Announcement

Collapse
No announcement yet.

3 wire thread measuring

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #31
    Keep it simple

    While the esoteric and "super/hi-tech" stuff is nice to know it is pretty well "knowledge for knowledge's sake" in many if not most instances.

    Further, if the references are not available to the HSM or small shop owner when they are needed, they are of no practical use at the time - if the machinist has the time or interest to read them.

    I am concentrating on the small and HSM shops where all that is available is what is to hand at time regarding tools, material and reference material. The bigger ("so called") "better" shops are "big enough between the tits" to have the resources to look after them selves. Many will respond just as the smaller shops will though.

    In other words, I am focused on good or normal or reasonable shop practice.

    Specially made or "mated/matched" super-high precision (and cost to suit) screw-nut pairs are either not available or relevant to many HSM or small jobbing shops - even if they were affordable.

    These sort/s of high-precision nut/screws would be even less applicable to lead-screws and "half-nuts" - without even considering the "end float" that will result from clearances in the half-nut actuating mechanisms.

    So it is not "OT" at all.

    Comment


    • #32
      Originally posted by biometrics
      I think I just found "the" answer to 3-wire measuring of threads on Tony's Lathe website. The photo tells it all:

      The photo does not tell it all. There is NO WAY that the three segments can assume the proper conditions for an accurate measurement. If the wire is bent to the exact dimensions of a correct thread, then an incorrect thread would have them at angles to each other, not parallel as the three wire method assumes and requires. Likewise, if the wire is bent even slightly wrong for the correct thread, then the three segments will never be in the correct position when the thread is correct.

      I would never use this. For an accurate thread measurement.
      Paul A.
      SE Texas

      And if you look REAL close at an analog signal,
      You will find that it has discrete steps.

      Comment


      • #33
        Originally posted by Paul Alciatore
        The photo does not tell it all. There is NO WAY that the three segments can assume the proper conditions for an accurate measurement.

        I would never use this. For an accurate thread measurement.
        Besides......

        Is that paper clip the "best diameter" for your thread pitch?



        For ID thread measurement, your best bet is to make a gage, using the 3 wire measurement to check it...... or buy a gage, if the size and tolerance is a recurring issue.

        A proper gage will have a "go" thread which must screw in, and a "no-go" thread which must NOT. Those are determined by the tolerances.

        If the measurement is that important, you can afford the time or money for the gage, and can lap it to size/finish.
        CNC machines only go through the motions

        Comment


        • #34
          Originally posted by Paul Alciatore
          The photo does not tell it all. There is NO WAY that the three segments can assume the proper conditions for an accurate measurement.
          That was a joke Paul -- that's a paper clip
          "Twenty years from now you will be more disappointed by the things that you didn't do than by the ones you did."

          Comment


          • #35
            Originally posted by oldtiffie
            I am concentrating on the small and HSM shops where all that is available is what is to hand at time regarding tools, material and reference material.
            Sigh. You said you didn't know of any way that even Job Shops could measure internal threads. Job Shops use Repro-Rubber, which is $100 for a complete kit from Enco, MSC, or J&L. Or if they're a machinist, they'd use Cerro-Safe, which they probably already have, which is $35/lb.

            Originally posted by Tiffie
            Specially made or "mated/matched" super-high precision (and cost to suit) screw-nut pairs are either not available or relevant to many HSM or small jobbing shops - even if they were affordable.
            My point was that even Moore's ultra-precision leadscrews use the same method that a lot of Home Shop guys do to get a low-backlash fit: cut the leadscrew, and make a nut to fit (or vice versa).
            "Twenty years from now you will be more disappointed by the things that you didn't do than by the ones you did."

            Comment


            • #36
              "Going good"

              I have no problem with Repro-Rubber or Cerro-safe as a viable method as they are well-proven.

              They will be of no value if not at hand when needed within the time-frame for the job. I doubt that many HSM-ers or small shops would have it on hand "just in case". But if they do and they use it, it will be as good as you say.

              "Moores" screw-sets are, as you say, lapped which infers lapping compound. If I were to even try to emulate that method and I was using bronze nuts, I'd be less than certain that I could get rid of ALL of the lapping media to ensure that it did not keep lapping under use. Bronze is too easily embedded with grinding/lapping compounds and too difficult to remove. While a good microscope - if I had one - might pick up the grit in the steel screw (where it is visible) - I doubt that I could see let alone remove any lapping compounds in the internal threads in the nut. This is of particular concern given that most lapping jobs work from a relatively "coarse" "grit" to a very "fine" one and that "grit" tends to bind with/to the bronze and is difficult at best to remove.

              Never the less, the end result is, as you say, a matter of the Machinist making the nut-screw "fit" fit to suit himself - often by "feel" and "experience" with no "measurement/s" at all.

              The main problem with many lathes - and mills (other than the "X" screw) is that there is no provision for eliminating "end play" at both ends of the lead-screw as is the case say on the "X" screw on a CNC-ed mill table with a ball-screw.

              As a point of interest. Put a magnetic or other base on your lathe bed and mount a good dial indicator on it with the "tip" on your tool post. Engage the half-nuts fully.

              Method 1:
              Move the carriage left with the hand-wheel and re-set the dial indicator to zero. Now move the carriage right until it is stopped. Record the dial reading. The difference between the readings is the total effective end-play and back-lash in the lead-screw in its mountings (end play) and lead-screw back-lash plus similar quantities in the half-nuts and their clearances in their slides.

              Method 2:
              With the carriage locked/clamped to the bed and with the indicator still on the tool post, physically pull/push the top-slide (side-ways) along the lathe bed axis. This is the side-play/"slack" in the cross-slide as well as any end-play in the top/compound slide lead-screw etc. and is in addition to method 1 in terms of "play" etc. at the tool-post.

              Method 3:
              Repeat methods 1 and 2 but in this case instead of physically clamping the saddle to the bed, move the saddle/carriage by rotating the lathe spindle. Record the positions on the chuck which correspond with the "start" of dial indicator movement in each direction. This is the total effective back-lash etc. between the job and the cutting tool at the position of cut - and includes the amounts contributed by all gearing - back-gears (if fitted) geared head (if fitted), gear-train and quick-change gear-box (if fitted).

              So all in all, their is no real case for total elimination of all these "errors" in a screw-cutting environment on most lathes - except perhaps for CNC lathes where the relativity between the spindle and the lead-screw are electronically controlled and the lead-screw is a precision ball-screw with all back-lash and end-play eliminated.

              This picture is not a disaster at all. It merely points out the despite all these potential "errors" and "problems" that in the right hands just about any lathe can make pretty good screw-threads. It can also make a pretty good job of any other turning required made of it as well. Same applies to mills with or without super-precision lead-screw/nut combination and elimination of end-play or DRO's etc.

              This is shown every day of the week on this forum and in most HSM and other shops.

              This is not meant to say that better is not better - it is. But it may not always be essential and a skilled machinist can "work-around" many such perceived "problems".

              I hope that this is a "boost" to some who might not realise just how well they are doing in their shops with what they have.
              Last edited by oldtiffie; 06-17-2008, 08:08 PM.

              Comment


              • #37
                Originally posted by oldtiffie
                I have no problem with Repro-Rubber or Cerro-safe as a viable method as they are well-proven.

                They will be of no value if not at hand when needed within the time-frame for the job. I doubt that many HSM-ers or small shops would have it on hand "just in case". But if they do and they use it, it will be as good as you say.
                I have about 3 lb of the stuff, and it works pretty well. I don't think I would trust it to tenths, but others might. Main issue is it's soft, and you SHOULD probably use a comparator, not a contact measurement means

                Gages work.

                But, anyone who believes that they can CUT threads on a lathe and work to sub tenths accuracy may need to re-think it..... that's grinding territory, for a reason. Your worn-out lathe *isn't quite* as accurate as it once was, and that was probably not THAT accurate as to the leadscrew.

                Then the tool tears its way through the material, leaving boulders and globs (on a microscopinc scale) that affect the thread sizing and measurement. You need to grind to clean it up, unless you have a 10EE, which has a chance at making a pretty darn good thread if the tool etc is right.

                As far as leadscrew end-play, best to get rid of it, which isn't that hard if you sieze ONE end of the leadscrew with oppositely preloaded thrust bearings. There are still lots of "camming" and other errors that crop up, aside from the basic leadscrew accuracy errors, and those affect accurate work. Sloppy half-nuts etc will not be affected.

                OTOH, if the leadscrew is always loaded in one direction, in tension against the cut, backlash in the OTHER direction won't matter unless you cut in such a way as to allow the tool to "take charge" and push the carriage ahead..... for one, doing a straight plunge-cut with crosslide, instead of advancing the compound at 29.5 (or the like).

                So, as you likely already know, much of your example is not as important as you make it seem. That backlash is only important if you cut moving both ways, in which case your thread will be sliced off and become just a bumpy spot on the work.

                Where it CAN be quite important is if it does not "repeat", if the backlash in the forward direction varies with carriage position, leadscrew rotational position, etc, etc. That causes a problem, particularly severe with finer threads.
                CNC machines only go through the motions

                Comment


                • #38
                  None needed.

                  Good reply JT - very good indeed - and appreciated.

                  Now I am going to perhaps really throw gas on the fire - with some anyway - and show that three wires of a prescribed - or even same - size are not required at all for a very satisfactory measurement of a thread (external) to acceptable limits to a standard.

                  Not two either thought that will do the job too.

                  Same for just one wire.

                  Or perhaps none at all!!!

                  All that may be required is a serviceable "1 thou" micrometer and a reasonable-to-fair lathe. ("Tenth" micrometer is not needed).

                  It is all so simple that you may not believe it.

                  And it may cost nothing either.

                  Just about everyone had all that is needed and if not it is either free or costs very little on the "net".

                  Later - we have to go out for a while - perhaps tomorrow - or the day after.

                  Think/cogitate about it.

                  Comment


                  • #39
                    I have a board above the lathe with a selection of decent nuts on it of various sizes.
                    I cut the thread to fit the nut then put it back.
                    Sod the thread wires.
                    It's too much hard work and none of mine has to pass any Aerospace examination.

                    The proof is the damn nut HAS to fit regardless of what a thread mic or wires tell you.

                    I am more concerned about getting a good thread. You can have a ripped up thread that measures OK on wires but it's not OK to send that job out.

                    [Edit] I do have two sets somewhere or were they small pin punches ?

                    .
                    .

                    Sir John , Earl of Bligeport & Sudspumpwater. MBE [ Motor Bike Engineer ] Nottingham England.



                    Comment


                    • #40
                      Originally posted by lazlo
                      That was a joke Paul -- that's a paper clip
                      Oh! OK. Ha, ha.
                      Paul A.
                      SE Texas

                      And if you look REAL close at an analog signal,
                      You will find that it has discrete steps.

                      Comment


                      • #41
                        rubber hands I could do with them to go with my flexible feet Alistair
                        Please excuse my typing as I have a form of parkinsons disease

                        Comment


                        • #42
                          An alternative wire-measuring method for screw-threads

                          Further to my post #38 at:
                          http://bbs.homeshopmachinist.net/sho...0&postcount=38
                          where I said that I would address the use of a thread wire measurement that did not conform with the traditional "3-wire method", and I was "running late", so I thought I'd better get on with it.

                          Here it is.

                          I will work in "inch" units to cater for the majority of readers. Metric equivalents will work just fine.

                          First a couple of pre-requisites:

                          You will need an open mind and some lateral thinking.

                          You will need a CAD program - pretty well anything between free, cheap, "not big", "simple" etc. and the biggies like AutoCAD and the like. The "little/free" ones should do the job easily.

                          You will need to be able to use CAD for the basics as regards construction and a smattering of geometry.

                          You will need access to thread detail and "sizing" details ("Machinery's Hand-Book" and lots of "tables" that are readily accessible).

                          You will not need a "3-wire" set but you will need ground cylinders that are round and straight, but need not be all or any size other than they must fit into the thread "grooves" and must touch the thread sides/flanks.

                          All wires must project above the outside diameter of the shaft on which the thread is being cut (so as to contact the micrometer).

                          You will need a fairly good "1 thou" micrometer - it will do to read to "tenths" (by eye/estimation) if you think you need to.

                          All thread angles are presumed to be precisely correct as regards angle and symmetry with the line of symmetry at right angles (90 arc degree) to and intersecting the centre-line of the screw-thread under test. This is a "big ask" as many tools are hand-ground and are set with and set with a "Thread Angle Guage" (aka the "fish"). But that seems to be the under-lying assumption/presumption in all texts and tables used in/for 3-wire thread measurement.

                          First of all, for those that may not have seen them - the next pic is my "3-wire" set of sets of 3 precision ground wires. Use of these can be found in "Machinery's Hand-Book" and the like.

                          This table and wires are the traditional way without the use of CAD. This process may prove that the "old way" is not needed. As a check, the wires in the tables may be inserted in place or Wires 1, 2 and 3 in the sketch.



                          Next is the tables applicable to and for use with the "wires" for US and metric threads (60 arc degree):


                          Next is a free-hand (pretty rough - sorry) sketch that will assist to show the alternative methods - and will need referring to as you read the text., so perhaps two concurrent "copies" of this post will make it easy - just "click" either as required.

                          ("Why the Hell doesn't he use a CAD system" I hear you ask. Good question. Answer is that I have AutoCAD and another on the computer and haven't used them in a long time - I have another new one a box and its been that way for two months. Plus I've already sketched it out and it will show some that CAD is not always needed as a sketch will often do).

                          In the method I have in mind the sizes of the wires need not be the same at all - they can all be different.

                          The construction to be used in the CAD process should be self-evident - and must be accurate as regards angles and distances -a "snap" in CAD really (sorry for the intended pun).

                          I hope that you will realise that this method can be used with any thread form/profile/pitch/lead/diameter whether "inch" or "metric" or any other.

                          It, of necessity only applies to "out-side" threads - unless for "internal"(female) threads you use "Repro-Rubber" or "Cerro-Safe" or the like as discussed on threads 35, 36, 37 and 38 on page 4 of this thread:
                          http://bbs.homeshopmachinist.net/sho...t=29487&page=4

                          OK - on with it.



                          As shown in Sketch 1 - the (orange/brown) "wire 1", the "3-wire" approach tries to get all three wires (same size) with their "touch" or "tangent" points as near to the thread pitch line as possible. It is "difficult" to juggle all three wires and the micrometer into correct alignment at once for measurement - which makes accurate and repeatable measurement even more difficult. If all wires (1, 2 and 3) were the same size and used simultaneously, we would have the traditional "3-wire" method in use.

                          If in Sketch 1, all wires are the same diameter (simulating the "3-wire" traditional method) it is only necessary to "measure" the distance from the "top" of wire1 to the mid-point of a line tangent to wires 2 and 3 and parallel to the "screw" centre line.

                          For the "2-wire" method, use say wire 1 with either wire 2 or wire 3 - both can be any diameter and either the same or different diameters. Draw a a line between the centre of the 2 wires used - say wires 1 and 3 (as shown). Extend the line to the outside diameter of each "wire". Measure "over the wires" by using CAD to get that dimension.

                          For the "1 -wire" method, just use one wire - say wire 1 - and "measure" from the outer part of the diameter of wire 1 to the "opposite" side of the diameter of the "screw" perpendicular to the screw centre line to intersect the outside of the "screw" mid-way between the "grooves" (in which wires 2 and 3 are sketched but not used).

                          Sketch 2 shows the use of a gear caliper measuring the width of the screw thread ("gear tooth") at its thread pitch line position (ie "Dimension "A"). But this in practice will need to be "off-set" by the thread helix angle (computed at "Sketch 3") and applied as at Sketch 4.

                          Note that all methods as shown are the theoretical sizes as in a CAD drawing/file. The actual outside diameter of the screw shaft under test will have to be physically measured and inserted into the CAD file for both the "1 - wire" and "no wire" ("gear-caliper") methods as the OD of screw shaft is used.

                          Any or all of these methods are applied to the job and micrometer (all "wire" methods) and gear caliper ("no - wire" method), and are the aim to be achieved on the lathe as screw-threading progresses.

                          Comment


                          • #43
                            The only issue with any method which relies on the major diameter of the screw is that it relies on the major diameter of the screw.

                            Otherwise, it works.

                            The thing about the wires, or the mic, triangles, etc, is that it measures ON the pitch line (if wires or mic etc are correct) and so is independent of the major diameter, which may be incorrect on the screw as-formed.

                            However, it is perfectly true that 2 wires of one size and one of another, OR a single wire, can be used, with suitable calculation, and allowing for the fact that the MD may not be correct. if you MEASURE the MD first, you can somewhat compensate for that error.

                            You do need to realise that the MD is not a "primary attribute", and may be in error by amounts that dwarf the pitch diameter spec.

                            On Geometry......

                            Tiffie's remarks about CAD remind me that CAD is not a substitute for knowing how to figure out something like this geometrically.

                            You can figure out the seating depth of a wire in a certain pitch of thread relatively easily, or the best size of wire.

                            And you can figure out the target measurement over the wires if you lose your handy little reference sheet also.

                            All by relatively simple geometry.
                            CNC machines only go through the motions

                            Comment


                            • #44
                              That too

                              Thanks JT - your feed-back is excellent and very much appreciated - truly.

                              The wires of course can be any size so long as they touch both sides/flanks of the thread groove and project above the screw OD. If the groove angle is exactly the correct angle, is symmetrical about its axis all will be OK. But as I said that's a huge assumption. But if that is so then the wires should give a correct reading no matter what their size and no matter whether they are tangent at the pitch line.

                              As an extension to the previous paragraph, I guess that a user could "run" with varying errors of angle and symmetry and see what the errors might be with an incorrectly ground or mounted/positioned screwing/grooving tool. I expect that accuracy will need to be a lot better than a usual/crude "degree" graduations on many grinding aids and tools!!

                              The system - as I guess you realised - will work for any thread form of any pitch or lead - even for plain grooves (threads, or "V"s with zero pitch).

                              I had to use the major diameter of the screw for several of the methods. It should not be too hard for the average machinist to take into account - even to the extent of taking several spaced readings and perhaps averaging them out.

                              Almost anything has to be better than trying to use the "3-wire" method let alone getting it accurate and repeatable.

                              I could see the process that I came up with some while ago but I did nothing about it until this thread came up.

                              It seemed that the "3-wire" method was both "cast in stone" and "Holy Writ" which everyone seemed to complain about, accept and never question. That situation is like a red rag to a bull to me!!!

                              That is not my way - I just went that "extra yard" - as both a challenge and as a necessity I guess.

                              It just seemed a logical use of CAD and computers.

                              It also seemed a good way for some to get to the end result with a few more options and less cost and not to have the expense of buying a "3-wire" set.

                              I haven't got as far as seeing if it will or can apply to gears or racks yet. I was thinking about using wedges which are double the pressure angle of a gear - a rack will probably need "wires" or a modification to them. But that is for later.

                              I would really appreciate more advice regarding "fine-tuning" and knocking off the snots and rough edges.

                              Comment


                              • #45
                                If you are refering to knocking off the rough edges of the cut thread, I always use a brass wire brush. Use a light touch if threading aluminum or other soft metal. Any rough edges on the OD can really throw off any method that uses that surface for measurement.

                                While reading this, I had a thought. For a 60 degree thread, I wonder if hex stock would be better than round wires? You would have a line of contact instead of a point so there would be less chance of pressing it into the surface. Of course, you would need precision hex stock, not just any old allen wrenches.
                                Paul A.
                                SE Texas

                                And if you look REAL close at an analog signal,
                                You will find that it has discrete steps.

                                Comment

                                Working...
                                X