No announcement yet.

How to improve turning accuracy?

  • Filter
  • Time
  • Show
Clear All
new posts

  • How to improve turning accuracy?

    Most of the turning that I do is general purpose stuff (say +/- 4 thou) and has no need for great precision. However, lately I've been trying to turn shafts and bore holes to within a thou and am finding that a challenge. Typically I turn down to 5 or so thou over, measure with a micrometer, adjust and ... overshoot.

    The cross slide has some backlash but as I'm not changing direction I wouldn't have thought that would have a great influence. The dial is graduated to 1 thou increments on diameter, so I'm thinking that I shouldn't need to be putting the compound on odd angles, and the lathe is a big solid thing (12x30), so flex while possible shouldn't be a major consideration. This happens with carbide as well as (freshly sharpened) HSS.

    From those who regularly do this sort of thing, what are the rules/ tricks for this operation. Is it just practice? Should I be locking the cross slide once adjusted even if it is a light cut? Taking more than one spring cut before measuring? I could rough down closer to the desired dimension but suspect the radius on the cutter tip will start to come into play.


  • #2
    Always make sure backlash is out, and for a one-off part (as opposed to high-production) take TWO final cuts of the same depth-of-cut. Try to have your final cuts at least equal-to or greater-than any radius on the tip of the tool.

    By taking two final cuts you can make the first cut, measure deviation from target size, then adjust the infeed for the second cut accordingly.


    • #3
      You need to practice a bit more, and play with parameters: DOC and feed rate.

      If you're turning something down, taking a .030" cut, measure it after every pass to see if it's actually being reduced by .060". When you change to a .005" DOC, the part will almost certainly be reduced by more than .010", due to the spring in your setup. Lighter DOC -> less spring -> more DOC than planned.

      When you want to hit a given dimension, the best thing is to back off on your DOC well before the final pass. IE: go from a .030" cut to a .005", when still maybe .020" from final dimension.

      If you change your feed rate, say to get a smoother final finish, you'll also tend to overshoot and make your part undersize. Same reason as above.

      Again, experiment and you'll be able to do much better. That said, I'll sometimes take the last .001" off with a file. It doesn't take long, and you can hit the dimension right on the nose. Plus, if your lathe is cutting a slight taper, you can usually correct it at that time.


      • #4
        I'm not in a position to give advice, but just wanted to share my experience while turning a shaft to be fitted into a bearing. I did a first pass, measured the diameter and then did a second pass. The result was 20.00mm, as measured with a micrometer. But when I tried the shaft after a half hour, it was loose. I didn't take into account the heat expansion, which made the part appear thicker than it really is.


        • #5
          It's always worth remembering that when you measure the diameter, you measure the tops of the peaks of the rough turned surface.

          If you're after .500 say, amd you measure .503 after a rough cut, the troughs are probably already below .499.

          If I want to know how I'm going, I'll take a finish cut at .520 to see what it really is. Then I'll rely on the dial to get to within a gnat's, even going for another rough cut and a finish cut afterwards.
          Richard - SW London, UK, EU.


          • #6
            Yes,spring back and heat can hurt. Depending upon several factors,like how long is the work,is it supported at the tailstock, how sharp is the cutting tool,etc. can cause the metal to spring back. Generally speaking,if you take another pass at the same setting,you will see the cutting tool removing a bit more metal. On a close fit,heat will swell the diameter enough to mess up a fit when it cools.


            • #7
              Originally posted by form_change
              The cross slide has some backlash but as I'm not changing direction I wouldn't have thought that would have a great influence. The dial is graduated to 1 thou increments on diameter, so I'm thinking that I shouldn't need to be putting the compound on odd angles, and the lathe is a big solid thing (12x30), so flex while possible shouldn't be a major consideration. This happens with carbide as well as (freshly sharpened) HSS.
              You might be getting a little "sticktion" on your cross slide that shows up as non uniform advances as you wind it in in small increments. Setting the compound over to do the last few thousands greatly helps over come this. FYI, the angle that seems most preferred is 5.7391704 degrees. This gives a ratio of 10 to 1 which makes the distance conversion a simple matter of moving the decimal point one over, nice and simple and darn near fool proof.
              Location: Saskatoon, Saskatchewan, Canada


              • #8
                Dont sneak up on a cut. Almost guaranteed to cut undersized.


                • #9
                  I would

                  1. take all cuts the same depth and feedrate...then load induced deflections are similar for each hit.

                  2. Aim for plus 2 thou.

                  3. Sort out the surface finish and size with a smooth diamond hone and/or smooth emery roll.
                  " you not think you have enough machines?"


                  • #10
                    Thermal is a good point, Cutting heats up work, your final cuts should be done after it returns to room tempature, especialy if you have hoged off a lot (really heats the work)

                    Spring back is another good point, in the work (Causes taper) and the lathe itself (Causes overshoot)

                    What i'll usally do is take 2, 3, sometimes even 4 passes at the exact same depth setting, untill the tool isent even shaving off dust.

                    Then measure it, adjust the dials for the next pass, finish diamiter + 0.004" or so.
                    Do that pass 2 or 3 times, measure again, and do another 2 or 3 passes at the final setting of finish diamiter (Sometimes a half thou off the +0.004 it should be from the last pass)

                    Watch out for taper when measuring it. If its work spring, it will be smallest at the chuck (And tailstock support if used and properly aligned)

                    Taking the final 0.001" or so off with sandpaper/files is supriseingly accurate and easy. Just measure along and where its big you mark with a felt pen, then just file the marks off, repeat. the trick is getting a decent lathe file, idealy with safe edges, and puting it down on the work flat without the edges touching first and leaving scratchs. (I am tempted to grind the edges of my lathe file to make it smoother on the corners, but then I can't make square shoulders with it.. Need another lathe file I guess)

                    And yea, it will *remove* taper if care is taken. thats right, some poor sob with a micrometer and a file can be MORE accurate then your 1000lb+ $3000+ lathe, easily.

                    Just realise the poor sob with a file sure won't beat you on material removal rates! its all about using the right tool for the job/task at hand! And yes doubters (I used to be one!) a file is often the right tool!
                    Last edited by Black_Moons; 07-20-2011, 07:10 PM.
                    Play Brutal Nature, Black Moons free to play highly realistic voxel sandbox game.


                    • #11
                      Originally posted by macona
                      Dont sneak up on a cut. Almost guaranteed to cut undersized.
                      I'm certainly not qualified to mentor on this issue, but that's been my experience too Macona.

                      I try to never find myself needing to take off less than 10-20 thou on the last pass. And as others have suggested, jot down the deviations from "dialed" cut increments for those last several preceding cuts, so I can sort of project how that final one is going to turn out.
                      Lynn (Huntsville, AL)


                      • #12
                        I am going to assume it's a worn lathe since you state there's slop in the cross slide.

                        When using a lathe with worn cross feed screw/nut you have to use a dial indicator on the cross feed. I use two mag bases. One with a dial indicator on it and another with a flat bar for the plunger to run against. That is the only way you will know how much your feeding into the work. You can't trust the dials on a worn or sloppy lathe.

                        Your just wasting time trying to use the dials on a worn lathe or any other machine that uses feed screws.

                        Sometimes you have to screw the compound all the way forward to remove any slop that can't be adjusted out.

                        I spent most my career in shops with worn out machines and had to learn how to make very accurate work using worn out machines and it wasn't easy.

                        I'll go set up what I used to do and post a photo.

                        Here's a photo of the setup I used. Of course I put the indicator on the right side when I was using it. The reason I put it on the left side here is because I didn't want to move the compound for the demonstration. You can use almost any mag base and indicator as long as you can put the plunger on a solid spot on the compound. The combinations are endless for sure.

                        Here's a photo of my collection of Mighty mag bases that I used a lot and sometimes needed several on a machine to do the job.

                        Last edited by Carld; 07-20-2011, 08:40 PM.
                        It's only ink and paper


                        • #13
                          There could be several things happening that could cause that, are the gibs on your crossfeed and crosslide properly adjusted? Better too be a hair tight than too loose. The depth of cut on your finish pass should be influenced by the alloy your working, some finish well to light passes of a couple thou and others do better with a cut of 10 thou or more. The setup of your cutting tool can have an impact too. Especially if your tool is below center. And experence and practice is always a plus. When I first started, my mentor had me turn stepped rounds of varying diameters and step heights until I could get them to within +/- .0005 of the spec set for them. It was great for experence in turning to a shoulder and cutting to a dimension and can be used to get familar with how different materials behave.
                          Typically, for straight turning, I use my crosslide for setting the depth of cut for my roughing cuts and only use my compound for the finish pass-es. If I'm not sure how the material will finish I'll take some practice finish passes with the compound while I'm while I'm still roughing, that let's me play with the speeds and feeds and check how my depth of cut compares to what's actually happening. That way I know how to achieve the finish I need before I'm ready for the last couple of passes. My finishing speed is almost always faster than my roughing cuts. If I've been taking heavy roughing cuts on a piece, as I start getting close to my dimension I make a couple passes with lighter cuts before my finishing pass.
                          I always leave my compound angle set at 30 deg. for threading unless I have to change it to cut a specific angle and there's no other way to easily make that cut. If I'm not threading my compound is always set be at the end of it's travel towards my work minus 1 full turn. That leaves me enough travel for finishing and gives me the most rigid cutting platform. I have the gib on my compound set tight enough so that it takes more effort to turn than my crosslide and has a slightly stiff feel to a turn of the handle. I set my tool post up so that my tool holders (qctp) are parallel to the face of my chuck for turning and parallel to my work piece when boring. I set my tools into the holders at a slight angle - about 3 deg. - to give me a little extra clearance when coming up to a shoulder. Remember that with your compound set at 30 deg. a 10 thou indicated turn of the of the handle will be a cut of less than ten thou. For precision I'll setup a dial indicator to show me the exact depth of cut on the compound.
                          I have a 10x jewelers loupe that I use to examine the cutting edges of my HSS and carbide when I'm sharpening them. It will let you see defects in the edges that you might miss otherwise. Especially if your freshening up a tool you've already used or your like me and your eyes arn't what they used to be.
                          This is just my opinion and I'm sure others here their own way that's just as good or better, so I'll be following this thread along with you to see what I can learn.
                          I cut it twice, and it's still too short!


                          • #14
                            The problem with a worn lathe is the feed screw and nut on the cross feed and compound are not worn evenly the length of the screw. That makes you get different DOC with the dials.

                            You really can't trust the dials on worn lathes. You have to use a dial indicator to know exactly how much you moved the cross feed or compound.

                            On some lathes I had to use a dial indicator on the cross feed and the compound when cutting a thread because the dials would lie to me.
                            It's only ink and paper


                            • #15
                              Just do like the Pix man say's. Keep it simple. john b.
                              John b. SW Chicago burbs.