscrews, there is some backlash, admittedly, it is measured in tenths or millionths. I believe the only reasonable solution for Chris is to modify the lathe to accept ballscrews. One drawback that he should be aware of is that they will tend to move on their own, from the cutting force. With CNC, this is not a problem, since a motor holds the screw in position, but on a manual machine, you will see the crossfeed dial start to move. There are small diameter ballscrews available, but the nut is usually too large to fit the channel in most small lathes, such as the SB or Logan. Another possibility, for short run, might be to use a nylon or other plastic nut, which can be cut slightly undersize. This might allow a nearly zero backlash nut, while still allowing some rotational movement. Of course, a truly zero backlash setup is virtually impossible, since it would too rigid to allow rotation. Oh, and by the way, making a comparison between the South Bend 9" and a Monarch 10EE is not exactly fair either. Scott Logan (1787)
If I can't get satisfactory results on the little lathe I will move the job to my 13", on which conversion to a ballscrew is about half finished. Unfortunately that lathe lives 45 minutes away from home... which makes using the little one desirable. I may try the retract too far and advance trick, however I'm not sure what that will do to the dimensions if I don't pause the Z axis feed, or what it will do to the surface finish if I do. Chris (1788)
Wrong -- you and he still don't understand but if you want to go out and blow money on "zero" back lash when you don't need to then you go right ahead but I have been doing this for a long time and you don't need to do this unless you just like spending money on a pointless upgrades that can be done without doing all this nonsense. You still have to compensate for the back lash and all CNC machines do this. The main reason for the ball screws as that they can reverse directions rapidly, and they handle the speed of a CNC machines better than conventional acme screws. I was trying to teach you guys something but maybe you will be ready to listen in 10 or 20 years when you figure out that ball screws don't make a machine more accurate or make it so you do not have to compensate for back lash. If you have the cross slide so stiff that it won't put pressure on the feed screw then the gibs are too tight and the machine is going to generate errors all on its own. As to your force vectors leave the BS math at home and lean how to machine something. The pressure on the feed screw is governed by tool geometry and is only near zero in a configuration that would not be used in machining the part described. Yasmiin (1790)
No one wants to blow money unnecessarily, so I will try some cuts using the screw which currently has .014 backlash. I do not believe the cutting force will reliably push the tool out when the screw is reversed, but will be interested to see the results. Ensuring that the machine stays on the 'pushing' side of the backlash is what I had in mind when I talked about a spring on the cross slide. I had originally thought about a weight on a string, but a spring is much more compact and one 3 or 4 inches long should provide a nearly constant force over the .125" of travel I actually need. It's a lot easier to compensate for, or simply not worry about, the backlash of a double nut preloaded ballscrew, since it should be .001" or less. Chris (1791)
I think I understand the weight /spring idea and that will tend to hold it against the screw but you should be able to move the cross slide to the position you want it then push the slide back to take up the back lash and re zero the dial. I don' t know the geometry of the tool that you are using but it sounded like you had slow changing diameters along the length of the cut. This would imply a rounded nose on the cutting edge which will provide quite a bit of force back on the lead screw. If the feed is always out then this should work fine. if its in and out then you will have to compensate for the backlash each time you change direction. The spring idea is a cleaver one and should help if you don't get the results you want with just the cutting edge force. I was reacting to the ball screw idea as even a small one for a crossfeed is several hundred dollars and really doesn't make you machine more accurate. If you really want zero back lash for a reasonable price then casting a new nut with moglice is a lot cheaper solution and .014 is enough to consider making a new nut. http://www.moglice.com As to back lash of a ball screw -- one used in a machine tool is usually about .0002 the .001 is an industrial grade used for linear motion in actuators and the like. Yasmiin (1797)
More backlash
I chose the Whitworth thread because it had no sharp points or gully's to fracture in the possible use of a plastic product. The moglice idea is very good. If you leave some mail with these people they will call you. They are VERY helpful and no customer is to small. I have talked to them on a number of occasions concerning some projects I am currently working on. There is one other material that should be mentioned. Babbitt. Your basic lead/tin/antimony. Get a copy of the Lindsay publication that covers this craft. You don't have a Lindsay catalog!? SHAME on you! If there is a way to make a simple mold that the screw fits into, perhaps you could pour a new nut out of babbit? Back to delrin for a moment. I use this stuff for everything. Drill and tap it and you get a tight fit. If you know what size you need, perhaps you can get the right-size acme tap from MSC? 1-800-645-7270. Could you describe the part your making Chris? It sounds like a long taper shaft. Ron (1804)
I should call the moglice folks - keep thinking of more places I'd like to try it, including the non-rotating guide bushing that travels just in advance of the tool to support my workpiece. I've thought of babbit. I've also thought about trying the roughly 280 degree low melting, zero shrinkage tin-bismuth allow Cerrotrue, which I happen to have a bit of on hand. More expensive tools... my guess is the moglice would be cheaper than the acme tap, but I'm not sure. Sure, I'm trying to develop the ability to make a number of 2 to 4 foot long tapered mandrels over which I draw thin wall brass tubing to form French horn parts. Typical diameter varies from .250" to .472" on one section, the next going up to maybe 5/8" at a lower taper. The tapers are not constant over the length of the piece, but rather a carefully designed curve. I'd like to automate the process as much as possible so that I can experiment with the design of the tapers, rather than being limited to just one or two by the high cost of making these tools entirely by hand. I think I've decided to switch to 12L14 instead of drill rod at least for experimental mandrels. Perhaps when I get a design I really like I'll figure out a way to cut it in air hardening drill rod (or send it out to a Swiss-turn shop) for durability. Chris (1806)
Beating a "dead horse", I seriously doubt Moglice or Babbitt would have the shear-strength for threads w/o some deformation/cracking. I would vote for a ball-screw if you want low-friction for production CNC, or an adjustable acme nut if you just want limited backlash. I've got to make a new nut (and maybe a lead screw) for my 9", and I'm looking at various adjustable nut designs. Maybe I'll post some candidates to get some feedback from the group. My two cents, and hey you got it for free. Paul R. (1807)
Actually, if you look at www.moglice.com they have rebuilt threads in positioning screws nuts for some huge machines. It apparently is strong enough if you do it right - boring a sleeve to the right amount above the major diameter and cutting shallow crossing right and left hand ID thread grooves to improve adhesion. Babbit was used to replace half nuts during WWII in factories that had been bombed and had whatever the normal material was melt out in the resulting fire. If necessary, one could cut an oversize nut out of a high-strength material so that the antifriction material was simple a coating on an underlying threadform. Chris (1808)
Moglice is formulated for exactly this application and is even used in new machines for nuts. It will handle the stresses just fine. Yasmiin (1809)
Chris, They use to use a material for stamping dies during WWII. I don't recall its exact name. It started with a "B". They would melt it onto a part to make part of a stamping die out of it. These were low volume dies 10,000 to 50,000 parts. I know there is a manufacture of hotrod parts and chassis in Ohio that uses this material. I lost the article that referenced it. I was wondering if anyone had info on exactly what it would be? Tom (1811)
This backlash thread (pun intended) has been very interesting to me. There are lots of good ideas and great suggestions. I am in the middle of fixing my cross slide to reduce backlash, so this is very timely. Small Parts Inc. sells lead screws and anti-backlash nuts that appear from the catalog photo to be of the spring loaded double nut type mentioned earlier in this thread. I am not suggesting that anyone should go out and buy from them, but the photo makes it easy to see how the system works. They have an online catalog at www.smallparts.com. Micro-Mark sells low temperature alloys of various types for casting. They sell tools and supplies for the model railroaders (I mean HO not 1/7 scale live steamers) so you can get small amounts (11.4 oz.) for under $15.00. I made a new nut for my cross slide last night. This was the third try but the learning process is part of the reason for doing this in the first place. I made a simple threading tool to cut the threads. I turned down a 3" x 1/2" round bar to under a 1/4" on 2/3 of it's length. I drilled and tapped a 6-32 set screw in the small end of the bar and about 1/2" from the end I drilled a 3/32" hole through the diameter of the bar. A piece of 3/32" drill rod ground slightly under sized to the thread on my cross slide made the cutting tool. It helps to grind a flat on the side of the drill rod with a slight bevel against the set screw to help hold it in place. The 1/2" end of the bar mounts up in my boring bar holder. You can replace the cutter as often as needed to finish the job or regrind it to change the profile of the cut. I think a tool of this sort is called a broaching bar when you use it to make keyways or rifling. I found that my new nut fits very tightly on the end of the screw where there is less wear and fairly loosely in the center part of the screw where most of the action has been. I have some Delrin rod stock, so I may try the spring loaded anti-backlash double nut system. I'll post the results when I'm done. Glen (1812)
Backlash fixes etc.
A little thought brings us all here but its fun to talk about it anyway: 1: there is different wear and thus "clearance" at every point on the leadscrew, even from new, but just how much is it?? this varies based on the wear and corrosion, original surface quality, etc. YMMV, look at the screw and see whether there is any hope for using a device that presumes a fairly constant degree of clearance, you can use a wire gauging technique to compare the end threads and middle threads, to see if the thread grooves in the middle have appreciable wear ( measure the width - not the depth of the grooves - the real wear is on the sides) - ALL single nuts will have this lash characteristic on a non-perfect screw fit, even single ball-type nuts that don't use a recirculating ball track will only follow their groove with the major load on one side while moving. all that radial preload does is drive the ball or the nut thread into the groove harder that the motive force is pushing it out. To maintain a constant accuracy with minimal lash you have to have a tapered groove and a follower that bears on TWO opposing sides at once. No use of Moglice, or any other fixed nut design will change this. Notice that the follower device does not have to bear on two sides of the same groove!! see next para 2. if you can use the concept with two nuts as shown and explained in the SHOPTASK EL DORADO page http://www.shoptask.com/feature3.htm using double loaded nuts, you have a chance at getting the lash down very low, but you have to live with the other space and load restrictions. this design bears the preload on two opposing sides and as long as the preload friction is less than the power needed to move the carriage riding on the nut - now you have the nut tracking the center of the space between the two groove surfaces and are getting a fairly constant track with very low lash, and the wear is not primarily on one or the other side of the screw thread surface. you can use this principle to defeat some of the lash in the bronze nut using a homemade Delrin nut added under the slide ( does NOT replace the bronze nut) and a spring around a small machine screw in the same space ( or just a spring attached to the bronze nut) but it has limited load capacity - a big cut will override the spring force. 3. Most SB with some wear have a problem with the space between the leadscrew handle end and the handle nut - this is where the leadscrew on the compound and crosslide get lash quickly as this is where the screw is fixed in place at the end, try putting a brass or delrin shim here and preload just a teensy bit with handle nut tension to get the big stuff out, then go after the little pieces. its easy to saw a thin split in the bronze nut and put a small clamp on it to close down the threads a little (radial preload!) to see if you can get some of the manufactured or worn clearance out without replacing the leadscrew if you really find too much lash in the nut. Easy way to make a tap for delrin ( or even bronze if you are only using it once or twice) is starting with a piece of the screw material - fairly cheap to buy a little piece - just cut a taper on one end, few slots, relief for the cutting edge, and grind a square on the other end - making sure you cut or polish it to just a few thou undersize so the delrin nut will have some inherent tension holding it onto the (worn) threads. If the nut is too loose you can polish the tap down and try again. polishing the space between the teeth is where you will get the best result, not very much from cutting down the diameter. Even mild steel will harden just a little bit but you could case-harden or get a screw length of drill rod or 0-1 if you get really serious. Soapbox sagging and windbag empty now, keep the chips outta your drawers and always tie your apron in the back! (1810)
I'm sure someone will jump in if I'm wrong, but my understanding is that you cannot use a spring-loaded anti-backlash nut unless the spring-force exceeds the maximum working force on the screw. To use a spring, it would need to be stiff enough to cause significant friction on the screw. The spring loading would be rigid in one direction, and springy in the other. Paul R. (1813)
Glen wrote about his machined nut: Nice job on machining ID acme threads... not something I'm sure I'm ready to tackle. Anyway, I just spoke to Bruce at Moglice. 100 grams of the stuff (about 3 cubic inches) goes for about $35 bucks. He recommended using tight stretched Teflon pipe thread tape (!) as a release agent. The idea is that it generates more clearance especially at the major and minor diameters. I asked about using new rolled screw material. Apparently the problem is that rolled screws have nice tooth sides, but iffy crests and roots. We talked about reaming out the moglice nut to increase the minor diameter tolerance, and I was thinking maybe one could grind down the outside of the screw slightly once the nut had been cast. Have to be carefully of generating burrs though. Or maybe chase the thread root somehow upsetting the crest of a thread on a piece of annealed screw stock while dressing the sides a bit narrower in that region to be sure not to cut them. I think I'm going to try the spring idea first, although $35 for Moglice + $25 for good rolled alloy screw stock + $.50 for Teflon tape sounds reasonable, too. If I do this I think I will go to 1/2" screw stock rather than the present 3/8". And maybe sneak a thrust bearing in there, too. On a related topic, I think my bronze nut is slightly loose in the top slide. Have to figure out how I'm going to tighten that up and/or make the shell for a Moglice one stay tight. Chris (1814)
I think the spring loaded nuts are made of low friction materials, so it's okay. For the add-on external springs, I guess one would have to decide between setting up for OD work or for ID work. For OD work, the screw pushes in and the spring pulls out with enough force to overcome dovetail friction/stickiness. That should be less than cutting force. For ID work the screw would pull out and the spring push in. At least that's how it works in the machine shop in my head. Chris (1815)
This may be a dumb idea because I have never tried it but how about taking the old nut and slitting the bottom of it. Then put a cap screw through the flat part so you could adjust the diameter of the inside of the nut. Then spread clover compound on the parts with less wear and lap them - turn the lead screw inside the now adjustable nut. You could check your progress with bluing periodically till the lead screw has a uniform thread down its whole length. Then make a new nut with Moglice. Yes the lead screw would have a non standard major diameter but who cares. Using the existing nut as a lap should maintain the length wise accuracy of the lead screw. Yasmiin (1816)
I think that we are all eluding to the fact that SOME backlash is acceptable as long as its CONSISTENT for the length of screw travel. that way the accuracy is maintained. We are assuming that the screw is ok, and only the nut is worn out. not true. consider that your lead screw is worn more near the headstock than at bracket end. I think someone pointed out this example. we are starting to sound like the old iron group debating the value and karma consequences based the dilemmas of choosing paint colors, the proper plastic knob colors, or the 'proper' brass rivets to reattach nameplates when restoring machines. Woodworkers always seem to get their panties in a twitch over these things. granted overall machine accuracy is a bit more important than the vanity of paint color match, but I digress... can anyone definitely state what the thread is on the screw? I would assume that it is standard acme. Is it some weird pitch like and odd number? (I admit, I haven't gone out to measure mine, and I know my screw is worn so I wouldn't trust my screw as the standard) I think Scott pointed out converting to a ball screw is probably best for accuracy. He might be right, and I don't want to seem like I am second guessing his expertise. however, consider that ball screws don't have enough friction to resist turning of not held in place with some sort shaft lock or stepper motor power. This could be a issue when using the compound off of the apron drive. Also, I don't think that manufacturing a ball screw retrofit is as easy as it might seem as I think that the ball screws and maybe the nuts are hardened. I will eat my words later talking about machining 4130 further on... with fixes like Moglice and babbit wearing surfaces i sounds like its almost too much trouble. Not only in material cost, but the time and detours you would need to take to do the job completely. And what do you do if the fix doesn't hold? you are only taking out play in the nut. you are not fixing the wear in the screw. And you are using a potentially worn screw as the mold for both and making that defect the reference surface for the entire screw. The spring loaded pair of nuts is a good idea. The spring force is an issue. as an alternate, consider the way Bridgeport are set up to take out backlash in the table screws. As memory serves me, the nut is long like a barrel and is split almost 1/2 to 3/4 through the diameter in the middle of the barrel. A screw in a boss on one half spans across the split and bears against a boss on the other half. by extending the screw, putting pressure on the boss on the other side you are effectively making the once parallel slit into a slight trapezoid. this in effect causes the barrel to 'deflect'. ASCII graphics... \|\| goes to / this causes the threads to be a bit misaligned, but the effect is that you have cocked the two nuts apart a hair and they bear on opposite sides of the threads. The amount of movement is VERY small. backlash is reduced.... granted I AM and engineer and my nature is to try and 'fix it better than it was made. At the risk of sounding like a marketing campaign....from green bay: a 4130 or 1018 acme threaded rod 2G, lead error of less than 0.001 /inch in "standard" sizes (0.25-16 up to 0.75-6/8/10/12) is anywhere from $3-$7 per foot. A round ampco or SAE bronze acme nut roughly 1" long is around $20 for the matching sizes. So for around the same $35 in materials you have the raw materials for what I think is a better chance of fixing the problem. To fix the screw: turn down the ends for the bearings. Turn, shoulder and thread the shaft for the drive. maybe silver solder a bushing to increase the shaft diameter. Slot a keyway for the drive (which can be done w/o a mill. manufacturing a new screw is pretty straight forward. The nut could be more tricky. You could turn it down to the 'right' size. chucking it 'flat' in a four jaw you can turn it and machine off the flat. drill and tap an old style nut. this should fit the channel. if you wanted to make a split barrel like I was describing, you cold turn into the oversize nut to a barrel. Maybe silver solder a couple of other sleeves or bosses on. Maybe turn a couple of flanges onto the ends of the oversize nut and and grind away 3/4 of their circumference to leave yourself the bosses. You would then need to bore out a few of the threads in the middle of the barrel. Split the barrel in the middle and drill and tap the mounting screws and install the take up screw. I think it sounds easier than it might be since it would be machining really small parts, but not impossible. This would have to be made to fit into the existing channel. I would think that it would be as durable as the original and adjustable. dennis (1817)
First let me mention one thing I found on the 9" lathe last night. I was getting endplay in the crossfeed screw bearing no matter how tight I turned the nut in the hand crank. Turns out the micrometer collar was slightly too short and the 5/16" bored crank was bottoming out on the shoulder to the 3/8" bore of the collar. A simply brass shim washer (between collar and crank) fixed that. It depends on your goal. For a manual machine, this is true - some backlash is okay, while consistency is paramount. For a computerized machine the opposite is actually the case - backlash is very bad (for some types of cuts) but consistency is not paramount as it can be compensated in software. Of course it's hard to make a tight screw that isn't a consistent pitch. I am in favor of replacing the screw, as long as I can find new stock that is as good as the original. This is partly as I'd rather carefully pack up the original components in a safe place and run my experiments (fit altering, polymer casting, etc) on easily replaced generic screw stock. On the 9" south bend the screw is 10 tpi. I thought it was 3/8" diameter, but it may be 7/16". I think I might try to fit 1/2-10 as a replacement. (The 13" is 5/8-8). We're assuming it is acme - one reason for replacing both screw and nut. I have a 5/8-5 double nut left hand ballscrew (about $200 for the set) partially fit to my 13" lathe. I had to mill out the channel in the carriage a little bit, using a ball nose mill to keep everything nice and round to avoid creating stress risers. If anyone else attempts this, be smarter that I was and put round stock under the V ways to set it on the mill table... (I tried to shim it level off the flat surfaces) I'm making a new screw-in bearing sleeve so I just remove the entire screw, sleeve, collar and handwheel assembly intact (can't get it apart anyway) and replace it with the ballscrew based one set up for a cog belt pulley instead of a handwheel. There is one thing I think that would make using Moglice in conjunction with a new screw easier than using a new nut. This is machining the top of the nut to fit into the hole in the slide. For the block the ballscrew fits into on the big lathe, I finally put the old screw in a 5c block holder on the mill table and indicated the nut boss. Then I switched to an appropriate collet holding a piece of bar single pointed with the same threads 15/16-whatever threads as on the outside of the ballnut and screwed the block that would become the nut mount onto that. Programmed the mill to circular mill the boss - one might be able to do it on a manual machine with a boring head having a cutter turned in rather than out. Contrasting with Moglice, I think one could machine a sleeve that fit the apron with an oversize bore in about the right place for the screw. Assemble it with some clay or something to dam around the screw. Take the carriage off the lathe, set it upside down on the workbench and inject Moglice through the hole that I think I remember is present in the middle of the carriage. Springs work best with a low friction nut material... Moglice is supposed to be lower friction than bronze. I'd get 2C class screw from McMaster, but otherwise agree. Bruce at Moglice said stainless screws are often better made than the carbon steel ones. I would be the alloy steel ones are pretty good, too. I've turned ballscrews with carbide - acme screws are deeper but I would guess still do able. Chris (1823)
I have been following the threads on "Backlash" with great interest. The latest posting by Chris prodded me into responding. If there is play (or "backlash" if you want) in the ball crank - graduated collar - bushing area, there is a proper fix that does not require shims. I have included the SBL sheet on fitting a new cross screw with the instructions as an attachment. These instructions are just as pertinent to fixing a screw that has a worn thrust shoulder on the screw and/or bushing. If one lacks access to a second lathe, the repairs can be accomplished on the lathe under repair. For example if one is repairing the cross feed screw, one removes the cross feed screw assembly (ball crank, collar, bushing and screw) from the cross slide. Then one tightens the gib screws on the cross feed slide to lock it in place. Then one can swing the compound to "zero" or straight in and use the compound as the cross feed. One doesn't need power feeds for this repair. After the cross feed screw is repaired and replaced (and the gib screws are readjusted), the same procedure can be performed on the compound screw assembly. End play in the ball crank - collar - bushing assembly is common on these lathes if they have any time on them. It is often confused with "backlash" in the screw - nut fit and is easily fixed. Webb (1824)
Delrin is a DuPont trade name for a plastic that is strong, wear resistant, machineable, and very slippery. Check it out at this URL. Glen (1825)
Matt Is that all, mine has a half turn since I got it almost new in the late 60s. It is worse now but does not keep it from doing very precise work if you take the lash out before you start cutting. JWE (1827)
0.100 backlash
If I replace the crossfeed nut will this do away with my loads of backlash? I've looked back in the archives and see some guys trying to replace the feedscrew itself, is this where the problem lies? I thought the brass nut would take the wear? Right now I just make sure I'm always turning the feedscrew "into" the tension. Believe me though this novice puts more error into my projects than my old rickety lathe! But I'm having fun! Tim Q (3632)
A few thoughts: 1) If backlash varies depending on what part of the screw you are working, then that indicates the screw is worn. Typically, the screw gets worn in the middle. If the backlash is the same, independent of whether the cross slide is, the the screw is probably not worn. 2) 0.100" is a lot of backlash, a complete revolution if you have an indirect reading dial. There may be another source of backlash/play besides the screw and nut. For example, There might be axial play between the screw and collar. 3) The best way to determine the location and extent of wear is to disassemble your cross feed or compound. You will be able to see rounded, worn threads on the screw, and feel the looseness of a worn nut. 4) The method of disassembly depends on the model. If you have difficulty figuring out how to disassemble and reassemble your lathe, mention your lathe model, and what type of dials you have (the old, indirect reading, small dial vs. the new, direct reading, large dial). Someone on this board probably has a similar machine. Jon (3633)
Tim, I've had the same problem and found an e-z temp. fix and a more permanent fix. The temp fix is to simply cut a small strip from a piece of .001 brass shim stock that can be found at just about any hobby store. Place the strip through the crossfeed nut and replace the screw. If you nut is very worn you may have to insert more than one strip. The first time you run the screw through the nut it's a little stiff, but by the third or fourth time the screw is forming the brass shim stock to the nut threads nicely. Just use good common sense as to the amount of stock placed within the nut and you'll do fine. The permanent fix is almost as simple. I went down to the local industrial supply store with crossfeed screw in hand and bought an acme two stage tap of the proper size for around $60.00. A new nut was fashioned from a chunk of brass picked up at one of the scrap yards in town for about 10 or 15 bucks. ( I always get a little extra for the "third or fourth time is the charm rule") I have found that the old trick of rejuvenating a copper head gasket works real fine in taking up about 6 or 7 thou. of slop inside the nut if the screw is a little undersized in regards to the tap size. shod (3649)
Variable backlash
I have WWII era 9 in. model A with the large dial on the cross slide screw. My problem is the backlash in both directions behaves the same. At the start of the movement I get .003+" backlash, the slide moves about .005" and then skips another .010+" and then moves smoothly for any distance. Effectively I have about .018+" backlash by the time I can rely on the movement. This happens moving both toward the long axis and away from it. Jim (11539)
Sound like you could have slop in the feed-screw/nut interface as well as the saddle/feed-screw interface. You will always have some in both those places. I've got about 20 thou in the saddle/feed-screw and few thou in the feed-screw/nut. Hasn't really bothered me, but I plan on eventually fixing the major slop. Paul R. (11551)
Backlash adjustment
I got curious about the mention of making up leadscrews and nuts for low backlash so I went looking at commercial units. They are preloaded and some have 2 acetel nuts and a spring. Idea, take a second acme nut and fasten it to the first with shims in between to adjust the backlash. I should be simple enough to accomplish and fit in the allotted space of the cross slide. Its just a thought. JP (13972)
This method is used on the shoptask and some other machines of the 3-in-one and the multipurpose genre, as well as some others. its also used for some ballscrew applications. for many of us, the problem occurs when a leadscrew with variations in wear along its length is used, as we have in many of the legacy machines. this screw may easily have 20 thousandths wear over a 12 inch length near the headstock, and then the balance of it has less than 5 thou or so. if we take the same idea, but make one side of the bracket hold a spring or a spring washer with enough motion to take up the difference in wear but enough preloaded force to hold against the carriage pressure, we have the beginnings of a device that can closely track one side of the leadscrew groove. selection of which side is determined by the construction of the device and the bias force that the tool will need to resist. it may be that we do not actually need this much control unless going to CNC or some other type of motion control system, but it is nice to see that there is a mechanism to get there. it is certainly useful if one wants to reduce the lash in their system and make precision work a little more convenient. there is an alternative method achieved by taking a longer acme nut and splitting it lengthwise, then making a holder to preload the two halves in opposite directions whole holding them close against the screw, the result being that the nut is always held firmly against two surfaces of the screw groove by the clamp force. this also needs to be balanced against variations in screw wear but both methods have their proponents and features. BTW, the popular 7x10 minilathe uses the method of holding the nut ( in the compound nut adjustment scheme) in a slightly cocked position to make opposing ends bear on the screw groove, ( presumably on opposite sides of the groove?) and it seems to work OK for this purpose under limited force. it uses a brass or bronze nut running on a steel screw. I'm wondering when somebody will make ACME-thread form rollers and put a few of them in a nut assembly - potentially making a higher- precision retrofit without needing a full ballscrew. I did a dream- design on this a few years back but haven't the skills or time to pursue it. it's relatively simple and used spring biased axles to maintain the wheels fully engaged on the screw, not patented yet either. of course there are a few other secrets there in the design but that information is reserved for the joy of the maker - any takers? (13980)
A company in Reading CT makes these roller nuts on a large scale to move elevators. Square thread 4" and up. I saw it 25 or so years ago. JP (13981)
If you're thinking of the ones I'm thinking you're thinking of - these are called roller screws ( and nuts). they are heavy-duty variations of a ballscrew and boy do they do some fun things. compound rates, variable rates, planetary sets with opposed threads on the inside and outside - a lot of windy stuff. what I'm thinking of is a plain-Jane variant that as a cartridge or a basic building block - replaces a solid or split acme nut on an existing acme shaft. AKA - a retrofit in the same pricing category as a replacement part, just giving an upgraded motion control capacity. (13986)
Backlash Fixes
Webb I am interested in your reply awhile back on the end play at the crossfeed dial and bushing, you mentioned a repair from south bend and attached a sheet from them , but the jpeg and instructions were not stored, where can I find this or get this. your help needed. Fred (14946)
Fred, I've loaded the instructions onto the "SouthBendLathePix" sister site, in the files section. http://groups.yahoo.com/group/SouthBendLathePix/files/Cross%20Feed%20Screw%20Installation%20/ Webb (14950)
Webb, is there a possibility that the manual you have taken these from could be scanned into the files in total? And do you have one for the Nine as well? I for one would be happy to pay for a copy if that is the only way to go. Bill (14964)
Backlash
I acquired a Heavy 10 about 6 months ago, and have devoted myself to reading all I can about lathe work, including a religious reading of this site daily. It has been a great help. One topic I never hear discussed is backlash, although there is an occasional reference to adjustments to remove some of it. Would someone kindly provide a brief description as to the source/cause of backlash and how it can be adjusted/corrected/reduced? My crosslide has about .008 backlash (and an older small mill I recently acquired has about .030 in the table) Seems to me that backlash should be simply wear in the nut, which can only be fixed by replacing the nut. Dave (21008)
Wear in the nut and or lead screw causes back lash. Compensating for it can be done by always making your table/cross slide lead screw turn in the same direction for the set up of each cut. For example: If turning a shaft would would move the cross slide in an amount. (say .002"). As the cutter gets to the end of the cut you would want to back the cross slide out say one whole turn. Then you would reset the cross slide in one whole turn, plus the original .002" plus the next amount of cut(say.002" again) Thus by making the last adjustment in the same direction each time you compensate for the back lash. Naturally having a dial indicator on the crosslide is very helpful as a second means of cross checking. I'm sure others will want to add their methods as there is considerable experience on this site. Eric (21011)
The reduction portion I can't answer, but the cause is likely not the nut in either case. It is usually the screw itself moving against the mounting on either end. Gene (21031)
Is backlash the play in the for instance, compound dial handle between the two spots on the dial where there is resistance? In other words, when I grab the handle and turn it one way or the other it turns pretty freely in either direction until it snugs up. Is this what you call the backlash? Now, if so, how do I tighten it or do I really care? I mean, as long as I snug up the handle to the work before I begin a cut and zero the dial isn't everything fine at that point? Brian (22496)
100% Correct! That's what I was told g anyway I went ahead and tightened things up but I can see why it isn't a required tune-up procedure. Gene (22497)
Brian Backlash between the screw and the nut is the difference between driving forwards and pulling back. There will always be some, otherwise you would have difficulty in turning the handle. Play in the gibs (ways) can appear to be backlash as can other factors. The result of backlash can be minimized by bringing the tool to the work from an apparent clearance point greater than the degree of backlash. in other words in order to begin a cut of one scale division, back off say five divisions before advancing six divisions. Many other factors can be at play such as if you don't secure the non driven motions: vibration and load can force the tool away from the workpiece and result in a tapered cut. Usually this is less important for roughing and more important for finishing. Jim (22499)
It could be a concern if you are doing something like climb milling. Jim B. (22500)
Climb milling on a lathe? (22501)
I have. Bob (22502)
I would not mill on a lathe period. But there are people who do. Jim B. (22503)
If I had a mill I would use one, but my lathe works just fine as a mill. Bob (22504)
Brian, I think the most important thing to remember when it comes to backlash is always work in the same direction. Example, you take a cut on your work then back out the tool far enough to remove all backlash, bring carriage right, then turn in tool (using graduated dial to desired depth) and take your next cut. The same goes for boring but in reverse. If you were to take a stepped cut on your work and you moved the tool into the smaller diameter then used the graduated dial to move out to the larger diameter your second step would be smaller in diameter by an amount twice the backlash, to avoid this you would bring the tool out beyond the required diameter and backlash then bring the tool back in to desired/corresponding dial reading for that diameter and there should be no error. I am not a SB owner as of yet so I couldn't tell you what an acceptable range for backlash would be or if this can be adjusted but I would say .003-.005" hopefully someone else has this info. Russ(22520)
As long as you back off two full turns your dial readings will remain accurate when you come back in, even on a badly worn machine. Roy (22533)
The main problem with significant backlash occurs when parting off rather than normal turning. Its also a right PIA if you try milling in the lathe, with a vertical slide for example, but that's not a popular exercise these days; and can give problems if you have something mounted in the saddle for boring with a between centers bar. With any screw feed system the load is only supported the way the screw is pushing. Any clearance means that things can float slightly if the load reverses until the clearance is taken up so the movement hits a positive stop. When things are floating the tool is clearly out of positive control. No great worry if its only a thou or five but large backlash can cause serious problems especially with parting off when, as we all know, the tool can try to draw in under load. An extra self feed of several 10s of thous when the parting tool hits a hard bit or the chips start to jam up is usually bad news. For normal turning you almost invariably arrange things so the tool is pushing back against the feed so the only problem with working with large backlash is remembering how much there is. Especially as large backlash means well worn lathe so the backlash tends to vary depending on where in the feed you are. If you are not in a position to do anything about serious backlash yet its well worth fitting an extra gib screw with a thumb push handle to top and cross slides. This makes it easy to lock the unused slide and snug up the gib on the used slide during critical operations without disturbing the main adjustment. 2 BA is a nice thread, gives about 1/3 rd turn from full loose to lock-up with a reasonably adjusted gib. Well worth doing as a matter of course when overhauling any small lathe. Such locks once enabled me to do a fair job of screw cutting on a lathe with something like 80 thou of backlash in both top and cross slides by locking the slides for every cut. Not fun but it helped solve the problem of using a lathe to make its own repair parts. Rear tool post and inverted mounting for the parting off tool helps cope with backlash in the most vulnerable operation for two major reasons. Firstly, with a rear mounted tool, the feed screw is in tension so everything is pulled firmly straight making the assembly a lot more positive. In the usual arrangement the screw is in compression so it will try to buckle out of the way either by bending (the core diameter is very slender) or wriggling away into the other wear clearances that have accumulated (lots of backlash = well worn lathe). The screw is a helix so the wriggling has a sort of rotating movement too. Effectively you are putting the cut on through a spring, a strong spring but still a spring. Given half a chance the thing will start oscillating in the backlash region if the cutting load changes. We call the effect chatter, usually with associated shop language, especially when parting off. Secondly the geometry of the inverted tool means that its more likely to be pushed out of cut rather than drawn in. Because everything is in tension and all held firmly back against positive stops any load changes have to stretch the feed screw a bit more or a bit less to change the feed applied. Takes a fair old load change to stretch or un-stretch(!?) a feed screw by a thou so the rear mounted system is effectively invulnerable to load induced changes in cut. Clive (22540)
Clive, The backlash is the play in the feed handle where I turn the handle back and forth where there is no resistance, i.e., nothing but the handle is moving, right? Brian (22547)
Backlash and movement of carriage
I have a heavy 10 that has a ton of backlash in the cross slide and I also notice the carriage will move up and down. There is some wear on the ways . Is there a way to minimize this without having the bed reground? Frank (22867)
Hope you get a reply Frank. I have some of the same issues. Brian (22923)
Sounds like your nut and screw have had the schnitzel. Replace or renew. The lift I would guess is at the back of the machine. (You didn't state where.) There should be a flat (gib) bar that holds the saddle down in the back. Could be loose or missing. If the lift is in the front you really have a lot of wear on the ways and or saddle. There should be some slack but not in excess. Bruce (22953)
Mine had about .025 +/- I made a new screw and nut. I'm down to about .005. I use it pretty steady and have to work to tenths many times. Can't afford digs etc. No room to play games. I guess it depends on what you are planning to machine on your lathe. Lift should be as close to zero as you can get without binding anything. One of these days I plan to put a needle thrust bearing in and hope to eliminate some of the .005. Unfortunately there will probably always be some play. Bruce (22975)
Bill. Guess I'll need to go home and check to see if the dial is direct or otherwise on my lathe. Brian (22978)
Most of the play is indeed in the back of the saddle. There are bolts on the underside which I assume are holding the gib in place. If I remove them is there any criteria to check the gib. Next how tight should the bolts be. Sufficient to reduce play without having binding of the longitudinal feed? Also could use some how to on disassembling the crossfeed to check the nut and rod for play. If I back off the feed three or four turns I have to physically pull back the crossfeed a good 1/4 ". I know something is not right. Just haven't had time or guidance to tune this thing up. Retiring in May. Perhaps then. I have south bend parts book but it does not go into detail for this type of maintenance (23026)
Frank, What size lathe do you have? (23032)
Frank, I have to turn my crossfeed handle about halfway around one time before I get friction and the cross slide begins to move. Other than that I can't feel anything loose. Is that what you have or is there something else loose? Brian (23057)
Yes there is at least half a turn before the cross slide begins to move. After that if I bring the cross slide back out to clear the work I have to physically pull back the slide a good 1/4 ". I'd like to disassemble the thing but would like a little written guidance on just how to do that. frank (23072)
I have a heavy 10, single handle gearbox that came out of a school. It was built in 1953. Re-shimmed the headstock and replaced a few oilers and had to weld the broken shift handle but it works. A little noisy and not accurate for real fine work ( It bores slightly tapered holes, etc) I think because the ways are worn in that 8" area near the headstock Price was right $350 with two chucks and set of 8 collets and collet wrench. (23073)
Best guess. Your cross-feed screw looks more like a sharp "V" than an acme thread. William (23076)
Have to agree with William on that one. My cross feed screw on the saddle has about .030 play in the center and the acme threads look a lot sharper in the center of the screw then on the ends. I called Rose whew a replacement screw was almost $600.oo for the saddle and $500.oo for the compound. Anyone out there want to make a few bucks and makes us a couple of screws ? Ed (23078)
Backlash and accrue
Two of my favorite words. Both are present in just about any system be it natural or man-made. We go to some extreme efforts to eliminate them, only to find more. Any machine will accrue backlash of one sort or another. If you spend your time trying to eliminate 'lash, that's less time making that steam engine part you need. You must decide how much of it you can deal with and eliminate the rest. Regular PM checks and adjustment kept most shops humming during the war. Or why it is that some old machines that were discovered under a layer of cosmoline are in such good shape. We can still do this today provided we have the skill to do it. I build backlash and gradual inaccuracy into my model of production. This way, it becomes part of the solution, not part of the problem. I can remember a little cartoon in "American Machinist" showing 2 guys watching a 3rd worker with the caption of, "He has a closer tolerance then thou attitude". Hmmm, another funny word tolerance, "what your willing to put up with". Ron (23436)
Great, sensible words Ron, and a comfort to many of us. I have a mate who is a brilliant machinist, 70 years old and Royal Navy trained, who has just built most of a Stanley Steamer around a few original parts and chassis. He has a beat up old Harrison with slack everywhere, and yet the parts he turns out are superb. His philosophy, when I remarked this, was, in essence exactly the same as yours, "if you know it's there, you build that into your calculations". I coachlined this vehicle for him, and as payment, had my first SB, a 9" C that was stuck up at the back of his workshop covered in grease with years of dust attached. Len (23437)
I learned most of my 'machinist' skills from an 'old timer' who was in fact younger than myself (sadly taken from this world by cancer a few years ago). I watched him machine a outboard motor part (for fine tuning gearcase clearances on my 60 horse Johnson) on my brand new (to me) 75 year old South Bend 13 inch. I hadn't even got around to installing the proper drive to it. With an angle grinder and a piece of shafting we made a drive using a variable speed half inch drill - hand-held! He got that part re-sized within two tenths (not to mention the hand ground HSS tool bit that appeared like magic in his hands off a cheap and essentially garbage bench grinder and stuffed into a lantern type tool holder!) My lathe had/has slop everywhere! Last year I had to fabricate three MT2 tapers for my RT on my mill- drill. Using his words of wisdom and a practical approach I was able to achieve a perfect fit on these tapers first try (all three!). To be sure I had the compound gibs tightened down to the point of making the turning of the feed crank painful but you do what it takes. I just wish I could show them to him. Worrying about backlash can be counter productive. Learning how to work with backlash and slop is where it is at because those conditions can, at best, be minimized but never eliminated. I could have spent a month and mega bucks to minimize that overall slop and still not made a better job. If your equipment has state-of-the-art ball screws and all that you wouldn't be here in this group reading this. My shop is an old wooden structure built on pilings over the tidal beach in front of my house. Leveling is just not an option! The whole damn lathe wobbles back and forth as I walk about the place. Every job, depending on needed precision, is individually adjusted for. Sometimes I blow myself away when I hit the mark. Whew - didn't set out to write a book. Bottom line - you can achieve remarkable accuracy by simply thinking through the problem and applying simple principals. It isn't the machine that makes the part - it is you. Hope this helps someone. Ken. (23467)
Ken As you say a lot of people waste bunches of time and money chasing inconsequential things. The one that made me laugh the hardest and longest was the reinforced concrete bench. JWE (23474)

Index Home Page