Uneven Grinding Chatter Marks: Why They Appear on Only Part of the Surface
Most chatter troubleshooting assumes the marks are spread evenly around the workpiece. So the operator balances the wheel, checks the spindle, adjusts parameters — and eventually gives up. If you are looking at marks that cover only half the circumference, or only one band along the length, that advice will not get you far.
Uneven chatter is a different family of problems. The pattern itself is diagnostic gold. This guide walks through how to read it: classify the distribution, calculate the frequency, and run two field tests that cost nothing and narrow the search dramatically.
Step One: Classify the Pattern Before Touching Anything
Turn the part slowly under good light and answer two questions. First, is the marking circumferential or axial — does it follow the rotation of the part, or does it sit along the length? Second, is it uniform or uneven?
Uniform marks — evenly spaced straight-wave patterns, sometimes called lobing or chatter lines, running the full circumference — usually come from a source that acts continuously: wheel imbalance, spindle bearing clearance, or parameters that have drifted into a resonance zone.
Uneven marks behave differently. There are three common shapes: marks on only part of the circumference (the classic half-marked part); marks all around but with clearly uneven depth or spacing; and marks that appear on only one section of the length. The logic behind all three is the same: the grinding condition changes during one workpiece revolution, or changes from one section of the stroke to another. The error fires only at certain angles or certain zones, so the marks appear only there.
| Defect | Appearance | Distribution | Cause Direction |
|---|---|---|---|
| Uniform chatter (lobing) | Evenly spaced lines, full circumference, depth usually under 0.5 µm | Along the axis, full circle | Wheel imbalance, spindle clearance, parameter resonance, regenerative chatter |
| Half / partial circle marks | Marks on part of the circumference, rest clean | Along the axis, partial circle | Center hole defects, centers, workpiece imbalance, uneven stock |
| Axial segment marks | Marks on one section along the length, rest normal | Segmented along the axis | Workpiece rigidity (slender shaft), local wheel dulling, guideways / hydraulics |
| Spiral marks | Continuous spiral, spacing about equal to feed per revolution | Helical | Wheel face not straight, uneven micro-edge heights, one-side contact, excessive feed |
| Fish-scale pattern | Rough, scaly surface | Local patches | Dressing not sharp, wheel dull, oil contamination on face |
| Scratches / tearing | Random, no periodicity | Random | Swarf, dirty coolant, loose grains |
| Burn | Spiral or spot discoloration | Heat-related | Excessive temperature, poor cooling, too-hard wheel, excessive depth |
Entry point: First confirm it is chatter. Random scratches are a tearing problem — go to coolant filtration and cleaning. Only periodic patterns enter the chatter analysis framework.
Expert Tip: Direction first, uniformity second. Getting the direction wrong wastes the entire investigation. Rotate the part through a full revolution by hand and mark where the pattern starts and stops before you touch a single dial.
Circumferential Unevenness: The Error Is Locked to Workpiece Rotation
Marks that follow the circle and cover only part of it tell you the error is synchronized with workpiece speed. Your first search area is the workholding side — not the wheel. This is the counterintuitive part, because most people reflexively blame the wheel.
The Number One Suspect: The Center Hole
An out-of-round, burred, dirty, or badly contacting center hole lets the workpiece axis wander as it rotates. Once per revolution, at the same phase angle, the part kicks outward, the wheel bites deeper, and the mark lands in that zone. The fix is to re-lap the center holes. And a simple habit pays for itself many times over: wipe the center holes clean before every setup.
Second: The Centers Themselves
A worn center, a taper shank that does not seat properly in the spindle socket, or tailstock pressure that is too tight or too loose all make the part rotate unstably. Too tight and the part turns unevenly and can be pushed into a bend; too loose and the system loses rigidity and starts to wander at low frequency. Aim for smooth rotation with zero axial float.
Third: The Part Itself
Eccentric geometry or uneven wall thickness in ring-type parts creates a periodic centrifugal force as it spins. Uneven stock around the circumference means the depth of cut swings from deep to shallow within a single revolution. Countermeasures: static balancing before grinding, and splitting the operation into rough, semi-finish, and finish stages with spark-out passes so the stock evens out gradually.
Fourth: The Drive
A single-point drive dog pushes the part at one spot, and the rotation becomes slightly uneven within each revolution. Uneven chuck jaw pressure does the same thing on chuck work. Both are common contributors to marks that show up on only part of the circle.
| Check Item | Common Defect | Reference Value | Action |
|---|---|---|---|
| Center hole roundness | Oval, polygonal, poor taper | Under 0.005 mm | Re-lap; grind if necessary |
| Center hole surface | Burrs, dents, dirt, dried grease | Ra ≤ 0.8 µm | Clean, apply fresh grease, wipe before setup |
| Center hole coaxiality | Two ends not on the same axis | Per part precision requirement | Re-machine or re-lap |
| Center runout | Wear, taper scoring | About 0.003 mm or less | Replace or repair center |
| Taper shank fit | Insufficient contact with socket | Contact rate ≥ 80% | Clean mating surfaces; repair socket if needed |
| Tailstock pressure | Too tight or too loose | Smooth rotation, no axial float | Readjust tailstock |
Expert Tip: Before every setup, wipe the center holes with a lint-free cloth and reapply fresh grease. This habit costs thirty seconds and eliminates an entire class of partial-circle marks. It is the cheapest fix in cylindrical grinding.
Axial Unevenness: Something Changes Along the Stroke
If the marks are distributed along the length instead of around it, shift your search to a completely different set of suspects.
Marks in the middle with clean ends is the classic symptom of a slender workpiece. Rigidity is lowest at mid-span; grinding force deflects the part, it springs back, and the vibration concentrates in the middle band. The remedy is steady rests, a lighter depth of cut, and a lower workpiece speed. Length-to-diameter ratios above 10 carry high risk and generally require a steady rest.
Marks near one end, or worse near the reversal positions, point to the table and hydraulics. An oil film that is too thick on the ways lets the table float and sway. Air trapped in the hydraulic system causes stick-slip crawling at low speed. Reversal shocks stamp themselves onto the corresponding section of the part. Lower the way-lube pressure, bleed the hydraulics, and check the reversal damping.
One band along the length suddenly turning bad while the rest stays acceptable suggests a local wheel problem: a patch of dull or loaded abrasive, or a wheel face that has lost its straight form after grinding through a hard spot. Dress the wheel, keep the profile true, and break the corners at both edges so they do not plow.
Headstock-to-tailstock stiffness mismatch or thermal deformation can also tilt the workpiece axis under grinding force, so the contact state between wheel and workpiece changes along the stroke. Check headstock and tailstock height equality and rigidity, and for precision work, run the machine to thermal equilibrium before starting.
Expert Tip: Wide wheels that are not corner-broken can leave a spiral line — the field calls it “carrying the tool.” Breaking both edges of the wheel with a small chamfer or shoulder removes the sharp corner that generates the periodic impact.
Three Cheap Field Tests That Pinpoint the Source
Test One: Count the Waves
Count how many waves sit around the circumference of the part, then multiply by the workpiece speed in revolutions per second. That product is your vibration frequency. If it matches the workpiece rotation frequency or one of its multiples, hunt on the workholding side. If it matches the wheel rotation frequency, hunt on the wheel side — balance, dressing, or the mounting flange.
Vibration frequency = Number of waves around the circumference × Workpiece speed (rev/s)
| Observation | Frequency Belongs To | Search First |
|---|---|---|
| Partial circle marks | Workpiece rotation frequency or low multiple | Center holes, centers, workpiece imbalance, uneven stock, drive dog |
| Uniform dense waves, full circle | System natural frequency (regenerative) | System stiffness, speed variation, wheel hardness and dressing |
| Wave count matches wheel rotation | Wheel rotation frequency | Wheel static/dynamic balance, flange and spindle taper fit |
| One axial segment marked | Related to stroke position | Workpiece rigidity, local wheel dulling, way oil film, hydraulic crawl |
| Pattern unchanged after speed change | Unrelated to workpiece speed | Machine structure, external source, wheel side |
Test Two: Change the Speed
Raise or drop the workpiece speed by ten to twenty percent and look again. If the pattern shifts with the speed change, the error is locked to workpiece rotation and you go back to the center holes and centers. If nothing changes, you are dealing with system chatter at a natural frequency, or something on the wheel or machine side. Ten minutes, zero cost, and it narrows the field dramatically.
Test Three: The Dressing Comparison
Dress the wheel once and watch what happens. If the marks fade but come back quickly, the wheel is dulling, loading, or the grade is too hard for the job. If dressing changes nothing, stop wasting time on the wheel and go inspect the setup and the machine.
Expert Tip: Speed variation of ±15% is the most effective emergency measure on the shop floor. It breaks the regenerative loop without any hardware change. If it does not help at all, the problem is not regenerative chatter — stop chasing that direction.
A Sensible Troubleshooting Order
Putting it all together, cheapest first. Do not start by tearing the machine apart.
| Step | Action | Problem Domain |
|---|---|---|
| 1 | Classify the pattern: circumferential or axial; uniform or uneven | Sets the search direction |
| 2 | Circumferential unevenness: clean and re-lap center holes → check centers and tailstock pressure → workpiece balance and stock → wheel last | Most partial-circle marks resolve in the first two steps |
| 3 | Axial unevenness: steady rest for slender parts → dress wheel to maintain face straightness → check way oil and hydraulic bleed → headstock/tailstock alignment | Rigidity, wheel, and stroke |
| 4 | Parameter review: reduce depth of cut and longitudinal feed; workpiece speed 1/60 to 1/100 of wheel surface speed | Reduces excitation energy |
| 5 | Deep diagnosis: speed variation test, vibration analyzer if needed | System chatter and hidden sources |
Key Reference Values
| Item | Reference Value | Note |
|---|---|---|
| Center hole roundness | Under 0.005 mm | Error transfers directly to the OD |
| Center hole roughness | Ra ≤ 0.8 µm | Grease reduces friction and runout |
| Center radial runout | About 0.003 mm or less | Replace when worn beyond limit |
| Taper contact rate | ≥ 80% | Check with marking compound |
| L/D threshold | Above 10 | Steady rest required |
| Table longitudinal speed | 0.5–3 m/min | Too high promotes spiral marks |
| Finishing depth of cut | Under 0.005 mm | Final pass especially small |
| Speed ratio | Workpiece speed 1/60 to 1/100 of wheel surface speed | Reduces self-excited vibration risk |
| Speed variation range | ±15% of spindle speed | Most effective emergency measure on site |
Expert Tip: If your part shows marks on half the circumference, go look at the center holes first. That trip usually is not wasted. It is the highest-probability, lowest-cost fix in the entire partial-circle family.
Frequently Asked Questions
Why do my chatter marks appear on only half the circumference?
✅ Partial-circle marks mean the error is synchronized with workpiece rotation and fires at the same phase angle each revolution. The most common source is a center hole defect — out-of-round, burred, or dirty. Check the center holes first, not the wheel.
How do I calculate chatter frequency on the shop floor?
✅ Count the number of waves around the workpiece circumference, then multiply by the workpiece speed in revolutions per second. If the result matches the workpiece rotation frequency, search the workholding side. If it matches the wheel rotation frequency, search the wheel side.
What if the marks are in the middle of a slender shaft, with clean ends?
✅ That is the classic signature of workpiece bending vibration. Rigidity is lowest at mid-span, so grinding force deflects the part and vibration concentrates there. Add a steady rest, reduce depth of cut, and lower the workpiece speed. Length-to-diameter ratios above 10 generally require a steady rest.
What is the fastest field test to narrow the source?
✅ Change the workpiece speed by 10–20% and look again. If the pattern shifts, the error is locked to workpiece rotation — go to the center holes and centers. If nothing changes, it is system chatter or a wheel/machine-side source. Ten minutes, zero cost.
When should I stop looking at the wheel?
✅ Run the dressing comparison test. Dress the wheel once and observe. If the marks fade but come back quickly, the wheel is dulling or loading. If dressing changes nothing at all, stop wasting time on the wheel — the problem is in the setup or the machine.
Key Takeaways
- Uneven chatter is a different family from uniform chatter. Classify the pattern first.
- Circumferential unevenness points to the workholding side, not the wheel — center holes first.
- An out-of-round, dirty, or burred center hole makes the axis wander once per revolution. Re-lap it and wipe it clean every setup.
- Axial unevenness points to workpiece rigidity, wheel face condition, or guideway and hydraulic state.
- Count the waves and multiply by workpiece rev/s to get the vibration frequency. Match it to a source.
- A speed change test is the fastest way to separate workpiece-synchronized errors from system chatter.
- A dressing comparison test tells you in one pass whether the wheel is even worth investigating.
- Key references: center hole roundness under 0.005 mm, center runout about 0.003 mm, taper contact ≥ 80%, L/D above 10 needs a steady rest.
- Speed variation of ±15% is the most effective emergency measure against regenerative chatter.
- Work cheapest first. Do not tear down the machine before you have wiped the center holes.
More grinding questions? Need a custom wheel solution?
Contact Us →