A linear motor applies force directly to the carriage, with no screw and no belt in between. This article answers one question: why can direct drive reach sub-micron, and what is the cost?
The answer: direct drive eliminates the screw’s lead error and backlash, so positioning precision is set by the optical linear encoder and rails instead — that is how it reaches sub-micron. The cost is higher price, mover mass affecting acceleration, and a magnetic field that needs shielding.
This article covers the direct-drive principle, the difference between coreless and iron-core movers, the optical linear encoder and Abbe error, and how to read the numbers of Newport’s three series.
Figure 1 | Generic coreless configuration with a fixed magnetic track and moving coils (not a specific product cross-section). The scale and readhead measure the carriage’s displacement relative to the base. Click for full size.
1. Direct-Drive Principle: Force Applied Directly to the Carriage
A linear motor is a motor unrolled along the travel: a magnet array and coils move in linear relative motion, force is proportional to current, and the carriage does not go through a screw[1]. A screw stage must convert motor rotation to linear motion through a screw, with lead error, backlash, and coupling compliance in between; a linear motor removes all those transmission elements and applies force directly to the carriage.
This is the root reason it can reach sub-micron: the precision limit is no longer the screw, but the optical linear encoder’s resolution, the rail’s straightness, and mounting and temperature. A screw stage’s precision ceiling is stuck at the screw C-grade and backlash (micron class); once a linear motor removes the screw, the ceiling moves to the optical linear encoder (nanometer-class resolution) and the rail (micron-class straightness).
Force proportional to current (F = Kf·I) also lets a linear motor accelerate without being limited by screw lead: for the same force, a screw stage must go through the screw’s speed ratio, while a linear motor acts directly on the carriage, so the acceleration can be very high. This is why it suits high-dynamics applications (fast positioning, scanning).
2. Coreless vs Iron-Core, Detent Force
Linear motor movers come in two types, differing by whether the mover contains iron:
| Coreless | Iron core | |
|---|---|---|
| Mover structure | Coils only, no iron | Coils + iron core |
| Detent force | None (smoothest) | Present (must be managed) |
| Force density | Low | High |
| Heat | Mover does not heat; coils can be fixed for cooling | Mover heats; thermal management needed |
| Typical use | High precision, short-to-mid travel | Heavy load, long travel |
Detent force is the periodic attraction that occurs as the mover passes over the stator poles, causing small vibration and positioning error during positioning. A coreless mover has no iron, so no detent force — the motion is smoothest, suitable for nanometer-class resolution; but force density is low, so the same force needs a longer magnetic track. An iron-core mover has high force density, but its detent force and attraction must be managed (e.g., by magnetic-circuit design or compensation).
A coreless mover has another advantage: the coils can be fixed to the base while only the magnets move (or vice versa), moving the heat-dissipation problem of the heating coils from the moving part to the fixed part[1].
3. Optical Linear Encoder Closed Loop + Abbe Error
A linear motor’s precision relies on an optical linear encoder closed loop: a glass scale is ruled with equally spaced lines, and a readhead reads position by interferometry or moiré, with resolution down to the nanometer range. The readhead measures the carriage’s actual displacement relative to the base, and the controller corrects accordingly — this is more direct than a screw stage’s “reading the motor-axis rotation,” because it reads the carriage itself.
Abbe error is the main error source of a linear stage: when the readhead’s measurement axis is not on the same line as the carriage’s motion axis (there is an offset), any carriage tilt amplifies the offset into a position error. For example, with the readhead offset 10 mm from the motion axis and the carriage tilted 1 arc-minute, the Abbe error is about 10 mm × tan(1′) ≈ 2.9 μm. So a high-precision stage keeps the readhead as close to the motion axis as possible (coaxial), or measures the tilt to compensate.
Direct drive removes the screw error, but rail friction, optical linear encoder error, Abbe error, mounting deformation, and temperature still each need evaluation[1] — removing the screw does not remove all error.
4. How to Read the Numbers: XMS, ILS, IMS
Newport’s three linear-motor series, guaranteed values[1][2][3]:
| Series / magnetic circuit | Travel (mm) | MIM | Bidirectional repeatability | Axial positioning accuracy |
|---|---|---|---|---|
| XMS-S / coreless | 50 / 100 / 160 | 1 nm | ±0.040 μm | ±0.75 μm |
| ILS-LM / coreless | 100 / 200 / 300 | 10 nm | ±0.30 μm | ±1.5 / ±2.0 / ±2.5 μm |
| IMS-LM / iron core | 300–600 | 20 nm | ±0.25 μm | ±4.5–7.5 μm |
Conditions cannot be omitted: the XMS 1 nm MIM is for the XPS-DRV11 drive configuration, and the manual requires a controlled environment and a 5 μm flatness mounting surface[1]; the IMS-LM 1000 mm/s is the no-load, XPS-EDBL driver condition, and with the XPS-DRV11 the max is 500 mm/s[3]. MIM (minimum increment) describes control and sensing capability, not positioning accuracy; for positioning accuracy read the “axial positioning accuracy” column and confirm the travel and driver conditions.
5. Applicable Range and Limits
Linear-motor stages suit sub-micron-class, high-speed, long-travel applications. Their precision ceiling is set by the optical linear encoder and rails, typically in the sub-micron class; for nanometer class, switch to piezo (see a later article in this series).
Common Misconceptions
- “A linear motor is always more accurate than a screw”: Wrong. A linear motor removes screw error, but rail straightness, Abbe error, and mounting flatness still limit precision; a cheap screw stage with good rails can be more accurate than a poor linear-motor stage.
- “Coreless is always better”: Coreless is smooth and cogging-free, but has low force density; heavy loads or long travel need iron core — you cannot judge by smoothness alone.
- “1 nm MIM means 1 nm platform accuracy”: Wrong. MIM is the minimum increment (control and sensing capability); platform positioning accuracy is the “axial positioning accuracy” column (XMS is ±0.75 μm).
Self-Check for the Reader
- Can you state why direct drive reaches sub-micron (which error it removes)?
- What is the trade-off between coreless and iron-core movers in detent force and force density?
- With the readhead offset 10 mm from the motion axis and the carriage tilted 1 arc-minute, what is the Abbe error?
Back to the opening question: why can direct drive reach sub-micron? The answer is that it removes the screw’s lead error and backlash, so precision is set by the optical linear encoder and rails. Applicable conditions: the numbers here come from the Newport XMS/ILS/IMS datasheets, all specific model and driver conditions. Unresolved limits: direct drive removes screw error, but rail friction, scale error, Abbe error, mounting deformation, and temperature still need evaluation; magnetic shielding and mover mass are the costs to confirm when selecting.