Lathe accuracy checks

⌖How to inspect

• Installation and levelingCommon

  ToolsPrecision level, wedges and leveling bolts

  ProcedureColchester: put steel wedges under the edge of the base so that it sits against the floor with no gaps, then fix it with bolts and add grout if necessary. Haas ST series: raise it with the four corner screws to even out the tension, lift some screws to support it at three points, bring the bubbles of the levels in the Z direction and along the turret face to the center with separate screws, and adjust while moving the turret to both ends of Z until the twist disappears. Bring the lifted screws into contact by hand and then tighten them only 10°

  TolerancesHurco: with a level of 0.020 mm/m resolution, none of the 9 points may exceed 0.020 mm/m (call service if exceeded). On a lathe, place the level on a base bolted to the turret

  The Haas procedure says to center the bubble until the twist is gone, and gives no numeric tolerance

  ^[27][28][29]

• Level sensitivity, and how to check the level itself (reversal method)Common

  ToolsPrecision level (JIS B 7510 Class A)

  ProcedurePlace it on a surface plate, read the bubble, then turn it 180° at the same spot and read again. If the reading is at the same position, no adjustment is needed. If it moves the same number of divisions in opposite directions, that is the level's zero error; if it does not move but is off, the base is tilted. If the readings are inconsistent, take half the largest error and correct it with the level's adjusting screw, starting from the side with the larger error, then correct the base. Read the average of the two ends of the bubble. Let the level acclimate to the temperature of the place where it will be used. If you hold a 0.02 mm/m level in your hand for long, body heat shifts its zero

  TolerancesSensitivity: 1 division = 0.02 mm/m (about 4 arcseconds), 0.05 mm/m (about 10 arcseconds), 0.1 mm/m (about 20 arcseconds). One arcsecond is about 4.85 µm per 1 m

  ^[26][25]

• Parallelism between the spindle centerline and the carriage (Z-axis) motionCommon

  ToolsTest bar (alignment bar), indicator with 0.002 mm resolution, magnetic base

  ProcedureHardinge: mount the test bar in the spindle and fix the indicator on the tool post. At the tip of the bar, set the spindle to the middle of the runout and zero the indicator, then move the carriage 152.4 mm (6 in) toward the spindle and read. Do this in two directions: vertical (up/down) and horizontal (front/back). Haas: feed the Z axis and watch the change in reading on the top and side of the bar

  TolerancesHardinge HLV-H: vertical 0 to +0.012 mm (tip low), horizontal 0 to −0.002 mm (tip toward the operator). Haas ST/DS: 0.010 mm per 254 mm on both top and side

  Both are in-house standards for each manufacturer and model. The tolerances in JIS B 6202 and JIS B 6331-1 could not be confirmed because the standard text is paid

  ^[30][31]

• Tailstock alignmentCommon

  ToolsTest bar (supported between centers), indicator with 0.002 mm resolution

  ProcedureHardinge: fit a center in the spindle, extend the tailstock quill 88.9 mm, and support the test bar lightly between the centers (it should still turn by hand). Lock the tailstock with the clamp lever at 17 N·m and lock the quill too. Set zero at the middle of the runout at the spindle end, then move the carriage 152.4 mm toward the tailstock and read. Do this in two directions: sideways (horizontal) and up/down (vertical)

  TolerancesHardinge HLV-H: horizontal 0 to +0.0025 mm (tailstock side toward the operator), vertical +0.013 to +0.025 mm (tailstock side high). Haas: parallelism between the tailstock quill and the Z axis is 0.025 mm per 102 mm on both side and top; concentricity between the tailstock and the spindle is 0.025 mm at each end

  Hardinge says that if the horizontal reading is off, first check for chips or burrs. The tolerance is set so that the tailstock side is slightly higher

  ^[30][31]

• Correction for an indicator set at an angleCommon

  ToolsLever-type (test) indicator

  ProcedureAs a rule, set the stylus at a right angle to the direction being measured. When that is not possible, true value ≈ reading × correction factor

  TolerancesCorrection factors: 10° = 0.98, 20° = 0.94, 30° = 0.86, 40° = 0.76, 50° = 0.64, 60° = 0.50

  ^[33](1 source, for reference)

• How to measure bed straightness, and its toleranceCommon

  Tools—

  ProcedureNo procedure or tolerance could be found in public primary sources

  TolerancesNot confirmed

  (could not be confirmed in published sources)

• Tolerances for each test in JIS B 6202 and JIS B 6331-1 (spindle runout, parallelism between spindle and carriage, center height difference of the tailstock, etc.)Common

  Tools—

  ProcedureThe standard text is paid, and it was not found in any public commentary either

  TolerancesNot confirmed

  (could not be confirmed in published sources)

• Dialing in level with a test cutGeneral-purpose

  ToolsSharp tool, micrometer, dial indicator

  ProcedureTake a very light cut on two disks (spaced 12 in apart) or a bar held in the chuck, and measure both diameters with a micrometer. If they differ, adjust the leveling screws until half of the diameter difference disappears, then cut again and check

  TolerancesThe diameters at the two locations should be the same

  Presented as a method for when higher accuracy is needed, such as in tool room work

  ^[27](1 source, for reference)

• Parallelism between the centerline between centers and the bed (relation to thread-cutting accuracy)General-purpose

  ToolsTest bar, indicator, autocollimator

  ProcedureRead a test bar supported between centers while feeding the carriage, and adjust mainly the tailstock center side. The carriage tilt is read with an autocollimator by feeding in both directions

  TolerancesValues reached in the study (not tolerances): about 1 µm in the horizontal plane, about 8 µm in the vertical plane, carriage tilt about 1 to 3 arcseconds

  It concludes that non-parallelism in the horizontal plane strongly affects the thread pitch error

  ^[32](1 source, for reference)

• Gib (slide plate) adjustmentGeneral-purpose

  ToolsHex key

  ProcedureHardinge: loosen the adjusting screw on the large end of the tapered gib by one turn, tighten the nut on the opposite side slightly, then lock it again. Move the carriage and check the feel

  TolerancesIt should feel slightly heavy but not bind. Over-tightening does not improve performance (Hardinge)

  Colchester says that when chatter appears, first check whether the slide strips are adjusted correctly

  ^[30][27]

• Feed screw backlashGeneral-purpose

  ToolsHandwheel dial

  ProcedureTighten the adjusting nut of the feed screw nut to reduce backlash, then lock it with the lock nut

  TolerancesHardinge HLV-H: up to 2 dial divisions for the tool slide feed screw, up to 3 divisions for the tailstock

  ^[30](1 source, for reference)

• Re-leveling intervalNC

  ToolsPrecision level (0.020 mm/m resolution)

  ProcedureUsing the values the service technician set at installation as the baseline, check monthly for 6 months after installation, then every 6 months

  TolerancesWithin 0.020 mm/m at 9 points (value for Hurco's 3-axis machines)

  ^[28](1 source, for reference)

• Spindle runout (face and radial)NC

  ToolsIndicator (fixed on the turret face), spindle alignment bar

  ProcedurePut the indicator on the spindle face and turn the spindle one revolution to read the total runout. For radial, read at the base and the tip of the bar

  TolerancesHaas ST/DS: axial runout of the face 0.005 mm max, 0.005 mm at each end of the bar

  The measurement positions and tolerances in the standard ("at the spindle nose and 300 mm away") could not be confirmed

  ^[31](1 source, for reference)

• Circular accuracy (ballbar)NC

  ToolsBallbar and lathe accessories (arm for the turret and bar for the spindle)

  ProcedureRun circular interpolation in the ZX plane and measure two passes, clockwise and counterclockwise (adding acceleration/deceleration arcs before and after). With the lathe kit, use a 360° arc of 100 mm radius; if the stroke is not enough, use a 220° partial arc or the 50 mm radius small-circle kit

  TolerancesThe machine-side tolerances are in the text of ISO 230-4 and ISO 13041 (paid) and could not be confirmed

  The Renishaw lathe kit requires at least 220 mm of clearance in X from the spindle centerline, a 25 mm spindle diameter, a tool post that can hold a 20 mm or 25 mm shank, and circular interpolation in the ZX plane

  ^[34][22][23]

• Positioning accuracy and repeatabilityNC

  Tools(Standard text not confirmed)

  ProcedureMeasure each axis directly and repeat the measurement at each position (outline of ISO 230-2). For lathes, ISO 13041-4 / JIS B 6331-4

  TolerancesNot confirmed (standard text is paid)

  ^[18][8][9]

• Thermal displacementNC

  Tools—

  ProcedureThermal displacement is caused by temperature change around the machine, heat from the machine, and heat from machining (Okuma). The test method is generally covered by ISO 230-3 / JIS B 6190-3; for lathes, ISO 13041-8 / JIS B 6331-8

  TolerancesNot confirmed

  ^[35][20][14]

≡Related standards

• General-purposeISO 1708:1989 Acceptance conditions for general purpose parallel lathes — Accuracy testing / JIS B 6202:1998 Engine lathes — Test methods for accuracy (adopted from ISO 1708 with modifications = MOD)ISO 1708 was reviewed and confirmed in 2022 and is current. It covers accuracy tests only; running tests such as vibration, abnormal noise, and stick-slip are outside its scope as they are done before the accuracy tests^[1][2]
• NCISO 13041-1:2020 / JIS B 6331-1:2022 Test conditions for numerically controlled turning machines and turning centres — Part 1: Geometric accuracy of machines with a horizontal workholding spindle (identical = IDT)The JIS table of contents lists the machine leveling, test sequence, measuring instruments, and tolerances, followed by tests in this order: workholding spindle, straightness of motion, relation between linear axes and spindle, tailstock, tool turret^[3][4][5]
• NCJIS B 6331-2:2022 (geometric accuracy of machines with a vertical spindle, ISO 13041-2:2020 IDT), JIS B 6331-3:2013 (machines with an inverted workholding spindle, ISO 13041-3:2009 MOD)^[6][7](1 source, for reference)
• NCISO 13041-4:2004 / JIS B 6331-4:2006 Accuracy and repeatability of positioning of linear and rotary axes (IDT)^[8][9]
• NCJIS B 6331-5:2019 Accuracy of feedrates and interpolation (ISO 13041-5:2015), -6:2013 Accuracy of finished test pieces (ISO 13041-6:2009), -7:2006 Evaluation of contouring performance in the coordinate planes (ISO 13041-7:2004). All are IDT^[10][11][12](1 source, for reference)
• NCISO 13041-8:2004 / JIS B 6331-8:2006 Evaluation of thermal distortions (IDT)^[13][14]
• CommonISO 230-1:2012 / JIS B 6190-1:2016 Test code for machine tools — Part 1: Geometric accuracy of machines operating under no-load or quasi-static conditions (IDT)JIS B 6190-1 supersedes the former JIS B 6191:1999. Its table of contents includes "installation of the machine before testing", "machine condition before testing", and "runout of rotating parts"^[15][16][17]
• NCISO 230-2:2014 / JIS B 6190-2:2016 Determination of accuracy and repeatability of positioning of numerically controlled axes (IDT, including Amendment Amd 1:2016)A method that measures each axis directly and repeats the measurement at each position. It can be used for acceptance, periodic inspection, compensation, and so on^[18][19]
• NCISO 230-3:2020 / JIS B 6190-3:2023 Determination of thermal effects (IDT)^[20][21]
• NCISO 230-4:2022 / JIS B 6190-4:2025 Circular tests for numerically controlled machine tools (IDT)^[22][23]
• CommonJIS B 6190-7:2019 Geometric accuracy of axes of rotation (ISO 230-7:2015 IDT). Related to spindle runout (error motion of the axis of rotation)^[24](1 source, for reference)
• CommonJIS B 7510:1993 Precision levels (Daibishi Keiki's manual shows the JIS mark on its Class A product)^[25][26]

The standard texts are paid, so tolerances are written from sources such as in-house standards published by machine manufacturers. For the correspondence between standards, see the standards correspondence table as well.