AutomateDirectUSAMotion LibraryShop the parts catalog

Troubleshooting

Tracing Lathe Chuck Cylinder Vibration Above 2300 RPM

Pneumatics and hydraulics Published 9 min read

Short answer

A rear-mounted rotating chuck cylinder that shakes above 2300 rpm is unbalance or runout until the checks say otherwise. Hold adapter and cylinder runout under 0.0005 in TIR — a 0.015 in TIR reading at the sensor end is thirty times that — and check mounting-bolt burrs, draw-tube threads and clamp pressure before condemning spindle bearings.

On this page
  1. Why does the chuck cylinder vibrate above 2300 rpm?
  2. How much runout is acceptable on a rotating chuck cylinder?
  3. What should I rule out before pulling the cylinder?
  4. Are the mounting bolts and adapter plate sound?
  5. Where do I measure runout, and in what order?
  6. How much play is too much in the spindle and cylinder?
  7. Can the draw tube or clamp pressure be the real cause?
  8. When is the fix a rebuild instead of an adjustment?
  9. Questions asked

A rotating hydraulic chuck cylinder that starts shaking above about 2300 rpm is a balance and mounting problem until proven otherwise. Anything wrong behind the spindle — a burr under the adapter, a stretched bolt, a bent draw tube — becomes rotating unbalance, and the rear position sensors and spindle bearings take the load. The runout target is under 0.0005 in TIR; a 0.015 in reading at the sensor end is thirty times that. Work the checks in order: each names the reading to take and what it decides.

Why does the chuck cylinder vibrate above 2300 rpm? #

Because force from a fixed eccentricity grows with the square of speed: from 2300 to 3200 rpm the shaking force multiplies by (3200/2300)² ≈ 1.9, and any runout behind the spindle nose delivers it straight into the rear sensors and spindle bearings.

Two mechanisms produce a vibration that appears at a speed and worsens above it:

  • Unbalance or runout. A fixed eccentric mass pushes once per revolution, and the force rises with speed squared. A cylinder tolerable at 1800 rpm carries nearly double the load at 3200 rpm.
  • Resonance. A loose or cracked joint lowers the stiffness of the spindle-plus-cylinder assembly and drags a natural frequency into the operating range. This looks like a band of bad speeds that may ease off above it.

Before touching hardware, sweep the spindle in steps — 1500, 1800, 2100, 2300, 2600, 3000 and 3200 rpm — with a magnetic-base accelerometer, or at minimum a dial indicator on the housing, at the spindle front, the cylinder housing and the motor. Record three things: the speed where the shaking starts, whether the dominant frequency equals spindle speed (2300 rpm ÷ 60 = 38.3 Hz; 3200 rpm = 53.3 Hz), and whether it changes when clamp pressure is applied. To tag the once-per-rev line, measuring rpm with a proximity switch on a pulley gives a clean reference signal.

Calculated example, not a measured force. Assume a 25 kg cylinder showing 0.015 in TIR at the rear. Eccentricity e = TIR/2 = 0.0075 in = 0.19 mm = 1.9 × 10⁻⁴ m, and rotating force F = m·e·ω²:

  • At 2300 rpm: ω = 2300 × 2π ÷ 60 = 241 rad/s, so F = 25 kg × 1.9 × 10⁻⁴ m × (241 rad/s)² = 276 N ≈ 62 lbf.
  • At 3200 rpm: ω = 335 rad/s, so F = 25 kg × 1.9 × 10⁻⁴ m × (335 rad/s)² = 533 N ≈ 120 lbf.

The mass is an assumed input; weigh your cylinder or take its mass from the drawing. The decision from the sweep: cylinder end dominant and frequency locked to 1× spindle speed means keep reading; motor end dominant and frequency free-running means start with the drive-end checks two sections down.

How much runout is acceptable on a rotating chuck cylinder? #

Under 0.0005 in TIR is the working target for a chuck cylinder and its adapter; 0.0001 in is a precision-spindle figure that a cylinder with its deliberate clearance to the adapter will not hold, so do not chase it.

The 0.0005 in line comes from workholding practice: below it, the usual route on an installed rotating part is machining it in place in its own bearings, which is not something you do to a chuck cylinder on a production lathe. Sub-0.0001 in TIR belongs to another class of machine entirely — designing ultra-precision spindles for sub-0.0001 in TIR applications is a separate discipline. On a cylinder that trims against its adapter with bolt torque, plan on holding under 0.0005 in and spend the remaining effort on the checks below.

What should I rule out before pulling the cylinder? #

Four cheap checks come first: belt tension and condition, spindle motor mounting bolts, machine level with anchor bolts, and the pump coupling in the hydraulic unit, because any of them can put a threshold-type shake on the spindle and none needs the cylinder off.

  • Belts: confirm tension, watch for whipping at speed, and file burrs out of the pulley grooves. A whipping belt shows up at a frequency that does not track spindle speed.
  • Spindle motor mounting bolts: tighten them; a loose motor mount reproduces the same speed-threshold shake.
  • Machine level and anchors: re-level, tighten the anchor bolts, and align a bar feeder to the spindle if one is fitted. If the mounts under the machine are corroded past adjustment, replace them with machine leveling feet before spending time behind the spindle.
  • Hydraulic power unit: inspect the pump coupling; one that had shed its rubber element turned up on a machine with this same symptom.

Treat these as checks, not guaranteed fixes. On one documented lathe, leveling, feeder alignment, new motor bearings, a rebuilt hydraulic unit, new belts and filed pulleys all failed to remove vibration above about 2300 rpm, while the machine held tolerance and finish throughout. If the motor end reads highest in the sweep and the frequency does not track the spindle, work the drive end first; separating bearing noise from other rear-of-motor sounds follows the same procedure as servo motor rear noise: brake or bearing.

Are the mounting bolts and adapter plate sound? #

Inspect them with the cylinder off: raised burrs or fretting around the adapter mounting holes and the tapped holes in the spindle rear, plus stretched bolts from a crash, tilt the adapter and show up directly as runout.

Do the teardown under your lockout/tagout procedure (in the US, OSHA 29 CFR 1910.147): spindle drive locked out, hydraulic pressure relieved, and the chuck supported before the draw tube comes apart.

With the cylinder off, look for:

  • Raised burrs or fretting at the adapter-to-spindle mating faces. A burr holds the adapter off the face and tilts it — pure runout, and no amount of balancing fixes it.
  • Stretched or damaged bolts. A crash that stops the spindle abruptly overstresses the bolts threaded into the spindle end; they then shift like a wheel with loose nuts, with occasional noises the shop background hides.
  • Elongated or damaged threads in the tapped holes of the spindle rear.

Decision: burrs or damaged bolts mean stoning the faces flat, fitting new bolts, torquing evenly and re-indicating — on one lathe, that single repair ended the problem. Clean faces and sound bolts mean the fault lives in the runout or the internals.

Where do I measure runout, and in what order? #

Indicate in this order: the adapter first, then the rotating front of the cylinder, then the rear sensor end, because each reading points at a different fault and the sequence keeps you from trimming a number that a burr upstream controls.

  • The adapter, bolted directly to the back of the spindle.
  • The rotating front of the cylinder, once the adapter reads true.
  • The rear of the cylinder where the position sensors sit — the mass actually whipping past the sensors, and the reading that matters most.

Where the adapter mounts with bolts passing through the spindle flange into a two-piece collar, alternating the torque on those bolts dials the runout in; the cylinder has a slight, deliberate clearance to its adapter so the same trimming is possible at the next stage.

Reading What it means Next step
Adapter runout high Adapter tilted or off-center: burr, stretched bolt, uneven torque Stone, re-seat, trim with bolt torque
Adapter true, cylinder runout high Cylinder not centered in its adapter, or a bent or worn rotating body Trim using the cylinder-to-adapter clearance; if it will not come in, check internal play
Front true, rear sensor end still high Bent or cracked draw tube, or worn cylinder bearings Go to the play and draw-tube checks

Stop when the front readings hold under 0.0005 in TIR. A rear end that stays high while the front is true is not an adjustment problem.

How much play is too much in the spindle and cylinder? #

Measure it, do not feel it: indicate spindle endplay and radial play, then cylinder endplay and radial play, and treat movement alone as acceptable unless it exceeds the maker's clearance limit or shows up as rear runout you cannot trim out.

Set a dial indicator axial and radial at the spindle nose, then axial and radial at the cylinder. Small play is built into these bearings; the condemning numbers are the spindle and cylinder clearances printed in the service documents for those exact assemblies. Two rules need no manual: radial play you can see at the rear sensor end while indicating, and rear runout that survives a true adapter and front, both point at the cylinder's own support. Excess spindle play sends the machine to spindle service. Silence proves nothing — one cylinder vibrated hard above 2300 rpm with no bearing chatter audible anywhere on it.

Can the draw tube or clamp pressure be the real cause? #

Yes, and it is the branch that is missed: a cracked or overstressed draw tube vibrates like a runout fault, and on one documented machine only the draw-tube thread failure — after every other rebuild — ended the vibration above 2300 rpm.

The sequence on that machine: leveling, feeder alignment, new motor bearings, a rebuilt hydraulic unit with a new coupling, a rebuilt actuator with new bearings and a seal kit, and position-sensor pots cleaned and adjusted twice. The vibration stayed. The machine held tolerance and finish until the draw tube pulled its threads out of the actuator, leaving remnants of the threads inside it; once the tube was out and repaired, the vibration was gone. The machine had run for a long period at a reported gauge pressure around 400 psi — a reported operating figure, not a rating.

What that case decides for you:

  • Whenever vibration survives the adapter and bearing checks, pull the draw tube and inspect its threads at the actuator end for cracks, wear and stripping before any more high-speed running.
  • Run the minimum clamp pressure that holds the work. The maximum at speed is stamped on the chuck and cylinder and given in the machine manual; that data is the limit, not habit.
  • Replace a broken pressure gauge before relying on it. Judging clamp force by collet creep only tells you grip is being lost, not what force you have.

When is the fix a rebuild instead of an adjustment? #

Rebuild when runout will not trim out at the cylinder and the play checks point at the cylinder's own bearings, or when the cylinder runs hot and loses fluid; repair in place when the fault is burrs, stretched bolts or a mis-seated adapter.

  • Burrs, fretting or stretched bolts: stone the faces, fit new bolts, re-seat and re-indicate. Done.
  • An adapter that trims in: set it under 0.0005 in TIR and confirm with the speed sweep — the 1× line at the cylinder housing should drop into the machine's background.
  • Runout that will not trim plus cylinder play: rebuild the cylinder with bearings and a seal kit together; half a rebuild repeats the teardown.
  • A cylinder that runs hot while the tank level drops: the seals are failing regardless of the vibration; rebuild before sustained running at speed.
  • Draw-tube thread damage: repair or replace the tube first. It is the failure that stops the machine mid-job.

The proof is the sweep after the repair, not the hours spent: unbalance falls with the square of speed, so a cylinder that reads true standing still and stays quiet through 3200 rpm is fixed.

Questions asked

How fast can I keep running with 0.015 in TIR at the sensor end?

Treat it as a stop-work condition for production speeds, not a tuning problem. Thirty times the runout target is already loading the rear sensors and spindle bearings. If a job must finish, stay below the speed where the shake first appeared — on the machine this fault was documented on, that meant running around 1800 rpm — and correct the runout before going back above it.

Can the cylinder be field-balanced instead of correcting the runout?

Not as the first move. Balancing adds or removes weight, but a cylinder tilted by a burr or a stretched bolt is displaced as well as unbalanced, and no added weight cures the runout itself or the sensor wear it causes. Bring adapter and cylinder runout under 0.0005 in TIR first; balance is a consideration only if the 1x vibration survives that.

Is the slight movement between cylinder and adapter a defect?

No — that small clearance is deliberate, and it is your adjustment. Loosen the adapter bolts part way, tap the cylinder with a soft-face mallet a few thousandths at a time, and re-indicate after each move; where the adapter clamps through a two-piece collar behind the spindle flange, alternating bolt torque alone trims the reading. Losing that adjustment range means worn internals, not a bad design.

The cylinder ran hot and the tank level dropped — is that separate from the vibration?

No, treat it as the same machine needing attention now. Heat with a falling tank level means hydraulic fluid is passing seals it should not, and hot, thin oil accelerates wear in the cylinder you are trying to true up. Rebuild with a seal kit, inspect the pump coupling in the hydraulic unit while you are in there, and find the leak before any sustained high-speed running.

Keep reading

  • Selection guide

    Pneumatic valve setup for double-acting robot grippers

    One 5/3 double-solenoid valve with a chosen center runs a single double-acting gripper cylinder: coil A closes, coil B opens, both off gives the idle state you ordered. Two spring-return valves on one manifold usually feed two functions — equalization, safety lock, cascaded exhaust — not one cylinder. Size the valve on peak flow: the worked example needs about 264 L/min at close.

    Pneumatics and hydraulics

  • Comparison

    Stopping a Vertical Air Cylinder on a Light Curtain Trip

    Use a 3-position center-pressurized valve with a higher regulated rod-side pressure — the only pneumatic center that balances a vertical load mid-stroke — then let a rod lock hold it. Verify every component in the trapped-air path against the deceleration pressure spike and confirm total stop time against the safety-distance formula Ds = 63 in/s × Ts.

    Pneumatics and hydraulics

  • Comparison

    Electric vs pneumatic butterfly valve actuator comparison

    Choose a modulating electric actuator when the valve must hold set positions between open and closed under a digital command and you would rather not run compressed air to it; choose pneumatic when the 90° stroke must finish in one to three seconds, when the process needs a spring-driven fail-safe action, or when the valve sits somewhere an electric drive would need certified explosion-proof construction. Size either type to exceed the valve's break torque by at least 25% at minimum supply pressure, on an ISO 5211 mounting interface.

    Pneumatics and hydraulics