CNC turning eccentricity elimination correcting spindle runout for precision work

Content Menu

● How Runout Actually Shows Up on the Part

● Measurement That Doesn’t Lie

● Mechanical Fixes — The Ones That Stay Fixed

● On-Machine and Control-Based Compensation

● Real-World Turnarounds

● When You Need Sub-Micron Territory

● Closing Thoughts

 

Spindle runout remains one of the most common reasons a turned part ends up out of round or fails concentricity checks on the CMM. In shops running tight-tolerance work — aerospace seal diameters, hydraulic spools, fuel injector bodies, or instrument bearings — even three or four microns of runout at the chuck face can push the finished feature outside 0.0002″ (5 µm) concentricity. The problem compounds quickly because runout doubles at the diameter: 5 µm TIR at the locating surface becomes 10 µm diameter-to-diameter error on the part. Over the past twenty-five years on the floor and in process engineering, I’ve seen the same handful of root causes repeat across dozens of lathes: Okuma, Mori Seiki, Doosan, Mazak, Haas, DMG, and even high-end Studer and Hardinge machines. The fixes are straightforward once you measure properly and attack the problem systematically.

Runout that creates turning eccentricity comes from three main zones: the spindle bearings themselves, the tool-holder-to-spindle interface (taper and face contact), and the workholding (chuck, collet, or fixture). If any one of them is off, the part spins on the wrong axis. Let’s walk through how to find it, prove it, and kill it for good.

How Runout Actually Shows Up on the Part

A 50 mm shaft turned on a spindle with 8 µm TIR at the chuck face will have a turned diameter that wobbles 16 µm peak-to-peak. On a CMM you’ll see a “banana” profile in the polar plot and concentricity errors that grow with distance from the chuck. I’ve measured crankshaft main journals that were 0.0012″ (30 µm) out because the spindle nose had a bent test bar reading of 0.0006″ after a minor crash no one reported. Another common sight: medical hip stems in 17-4PH where the bearing diameter ran 12 µm eccentric to the taper — the spindle taper had picked up a piece of swarf and seated 0.0004″ off axis.

Measurement That Doesn’t Lie

Stop trusting the cheap import test bar and the $29 dial indicator for anything tighter than 0.001″. Here’s what actually works.

1. Use a certified test arbor with ≤ 1 µm cylindricity (Lion Precision, TAC Rockel, or EDM-ground in-house).
2. Indicate two positions: 20 mm from the spindle nose and 150–200 mm out. The difference tells you whether the error is parallel (bearing preload or contamination) or angular (taper damage or drawbar issue).
3. Spin at 50 RPM for static checks, then step up to operating speed with a non-contact probe (capacitive or eddy current) to catch dynamic errors.
4. Always warm the spindle first — 30 minutes at 3000–4000 RPM is usually enough for the bearings to stabilize.

Real shop example: a tier-1 automotive supplier was scrapping 8% of camshafts because of lobe eccentricity. A quick warm-spindle test with a Lion capacitive gauge showed 9 µm growing to 22 µm after 20 minutes. The oil chiller thermostat had drifted; fixing it dropped runout to 3 µm and saved the line.

Mechanical Fixes — The Ones That Stay Fixed

Clean the taper. Seriously. Alcohol, lint-free cloth, and a taper brush. Then check contact with Prussian blue — you want 85–90% on the taper and full ring contact on the face. Anything less and the holder will walk under load.

Upgrade the gripping system. Standard ER collets and 3-jaw chucks rarely repeat better than 10–15 µm. Switch to:

• Schunk Tendo E compact hydraulic holders (≤ 3 µm repeatable)
• Rego-Fix powRgrip (sub-2 µm with the right collet)
• HSK63 or PSK (Capto) face-and-taper contact systems instead of BT40/CAT40
• Compensating pie-jaw chucks (SMW Autoblok, Kitagawa) for second-op work

Spindle rebuilds pay for themselves fast on high-hour machines. A typical 40-taper cartridge spindle with 15–20 µm runout can be brought back to 1–2 µm by grinding the taper, replacing angular-contact bearings, and re-balancing the drawbar assembly. One aerospace shop I worked with rebuilt two Mori NL2500 spindles for $18k each; the resulting 1.5 µm TIR let them hold 4 µm total on titanium landing-gear pistons and paid back in six months through scrap reduction alone.

On-Machine and Control-Based Compensation

Most modern controls (Fanuc 31i-B5+, Siemens 840D sl, Heidenhain TNC 640) have spindle error compensation tables. Measure a polar plot of the runout (magnitude and angle) at several speeds, enter the values, and the control nudges the X and Z servos in real time to cancel the error. A German toolmaker turning carbide form tools reduced effective runout from 7 µm to 0.8 µm using this on an Index C200.

For shops without built-in compensation, a simple macro can do the same thing: probe a turned test ring on the machine with a touch probe, calculate the center offset, then shift G54 X and Z by half the measured error. Repeat twice and you’re usually under 2 µm.

Real-World Turnarounds

Case 1 – Medical 316L implant shafts Original runout: 14 µm on a ten-year-old Citizen L20 Swiss lathe Fixes applied: cleaned HSK taper, switched to Albrecht precision collet (0.002 mm TIR spec), added 20-minute warm-up cycle Result: 1.8 µm spindle TIR, finished concentricity 3–4 µm, 100% first-pass yield

Case 2 – Aerospace Inconel 718 turbine spacer Original: 28 µm eccentricity on a DMG NTX1000 Root cause: cracked inner bearing race (machine had 42,000 hours) Fix: full cartridge rebuild by GTI Spindle Technology Post-rebuild: 1.2 µm TIR, eccentricity now 2–3 µm on 180 mm diameter

Case 3 – High-volume aluminum transmission valve bodies Problem: thermal drift of 18 µm over an eight-hour shift Solution: added spindle oil chiller setpoint control (±0.5 °C) and Fanuc thermal compensation package Result: drift limited to 4 µm, eliminated afternoon scrap entirely

When You Need Sub-Micron Territory

For optics, gage blocks, or fuel-system pilot diameters, air-bearing spindles (Professional Instruments, Moore Nanotechnology) or hydrostatic oil bearings (Hardinge, Studer) are the only realistic answer — mechanical bearings simply can’t stay under 0.25 µm consistently. But 95% of precision turning jobs live happily in the 2–5 µm runout zone if you maintain the machine properly.

Closing Thoughts

Eliminating turning eccentricity caused by spindle runout is less about exotic technology and more about discipline: measure accurately and often, keep tapers spotless, use the right gripping tools, warm the spindle, and rebuild before the bearings grenade. Do those things and you’ll turn parts that measure dead nuts on the CMM instead of coming back from inspection with red tags. I’ve watched shops go from 15–20% scrap on precision turned components to near-zero just by treating spindle runout as the enemy it is. Start tomorrow morning with a proper test bar and a warm spindle — you’ll be amazed how many “mystery” problems disappear.

Q&A: Questions I Get Asked Most on the Floor

1. How much runout is acceptable before I need to act?
   Under 5 µm at the spindle nose for most precision work; under 2 µm if you’re chasing 0.0002″ concentricity.
2. Will truing the chuck jaws in place fix spindle runout?
   No — it only masks grip errors. Runout is measured with a test bar, not soft jaws.
3. Is it safe to shim a toolholder to correct runout?
   Never. You’ll destroy repeatability and risk a crash.
4. Should I indicate the spindle every shift?
   Weekly minimum, daily on critical machines or after any crash.
5. Are Big-Plus or HSK really worth the extra cost?
   Yes — dual-contact systems routinely cut runout 60–80% compared to single-taper 40/50 holders.