Time Servro System: How Tool Supply Shapes Accuracy on Large Machine Tools

• 1
  Key takeaways
• 2
  What a time servro system actually controls
• 3
  Where the loop loses accuracy on large machines
• 4
  Which machine tools benefit and which do not
• 5
  Matching the drive to the axis
• 6
  Control system fit and installation notes
• 7
  Drive choice by tool supply application

A time servro system is a closed-loop drive that regulates three things at once: motor speed, shaft position and output torque. Older tool supply drives handled speed only. The operator turned a potentiometer, the motor ran at some nominal rpm, and the tool arrived wherever the mechanics put it. That works while the load is constant.

Tool supply is not a constant load. A gantry mill may pull a 40 kg head assembly out of the rack, carry it 2 m, then seat it against a spindle taper. Load changes at every stage. A speed-only loop overshoots on the light leg and stalls on the heavy one. The time servro system reads the actual position and corrects within the same move.

The control board in most of these drives is a 32-bit microprocessor running a current loop at a few kHz and a position loop an order of magnitude slower. The split matters. Fast inner loops keep torque smooth; the slower outer loop decides where the axis stops. When someone says the drive is accurate, they usually mean the outer loop, and that is the number worth checking on a datasheet.

Torque control is the part engineers overlook. On a tool supply axis, torque is what tells the controller the tool has seated. Current rises sharply when the taper bottoms out. The drive can stop on that rise instead of on a limit switch, which removes a mechanical failure point.

• 1
  Speed loopKeeps the motor turning at the commanded rpm under changing load.
• 2
  Position loopCompares encoder counts against the target and issues the correction.
• 3
  Torque limitDetects seating, jamming and tool-present conditions without extra sensors.

The drive can only correct what the feedback device reports. On a large machine tool the encoder is usually mounted on the motor shaft, not on the tool carriage. Everything between them, including the ball screw, the coupling, the gearbox and the linear guide, sits outside the loop. Backlash in that chain is invisible to the controller.

Thermal growth is the second error source. A 4,000 mm bed can move several hundredths of a millimetre as the machine warms through a shift. A time servro system cannot measure that unless the machine has linear scales. With scales, the loop reads real carriage position and compensates. Without them, the drive holds motor position perfectly and the tool still drifts.

Rack and pinion tool supply axes add a third problem. Pinion teeth wear unevenly, so the effective ratio changes along the travel. The controller sees a smooth motor rotation and a jumpy carriage. Regular backlash measurement catches this before it becomes a scrapped part.

None of these are drive faults. They are machine faults that a good drive exposes. If you are chasing tenths on a long axis, mount the feedback where the cutting happens, or accept that the servo is only as good as the mechanics behind it.

• 1
  Motor-mounted encoderBlind to screw, coupling and guide wear.
• 2
  Linear scaleReads true carriage position; costs more and needs protection from chips.
• 3
  Thermal driftRoughly 0.01–0.03 mm per metre of steel over a 10 °C rise.

Gantry grinders, gantry mills and large vertical boring machines are the natural home for a time servro system. These machines move heavy tool assemblies over long distances, and the tool supply function feeds directly into the first cut. A 0.02 mm seating error shows up as a taper or a step on the workpiece.

Machines with short travel and fixed tool stations gain much less. If the tool moves 200 mm and the load never changes, a basic speed-controlled drive with a hard stop will hold the same tolerance for a fraction of the cost. Adding a time servro system there is money spent on capability nobody uses.

Multi-station tool magazines on machining centres sit in the middle. The magazine itself is usually fine with an indexing drive. The spindle-side clamping is where the accuracy lives, so that is where the position feedback belongs.

A practical test: measure the spread of tool seating position over 50 changes with a dial indicator. If the spread is under 0.01 mm, the existing drive is doing its job. If it drifts past 0.03 mm, the loop is the first thing to examine, ahead of the mechanical alignment.

• 1
  Good fitGantry mills, gantry grinders, long-travel boring machines, heavy head changers.
• 2
  Poor fitShort-travel mills, fixed stations, light tooling with repeatable hard stops.
• 3
  Measure first50-change seating spread tells you more than any datasheet.

Sizing starts with inertia, not with torque. The drive has to accelerate the motor rotor plus everything the tool supply axis carries, reflected through the gear ratio. A head assembly that weighs 40 kg through a 10:1 reducer looks like 0.4 kg at the motor, but the reflected inertia goes with the square of the ratio. Get this wrong and the loop becomes unstable.

Next comes the encoder. For a target of ±0.005 mm on the carriage, work backwards through the reduction to find the required counts per motor revolution. A 2,500 line encoder with four-fold quadrature gives 10,000 counts per rev. Through a 10:1 reducer that is 1,000 counts per mm, or 0.001 mm per count. Enough headroom for a ±0.005 mm target.

Bandwidth decides how fast the loop settles after a disturbance. Tool seating is a step load. If the drive bandwidth is under 50 Hz, expect visible overshoot and a longer seating time. Most modern drives run 100–200 Hz on this kind of axis, which is plenty when the mechanics are stiff.

Finally, check the tuning interface. A drive that only offers two potentiometers will be hard to tune on a large machine with changing load. Look for software tuning with a step-response display, so the setup can be verified rather than guessed.

• 1
  Inertia ratioKeep reflected load inertia under 10× motor inertia where possible.
• 2
  ResolutionAim for at least 10 counts per 0.01 mm of carriage travel.
• 3
  Bandwidth100–200 Hz suits most heavy tool supply axes.

A time servro system has to talk to the machine control. Most drives accept a ±10 V analog speed or torque command, plus a few digital I/O lines for enable, fault and in-position. That interface works with almost any CNC, including older controls that have no fieldbus.

Pulse-and-direction is the other common option. It suits controls that already output step signals and keeps wiring simple, but it limits how much diagnostic data the drive can return. If you want to log following error over a shift, choose a drive with a serial or Ethernet link.

Wiring practice matters more than people expect on a long axis. Route encoder cable away from motor power cable, use shielded twisted pairs, and ground the shield at one end only. A noisy feedback signal produces an error that looks exactly like mechanical backlash, and engineers have chased that ghost for days.

Commissioning is four steps: verify encoder direction, tune the current loop, tune the velocity loop, then tune position. Changing the order creates oscillation that no amount of gain adjustment fixes. Record the final gains on the machine, because the next person to open the cabinet will need them.

• 1
  Analog ±10 VWorks with nearly any CNC; needs a separate enable line.
• 2
  Pulse and directionSimple wiring, limited diagnostics.
• 3
  Cable routingSeparate encoder and power runs; shield grounded at one end.