Diagnosis and Treatment of Big Noise Problems on a Tilted by Turn CNC
A slanted-bed lathe that suddenly gets loud is telling you something. This guide walks through the six faults we see most on a tilted by turn CNC, from chuck jaws to regenerative braking, and what to do about each one. For maintenance techs and process engineers who need to decide fast whether to keep cutting or stop the spindle.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
Symptom, likely cause, and first action
Work down the column that matches what you hear. Two symptoms can share one cause, so check them together.
| Symptom | Likely cause | First action |
|---|---|---|
| Knock once per spindle rev | Loose chuck jaw or worn scroll | Stop, indicate jaws, retorque to spec |
| Rumble that rises with rpm | Spindle bearing preload lost | Log rpm vs dB, then check preload |
| Squeal only during rapid moves | Ball screw or linear guide dry | Re-lube rail, check lube line pressure |
| Buzzing on deceleration only | Regenerative braking fault | Check drive alarm history first |
| Chatter at the cut, quiet at idle | Tool overhang or weak turret clamp | Shorten overhang, reduce depth of cut |
| Whine from the hydraulic unit | Air in oil or clogged suction strainer | Bleed the circuit, clean the strainer |
What to do next
Find the operating state that produces the noise, fix the one component or parameter that matches it, and re-test at the same rpm and feed. If the noise moves instead of disappearing, you have a second fault. Send us the rpm, the alarm history, and a photo of the setup, and our engineers will tell you whether it is a machine fault or a cutting-loop problem.
How to localize noise on a tilted by turn CNC
A slanted-bed lathe gives you one advantage: gravity pulls chips and coolant away from the ways, so the bed stays cleaner than on a flat-bed machine. It also gives you one disadvantage: the same slant loads the headstock, turret, and tailstock along a different vector than a horizontal lathe, so wear shows up in places you may not expect. Before you adjust anything, separate the noise by operating state. Idle, spindle-only, axis-only, and cutting are four different tests.
Run the machine through each state for 30 seconds and listen. If the noise is there at idle with no motion, the source is usually hydraulic or pneumatic, not mechanical. If it appears only when the spindle turns, look at bearings, belts, or the chuck. If it follows the axes, look at ball screws, linear guides, and the lube system. If it only shows up under cut, the problem is in the cutting loop, not the machine.
Write down the rpm or feed rate where the noise starts. That number is the single most useful piece of data you can hand to a technician or a machine builder. A rumble that starts at 1,800 rpm and disappears at 2,400 rpm points to a bearing passing frequency. A squeal that starts at 6,000 mm/min points to a dry guide or a loose thrust bearing.
Do not keep running a machine that has developed a sudden new noise. Bearing spalling, a loose chuck, and a failing ball screw all get more expensive by the hour. Stop the cycle, finish the part in the spindle if it is safe to do so, then diagnose.
- 1Idle testAuxiliary systems only. Hydraulic, coolant, and air.
- 2Spindle-only testNo axes, no cut. Listen from 500 to max rpm.
- 3Axis-only testJog each axis at 25, 50, and 100 percent rapid.
- 4Cut testSame tool, same material, step the depth of cut.
Chuck, spindle, and turret noise
Chuck noise is the most common complaint on a tilted by turn CNC, and it is also the easiest to rule out. A three-jaw scroll chuck that has been tightened on the same diameter for months develops a wear pattern in the scroll. When you clamp a new diameter, one jaw sits slightly proud and the part runs eccentric. You get a knock at spindle frequency plus a finish that looks like a wave. Indicate the jaws on a known-good test bar. If runout exceeds 0.03 mm, disassemble, clean, and regrease the scroll, or replace the jaws.
Spindle bearing noise has a different character. It is a rumble or growl that rises with rpm and does not depend on the workpiece. On a slanted-bed machine the front bearing carries a combined radial and axial load because of the bed angle, so preload loss shows up earlier than on a flat-bed lathe. If the rumble starts below 1,500 rpm, the bearing is likely already damaged. If it starts above 3,000 rpm, you may be chasing a preload or lubrication issue that is still fixable.
Turret noise under cut is often blamed on the turret when the real problem is the tool holder. A 20 mm boring bar with 80 mm of overhang will chatter at any depth of cut above 0.5 mm, no matter how good the turret is. Shorten the overhang, switch to a carbide bar, or reduce the depth of cut. If the turret itself is loose, you will hear a clunk on every index. Check the turret clamp pressure and the coupling teeth.
Tailstock noise on a long shaft is a sign of a worn center or insufficient thrust. Check the center for blueing, replace it, and set tailstock pressure to the low end of the recommended range for the shaft diameter.
- 1Chuck runout limitKeep jaw runout under 0.03 mm on a test bar.
- 2Spindle warm-upRun 10 minutes at 25 percent max rpm before heavy cuts.
- 3Tool overhangAim for 4:1 length-to-diameter or less on boring bars.
- 4Turret clampCheck clamp pressure and teeth every 2,000 hours.
Deceleration noise, drives, and parameters
Noise that appears only when the spindle decelerates is almost never mechanical. On a tilted by turn CNC the spindle and axis drives share a DC bus, and during braking the motor acts as a generator. If the regenerative resistor or the braking unit cannot absorb that energy, the drive faults or the bus voltage climbs and the control loop starts to oscillate. You hear it as a buzz or a growl that lasts as long as the deceleration ramp.
Pull the drive alarm history before you touch a wrench. Look for overvoltage, regenerative overload, or bus voltage high alarms. If those codes are present, the fix is electrical: check the braking resistor for an open circuit, check the resistor wiring and terminal torque, and confirm the deceleration time constant in the parameters. A deceleration time that is too short for the load inertia will trigger this every cycle.
Parameter drift is a real problem on older machines, especially after a battery replacement or a control swap. If the noise started right after a service, compare the current parameter set against a known-good backup. Spindle gain, velocity loop gain, and acceleration and deceleration time constants are the four values that matter most. Change one at a time and test after each change.
A failing encoder or a loose encoder coupling can also produce noise. The symptom is a growl that changes with speed and often throws a position deviation alarm. Check the coupling runout and the cable shield ground.
- 1Check firstDrive alarm history and bus voltage trend.
- 2Resistor testMeasure resistance against the drive manual value.
- 3Ramp timeLengthen deceleration if the load inertia changed.
- 4EncoderInspect coupling runout and shield grounding.
Pump whine, lube faults, and air in the oil
Hydraulic noise on a CNC lathe is usually a whine or a pulsation that is present whenever the pump runs. The two most common causes are air in the oil and a clogged suction strainer. Air gets in through a loose suction fitting or a low oil level. The pump cavitates, the pressure gauge needle vibrates, and the turret index gets rough. Top up the oil, tighten the suction line, and bleed the circuit at the highest point.
A clogged suction strainer raises the pressure drop on the inlet side and produces the same cavitation noise, but it will not go away after bleeding. Pull the strainer and clean it. If the oil is dark or smells burned, change the oil as well. On a slanted-bed machine the reservoir is often mounted low to keep the center of gravity down, so the suction head is shorter than on a flat-bed lathe. That makes a partially clogged strainer show up sooner.
Lube system faults produce a squeal or a dry rumble that only appears during axis motion. The lube pump may be running but not delivering oil because of a blocked metering unit or a kinked line. Check the lube pressure switch signal in the PLC, then pull one metering unit and confirm it is passing oil. A dry linear guide will fail within days once it starts squealing.
Coolant pump noise is a different problem. A noisy coolant pump is annoying but rarely affects accuracy. If the noise is coming from the coolant tank, do not spend your diagnosis budget there.
- 1Oil levelKeep above the suction inlet at all times.
- 2StrainerClean or replace every 1,000 hours.
- 3Lube pressureConfirm the switch signal in the PLC each cycle.
- 4Metering unitsPull one and confirm oil flow at the outlet.
Chatter that only appears under cut
If the machine is quiet at idle and quiet on rapid moves but screams the moment the tool enters the material, you are dealing with chatter, not a machine fault. Chatter comes from the cutting loop: tool, holder, workpiece, and fixture. On a tilted by turn CNC the most common trigger is a boring bar or a long turning tool with too much overhang. The bar deflects, the cut depth varies, and the vibration feeds itself.
The fix is mechanical. Shorten the overhang, step up to a larger bar diameter, or switch from steel to carbide. If the part allows, support the free end with the tailstock. Reducing depth of cut and increasing feed per revolution often breaks the resonance more effectively than slowing the spindle down. On a slanted-bed lathe, the bed angle helps chip evacuation but does nothing for tool stiffness.
Workpiece stiffness matters too. A thin-wall tube will chatter no matter how rigid the tool is. In that case you need a different strategy: fill the bore with a dampening medium, use a follow rest, or take two light passes instead of one heavy pass. Do not compensate by increasing the clamping force on a thin-wall part. You will distort it and the chatter will move, not stop.
Spindle speed selection is the last lever to pull. Every setup has a stability lobe. If you can, run a quick test at three speeds 10 percent apart and pick the quietest one. That single test often buys more than any hardware change.
- 1OverhangKeep boring bar length-to-diameter at 4:1 or less.
- 2Depth of cutHalve it before you change spindle speed.
- 3Thin wallsUse a follow rest or dampening, not more clamp force.
- 4Speed testTry three speeds 10 percent apart, keep the quiet one.
Step by step: diagnose and treat the noise
Follow the order. Each step narrows the fault and tells you whether to continue or stop.
- 1Stop the cycle and record the stateNote the spindle rpm, feed rate, and axis where the noise starts. Write it down or photograph the screen. This data saves an hour later.
- 2Run the four isolation testsIdle 30 s, spindle-only from 500 rpm to max, each axis at 25/50/100 percent rapid, then a stepped cut. Mark which state produces the noise.
- 3Check the drive alarm historyLook for overvoltage, regenerative overload, or position deviation codes. If present, go to the electrical branch before touching mechanical parts.
- 4Measure chuck and tool runoutIndicate a test bar in the chuck. Keep jaw runout under 0.03 mm. Check the tool holder taper for fretting or chips.
- 5Check lube and hydraulic pressureConfirm the lube pressure switch signal in the PLC. Clean the suction strainer. Bleed air from the hydraulic circuit at the highest point.
- 6Adjust one parameter or one part at a timeIf the noise started after a service, compare against a known-good parameter backup. Change one value, test, then decide.
- 7Re-test at the same conditionsUse the same rpm, feed, and tool as step one. If the noise moved but did not disappear, you have a second fault, not a failed repair.
- 8Log the result and set a recheck intervalRecord the fix, the date, and the running hours. Schedule a recheck at 500 hours for bearing or guide repairs.
Questions we get about tilted-bed lathe noise
Can I keep running the machine while I wait for parts?
It depends on the fault. A noisy coolant pump or a slightly worn chuck jaw can run for a while with monitoring. A spindle rumble, a dry linear guide, or a drive that faults on deceleration should not. Those three get worse fast and can damage the spindle or the ways.
If you must finish a job, reduce the spindle speed to below the point where the noise starts, take lighter cuts, and inspect the part finish after every few pieces.
Why does the noise only appear when the machine decelerates?
That pattern points to the drive and braking circuit, not to a mechanical part. During deceleration the motor generates energy that must go somewhere. If the braking resistor, the wiring, or the deceleration time constant is wrong, the DC bus voltage climbs and the control loop oscillates.
Check the drive alarm history for overvoltage or regenerative overload codes, then measure the braking resistor against the value in the drive manual.
How much chuck jaw runout is acceptable?
Keep it under 0.03 mm on a test bar for general turning. For finishing work where you need to hold ±0.005 mm, aim for 0.01 mm or better and warm up the spindle first.
If runout is high, clean and regrease the scroll before you replace the jaws. A worn scroll will make new jaws run out too.
The noise started right after a control battery replacement. What now?
Parameter loss is the likely cause. Compare the current parameter set against a known-good backup, focusing on spindle gain, velocity loop gain, and acceleration and deceleration time constants.
Change one value at a time and test after each change. Changing several values at once makes it impossible to know which one fixed the noise.
Is a slanted-bed lathe noisier than a flat-bed machine by design?
No. The bed angle changes how chips and coolant drain, not how the spindle or the axes sound. A well-maintained slanted-bed machine is no louder than a flat-bed one.
What does change is wear location. The bed angle loads the headstock and turret along a different vector, so preload loss and guide wear can show up earlier in those areas.
When should I stop diagnosing and call the machine builder?
Call when the noise is inside the spindle, when the drive faults on every deceleration, or when you have replaced a bearing or guide and the noise returns within 200 hours.
Bring the rpm or feed rate where the noise starts, the drive alarm history, and the date of the last repair. Those three items cut the response time a builder needs.
Parts turned on a machine that runs quiet
GreatLight machines turned and milled parts to ±0.005 mm with 100% inspection before shipment. Upload your drawing and get a quotation with free DFM analysis within 12 hours.
12-hour quote100% inspectionNo minimum order quantity
More machining notes
We publish setup notes, tooling trials and inspection data from the factory floor.
