Vertical composite grinding: how one machine finishes ID, OD and face

• 1
  What makes vertical composite grinding different from a plain grinder
• 2
  Turret columns, wheel heads and the indexing decision
• 3
  Speeds, feeds and coolant for composite grinding passes
• 4
  When vertical composite grinding is the wrong route
• 5
  Where the machine design is heading
• 6
  Vertical composite grinding compared with other finishing routes

A conventional cylindrical grinder holds a horizontal workpiece between centers. A vertical composite grinding machine flips that arrangement. The workpiece sits on a vertical axis, usually on a faceplate or a magnetic chuck, and the grinding heads sit on a column or turret that can index around it. Because the part is clamped once, every feature ground in that setup shares the same reference. Concentricity stops depending on how well two fixtures agree.

The composite part of the name means more than one grinding operation is available without re-clamping. Typical combinations are internal bore, external diameter, flat face, and conical seat. The wheel head indexes, the part stays put. Operators gain access to the grinding zone from above, so swarf falls away instead of piling against the wheel.

There is a practical limit. Vertical composite grinding suits parts that are more disc-like than shaft-like. A long slender shaft is still better on a horizontal machine, because gravity and center support work in your favor there. Once the length-to-diameter ratio climbs past about 3:1, vertical clamping starts to fight deflection.

Thermal behavior also changes. On a vertical machine the part mass sits low and the grinding zone is open, so heat leaves faster and thermal drift over a batch is smaller. That matters for bores held to tight roundness, where a few microns of spindle growth will show up in the last part of the run.

• 1
  One setup, shared datumID, OD and face come off the same reference, so runout stacks less.
• 2
  Open grinding zoneChips fall clear; coolant reaches the contact point without fighting the part.
• 3
  Best shape envelopeRings, flanges, bushings, discs. Not long shafts.

Most machines use a turret column that indexes through a fixed angle, often up to a full rotation, with grinding heads bolted to different faces. Indexing accuracy is the number that decides whether the machine is worth its floor space. If the turret repeats to a couple of arc-seconds, ID and OD stay concentric. If it drifts, you have bought an expensive way to move a wheel.

Wheel selection follows the same logic as any grinder, but the vertical layout changes the dressing routine. Because the heads are stacked, a small wheel diameter for internal work and a larger one for external work usually live on the same column. Dressing both on the machine keeps their relative position known. Off-machine dressing adds a setup variable you cannot inspect away.

In-process gauging is common on these machines. A measuring head touches the bore or the face, feeds a correction back to the control, and the next part compensates. This is what holds a bore size through a long run without an operator miking every piece. It also means the machine needs a stable temperature environment, because the gauge measures what the part is, not what it will be when it cools.

For very hard materials, creep-feed and peel grinding passes are used instead of many shallow sparks. A deep, slow pass puts more heat in one place but spends less time rubbing. On hardened steel above 58 HRC, that trade usually comes out ahead for cycle time, provided the coolant is aimed at the arc of contact and not at the part in general.

• 1
  Index repeatabilityAsk for the arc-second figure, not the resolution of the encoder.
• 2
  Dress on the machineKeeps both wheels in one coordinate frame.
• 3
  Gauging loopClosed-loop size correction holds a run, not just one part.

Wheel speed for vitrified bonded aluminum oxide on hardened steel usually sits between 30 and 35 m/s. For cubic boron nitride wheels on the same material, 60 to 80 m/s is normal, and the higher speed is what makes CBN economical on a production run. Below 25 m/s the wheel tends to rub rather than cut, and the bore burns before the size comes in.

Depth of cut depends on whether you are roughing or finishing. A peel grinding pass can take 1 to 3 mm of radial depth at a slow table feed, while a finishing pass is measured in microns. On a composite machine the roughing head and finishing head can be different wheels, so the same part sees both regimes without a changeover. That is the real advantage of the layout.

Coolant is not a detail here. In internal grinding the wheel contact zone is enclosed, and if coolant cannot reach it, the wheel loads and the bore tapers. High-pressure through-spindle coolant, often 20 to 70 bar, is used to break the air barrier around a fast wheel. On external passes a flood nozzle is usually enough, but it must be aimed at the contact arc, not at the top of the wheel.

Dress interval is set by size drift, not by a clock. If the control corrects more than a few microns per part over a short span, the wheel is dulling and the dress needs to come sooner. Tracking correction magnitude is a simpler signal than watching spark color or listening to the sound.

• 1
  Vitrified Al2O330-35 m/s, forgiving, good for mixed batches.
• 2
  CBN60-80 m/s, holds form longer on hardened steel.
• 3
  Through-spindle coolant20-70 bar for internal passes; flood for external.
• 4
  Dress by driftWatch correction size, not the clock.

If the part is a shaft with a bore through it, a horizontal machine with steady rests will usually hold straightness better and cost less per part. Vertical clamping of a long part invites chatter at the top end, and no amount of wheel tuning removes that. The shape decides the machine, not the other way around.

Low-volume work with loose tolerances does not need composite grinding either. If the bore and the OD only need to be within 0.05 mm of each other, a mill-turn center that turns, bores and faces in one setup will get there faster and cheaper. Grinding is for when the tolerance or the surface finish is beyond what a turning insert can hold.

Thin-wall rings are the tricky middle ground. They suit the vertical layout because the part is supported on a face, but they also deflect under clamping and under wheel pressure. Light passes, a low-clamp-force fixture and a soft wheel grade are the usual answer. If the wall is under about 2 mm, expect to spend more time on fixturing than on the grind cycle.

Finally, consider whether the geometry really needs one setup. Sometimes two simple machines and a good fixture give the same concentricity for less capital. Composite grinding earns its place when the tolerance stack across features is tight, the volume is real, and re-clamping would introduce error you cannot inspect out.

• 1
  Long shaftsHorizontal with steady rests is the better call.
• 2
  Loose toleranceMill-turn in one setup is faster and cheaper.
• 3
  Thin wallsFace support helps, but clamping and wheel force still bite.

The clearest trend is that composite machines are absorbing more operations. A single platform now turns, grinds and measures, which removes a handling step and a fixture. For a shop making hydraulic spools or bearing rings, that is a direct cost cut, because the part never leaves the spindle reference between the turning pass and the grinding pass.

The second trend is the shift from hydraulic and manual infeed to full CNC interpolation on every axis, including the wheel head dresser. That allows wheel profiles to be generated rather than formed by a template, so a form change is a program edit instead of a new crush roll. Small batches of complex profiles become practical.

A third trend is measurement integration. Gauging heads that correct size in real time are moving from a premium option to a baseline expectation. The engineering meaning is simple: the machine holds a distribution, not just a part. That is a different quality claim, and it is the one that matters when a bore is specified at a tolerance band of 10 μm.

None of this changes the physics. A vertical composite grinder still removes material with an abrasive wheel, still generates heat at the contact arc, and still needs a rigid setup. The gains come from fewer setups and faster feedback, not from a new cutting mechanism.

• 1
  More ops per platformTurn, grind and measure without re-clamping.
• 2
  Profile generationDresser moves under CNC; form changes are program edits.
• 3
  Closed-loop sizeThe machine holds a distribution, not one part.