CNC Following Excentric Trode Grinding: How the Wheel Keeps an Eccentric Form

• 1
  What "Following" Means on a CNC Following Excentric Trode Grinder
• 2
  Machine Elements That Decide the Result
• 3
  Where the Form Goes Out of Tolerance
• 4
  Which Parts Suit This Process, and Which Do Not
• 5
  Setting Up a Following Eccentric Grind in Practice
• 6
  Following Eccentric Grinding vs Other Routes

An eccentric trode is a form that is not concentric with the axis it turns on. The material to be removed is deeper on one side than the other, and the depth changes as the part rotates. A fixed wheel position cannot hold that shape. The grinder has to move the wheel while the part turns, and the two motions have to stay locked in phase.

That linked motion is what "following" means. The CNC reads the programmed eccentric profile, feeds the rotary encoder value, and drives the axis so the wheel stays a fixed distance from the surface. The contact point travels around the part, and the depth of cut stays within the allowance left by the previous operation.

The high performance digital control system is what makes the correction fast enough to matter. Older mechanical copying used a template and a stylus, so any template wear showed up in the part. A digital loop reads position, compares it to the profile, and corrects on the next encoder count.

Two things have to be true for this to work. The eccentric profile has to be defined as a function of rotation angle, and the wheel has to be dressed to a known form. If either is vague, the control loop has nothing solid to follow.

• 1
  Profile-driven pathWheel position is a function of part rotation angle, not a fixed offset.
• 2
  In-phase motionRotary and linear axes stay synchronized through the whole revolution.
• 3
  Template-freeNo stylus or master cam to wear out and drift.

The rotary table is the reference for everything else. On our grinding and machining cells we use a Ø400 mm rotary table, and the eccentric throw is usually small relative to that diameter. A larger table does not automatically help; what matters is encoder resolution and how much the table deflects under load.

Wheel head stiffness sets the practical depth of cut. A light pass of 0.005–0.020 mm per revolution holds form well. Push to 0.05 mm per pass on an interrupted eccentric surface and the wheel starts to bounce, which shows up as chatter marks on the flank.

Coolant delivery is often the weak link. Grinding an eccentric form means the contact zone moves around the part, so a fixed nozzle can starve the cut on one side. We aim coolant at the contact point and check for burn discoloration at the trailing edge after the first article.

For the parts we machine alongside grinding, tolerances run to ±0.005 mm and fine surfaces to Ra 0.2–0.8 μm. Those numbers are a reasonable target for a following eccentric grind, provided the setup above is respected.

• 1
  Rotary tableEncoder resolution and radial stiffness matter more than table size.
• 2
  Wheel headLight passes keep form; heavy passes invite chatter.
• 3
  CoolantDirect the stream at the moving contact zone, not at the part center.

Most form error traces back to three places: the profile definition, the wheel condition, and thermal drift. We check them in that order because the first two are cheap to fix and the third takes time.

Profile definition errors are the easiest to miss. If the CAM output describes the eccentric curve with too few points, the control interpolates straight segments between them. On a large throw this reads as flat spots. A point spacing of 0.5° or finer is a safe starting value for most eccentric journals.

Wheel wear is not uniform on an eccentric grind. The side that removes more material loads up faster and glazes. Dressing on a fixed interval, not a fixed part count, keeps the form stable across a batch. We log dress cycles against measured form so the interval can be adjusted with data.

Thermal drift shows up late in a run. The part grows as it heats, the wheel head grows too, and the two do not grow at the same rate. Letting the machine idle to temperature before the first article, then checking the last part of the batch against the first, catches most of it.

• 1
  Curve resolution0.5° point spacing or finer on the eccentric profile.
• 2
  Dress intervalTime-based, not part-count-based, on uneven stock removal.
• 3
  Warm-upIdle to temperature, then compare first and last article.

Following eccentric grinding earns its keep on hardened parts with a modest throw. Camshaft lobes, eccentric journals, pump rotors and offset pins are typical. The material is usually above 45 HRC, which is why turning or milling cannot finish the form directly.

Throw size is the first filter. A small throw, say under 5 mm on a 40 mm journal, follows cleanly because the wheel head travel stays short and the contact geometry does not change much. A large throw on a slender part deflects the workpiece, and the control cannot correct for a deflection it cannot measure.

Aspect ratio is the second filter. If the eccentric section is long and thin, the part bends away from the wheel and the measured form goes barrel-shaped. Supporting it with a steady rest helps, but the rest itself becomes a source of runout.

The process is a poor fit for very deep eccentric reliefs, for parts that are not hardened, and for one-off geometry where grinding setup time exceeds the value of the part. In those cases, 5-axis milling or wire EDM usually wins on total cost.

• 1
  Good fitHardened cams and journals with throw under about 5 mm.
• 2
  Poor fitLong slender eccentric sections, soft material, one-off geometry.
• 3
  Alternative5-axis milling for 3D forms, wire EDM for flat eccentric profiles.

Start from the drawing, not from the last program. The eccentric profile should be defined as a polar function with the throw, the base circle and the phase angle all stated. If the drawing only gives a chord length and a height, convert it before programming, not at the machine.

Dress the wheel to the required form and record the dress. Measure the first article on a profile projector or a CMM with the eccentric datum picked up, not on a simple micrometer. A micrometer over an eccentric form reads the chord, not the true radius, and that difference is larger than the tolerance.

Run the first article in three passes: rough at 0.02–0.03 mm per revolution, semi-finish at 0.01 mm, and finish at 0.002–0.005 mm with a freshly dressed wheel. Measure between passes so drift is caught before the finish pass, not after.

Once the setup is proven, hold it. Change one variable per batch and log the result. That log is what turns a following eccentric grind from a skill into a controlled process.

• 1
  Define firstConvert the eccentric profile to polar coordinates before programming.
• 2
  Measure rightUse a CMM or projector with the eccentric datum, not a micrometer.
• 3
  Three passesRough 0.02–0.03 mm, semi-finish 0.01 mm, finish 0.002–0.005 mm.