No announcement yet.

Level Accuracy

  • Filter
  • Time
  • Show
Clear All
new posts

  • Level Accuracy

    I need to get a machinist's level. What I want to know is just how accurate does it need to be? And what length is best? I have a South Bend 9 to set up and will likely use it on the mill-drill also. Any advice will be appreciated.
    Paul A.
    SE Texas

    Make it fit.
    You can't win and there is a penalty for trying!

  • #2
    The only machine manual have says use a " 'precision ground bulb' level. Carpenter's levels or combination square levels is not sensitive enough." There are 12" european levels that claim sensitivity of .0005" foot available from MSC, which may be quite satisfactory in the $125.00 price range.The only hitch with a long level is what most of us have for parallels for setup may be less than ideal. No connection other than being a paying customer.

    [This message has been edited by gvasale (edited 08-02-2003).]


    • #3
      As SENSITIVE as your work needs to be accurate, and as sensitive as you want the machine accuracy to stay........

      If you use an 0.005 per foot level, like a Starret 98 (?) then you may have a couple thous error per foot, in your leveling and maybe in your parts. If you use an 0.0002, your error is less.

      The machine will gradually wear to fit the saddle to the shape of the bed, so errors in leveling are reflected in uneven wear, and eventual "rocking horse" saddles.

      Takes a long time, depending on use, cleanliness, and lubrication but it happens.

      If your machine is used, the previous owner may have already worn it in in a bad way and leveling it may actually make things worse for you. But, it is the only standard for avoiding twist in the bed, so it's best to do it.

      0.0002 per foot is best, 0.005 is better than an HD carpenters level............

      I don't see where anything over 8" is handy at all, but the bed of my Logan is only around 7" wide. For a bigger machine, it might be handier to be longer in some cases, but then you can always use a beam parallel. It's hard to saw off a bit of a level..........


      • #4
        I think the emphasis on getting a machine way level is really an effect of trying to get the ways of a machine straight. With a long and not very stiff way such as is the case with conventional lathes it is important to minimize warping of the ways. One way to try to do this is to use a level and check level at increments along the ways. By adjusting the supports until the ways are in the same plane at all test points you have some idea that the ways are straight. I am going over this because my experience has always been with super precision machines that are so stiff and short coupled that we never worried about whether the machine was exactly level but we did worry a lot about slide straightness. All the straightness and squareness measurements were made with laser interferometers. Once we had an application that required sub-arc second resolution on uniformity of angle of an extended surface. The surface was horizontal so "level" applies to the measurement. The required resolution was better than any of the electronic levels we knew of. A creative PhD type thought up an ultra level for the application and we built it and used it effectively. It consisted of a hollow column of steel with a three point base that was ground and lapped flat and nominally square to the axis of the square column. A thin gage foil of steel was suspended from the top of the column with a weight attached to the end to keep the band of foil taunt. A capacitance sensor was positioned to read the location of the foil. The capacitance sensor had a resolution adjustable to as fine as 1/100th of one millionth of an inch. The the "level" thus was a sealed container made of one material and insulated from transient heat. The instrument was positioned and the capacitance gage was zeroed. Measurements at other positions were recorded relative to the first position. Any angular change of the base was indicated as change relative to the foil which remained pointed at earth center. The column had a vacuum port on it to allow extraction of the air but we found that not to be necessary. I have thought since then that a similar "level" could be built which would use a dial indicator. The height of the column was about 12" and the gage strip was about 1/2" wide and about 10" long. Obviously the longer the hanging foil with weight the higher the resolution but also the heavier and more awkward the instrument. The capacitance gage senses on an area so it was not sensivtive to centering which might effect a dial indicator. The capacitance gage is an expensive device costing about $10k at the time (a little cheaper now) so it is too expensive for most people. If you built something like this with a dial indicator or better with an electronic digital indicator I think it would be a useful tool. Laser straightness interferometers would do the lathe line up job more effectively though and autocollimators and alignment telescopes also do a good job.


        • #5
          Here's something to think about.

          I bought a $10 Chicom laser level at HF for a carpentry application. It
          projects a laser spot over a considerable distance. I set the level on my
          (12" long) surface plate so the laser spot hit a vertical scale fifteen feet
          away across my shop. I slipped a 0.004" feeler gage under one end of the
          level and noted that the laser spot moved ~1/16" on the scale. Now,

          sin x = 0.004/12

          But the angle 'x' is tiny so sin x ~= x. So:

          x ~= 0.004/12 = 0.000333 radians = 333 microradians (urad)

          333 urad is about 68 arc seconds (or, obviously, 0.004" per foot). Detecting
          a one arc minute slope for $10 invested in tooling is pretty cool and you
          don't need to worry about the vials in the level.

          I found that, at 15', reading the beam location to better than 1/16" is
          difficult. I have an idea, as yet untried, on how to deal with that. If I
          build a simple circuit consisting of little more than a biased
          photo-transistor powered by a 9v battery, I can use it as a light amplitude
          detector. (The voltage across the biasing resistor will be proportional to
          the light intensity falling on the photo-transistor.) If I attach this to a
          height gage, I'll be able to measure the height of the maximum signal coming
          from the laser light. This will take my bifocals and the laser speckle
          problem out of the equation.

          I haven't pursued this as a technique for detecting lathe twist but I think it
          has potential as an economical intermediate between a coarse, inaccurate level
          and a super-expensive machinist level. Perhaps some of you clever guys may
          want to explore it.
          Regards, Marv

          Home Shop Freeware - Tools for People Who Build Things

          Location: LA, CA, USA


          • #6
            one wat to double check your level is to flip it end for end on your surface and see if your bubble is in the same spot, or it moved. It won't tell you how sensitive it is, just if you should throw it away or not!


            • #7
              This might work for you out there that are electronically endowed. In the May issue of "Nuts & Volts" there is an article on how to build a "tiltometer" by using 2 small cups of mercury 24" apart that can measure the tilt of the earth of 125 mills in 200 MILES. Now that's a level!
              Don Warner


              • #8
                The Starrett #199 is also usable as a straight edge when scrapping. They have a seasoned hand scrapped cast iron base. You my find it too accurate for many applications, but they are ideal for machine set-up. Used them a lot on plastic Injection Molding machines.


                • #9

                  Even on a high quality laser beam there is more divergance in that distance than you claim to be able to differentiate. Note that a Starrett #199 is 15 arc seconds - accuracy very near to what you suggest the cheesy laser is.

                  I don't think so.


                  • #10
                    An average quality autocollimator has a resolution of 0.1 arc second. That was the standard alignment tool before laser interferometers. The Brunson web site has a tutorial on how to use one.


                    • #11

                      Gosh, I wish you'd told me that what I was doing was impossible before I went
                      ahead and wasted all that time doing it.

                      15 arcseconds is "very near" to 68 arcseconds according to you? I don't call
                      a factor of 4.5 very near even when I'm doing woodworking. The Starrett Master
                      Precision Level claims 10 arcsec resolution, but it costs $520 - kind of steep
                      for the hobbyist. Rather than paying that kind of money, I'd be happy to get
                      within an arcminute cheaply and then take out the last few thousandths of
                      twist using the standard two-collar test bar.
                      Regards, Marv

                      Home Shop Freeware - Tools for People Who Build Things

                      Location: LA, CA, USA


                      • #12
                        Hey mk,

                        double or triple your resolution capability. Throw some surface mirrors in there and artificially lengthen the 'run'. of course you'll have to go back and forth realigning mirrors when the angle is too great. used to do this on a small scale for sensitivity (gain) adjustments on oscillographs. try to collimate the laser spot too with adjustable focus lens. no longer $10.00. but fun to mess with.


                        • #13
                          mklotz: You can increase your resolution tenfold or reduce your target distance tenfold. Instead of targeting a flat surface, target a sloping surface. I've done this and shrunk the size of an optical fixture from 72" to 7". Fits a lot nicer on my bench. You'll have an long elliptical spot but who cares (in many cases).

                          thrud: Missed your reply ... all you gotta do is use the center of mass of the laser spot (after calibrating it and your eyeballs somehow ).

                          [This message has been edited by nheng (edited 08-05-2003).]


                          • #14
                            Cass, It occurs to me that the "Super level" you describe is probably sensitive enough to be affected by the gravitational attraction of nearby equipment. Did you try to make any such determination?

                            Slightly off topic, I built a Lehman style seisomgraph back in the 80's which is still set up and operational in my basement. A Lehman seismograph is a simple instrument and is basically a tilt meter. Mine is sensitive enough to detect magitude 5 earthquakes or so anywhere in North America and mag 6 just about anywhere in the world. It is so sensitive that it will detect someone merely walking around the outside of the house or a car pulling in the driveway 150 feet away. It easily detects six second microseisms that are thought to be either wave action on the coast 500 kilometers away and/or the effect of atmospheric pressure on the earth's surface. If that sounds unlikely the math shows how it may be so. A change of 25mm on the barometer(about one inch)is about 39 grams per sq cm or about 390 kilos per sq metre or about 3,900 metric tonnes per Hectare (about 2.5 acres) or about 390,000 tonnes per sq kilometre. I can clearly see approaching low pressure cells with the instrument.

                            Oh yeah, I forgot to mention that I used my SB9 to wind the pickup coil on the siesmograph. 10,000 turns (as per turn counter) of #70 wire as fine as a hair. I managed to do it with out a break on the first try.

                            [This message has been edited by Evan (edited 08-05-2003).]
                            Free software for calculating bolt circles and similar: Click Here


                            • #15
                              robert phillips -

                              Actually I did fold the path back on itself with a surface mirror so the beam
                              displacement was ~1/8" for a 0.004" shim under the level. That has the added
                              benefit of putting the scale where the beam position is noted close to the
                              operator although that's not of great consequence for a one-time operation like
                              unwinding a lathe.

                              nheng -

                              I like the tilted approach. Gotta try that. Despite Thrud's encouragement,
                              I'm going to build the height gage-mounted beam maximum detector as well.
                              Regards, Marv

                              Home Shop Freeware - Tools for People Who Build Things

                              Location: LA, CA, USA