By Sun Jie, Welding Positioner Lead Engineer · EVST Welding Positioner Team · · Reviewed by EVST mechanical and quality engineering
Wrong positioner and the robot can be dead-accurate yet the cell never hits cycle. Three numbers are the real gate: payload, turning torque, and indexing accuracy. EVST’s EVS-SWP single-axis family covers 200 to 2000 kg, the EVS-DWP dual-axis family goes to 5000 kg, indexing accuracy holds better than ±0.3°, and two-station layouts double arc-on time. Five layout choices — turntable, L-type, headstock-tailstock, H-type, Ferris wheel — match part type. Pair with W500–W800 travel rails on long seams as coordinated external axes. This is the buyer guide that gates the three specs that actually matter.
Key takeaways
- Three specs are the real gate: payload, turning torque, indexing accuracy.
- EVS-SWP: single-axis 200–2000 kg.
- EVS-DWP / EVS-DWP-U: dual-axis to 5000 kg.
- Indexing accuracy < ±0.3° preserves the ±0.5 mm seam-tracking margin.
- Five layouts match part type — turntable / L-type / headstock-tailstock / H-type / Ferris wheel.
- Two-station arc-on: while one side welds, the other loads — ~2× cell output.
- Pair with W500–W800 travel rails on long seams; both join the controller as coordinated ISO 9283 external axes.
This article is for welding cell designers, integrators and plant engineers picking a welding positioner for a new or retrofit cell. It covers EVST EVS-SWP / EVS-DWP positioner sizing and the five layout choices; it does not cover the robot itself, the travel rail, or weld-procedure details — those have separate guides.
Why three specs, not just payload
Most positioner spec sheets lead with payload. Payload alone is wrong because two cells with the same payload class can run wildly different throughput and seam quality. The three specs that actually gate cell performance:
- Payload decides whether the positioner physically holds the part.
- Turning torque decides how fast it indexes — and therefore the cycle time.
- Indexing accuracy decides where the seam lands — and therefore the weld quality.
Pick all three together. Skipping any one of them lets a cell pass on paper and fail in production.
When the right positioner pays back — and when it doesn’t
The decisive factors are part orientation count, weld face count and cycle gain. Use this frame:
| A positioner pays back when… | Reconsider when… |
|---|---|
| Part has 2+ welding orientations | Single-orientation flat parts |
| Manual re-fixturing currently dominates cycle | Robot welds flat, no reposition needed |
| Heat input control matters (multi-pass, thick) | Single-pass thin sheet |
| Two-station layout fits the floor | Floor too tight for two-station |
| Part family stays consistent in mounting | Mounting changes every shot |
EVST scopes the positioner with the Three-Spec Match method: payload sized to part with 30% margin, torque sized to eccentricity × weight × angular acceleration, indexing accuracy sized to seam-tracking margin — all three together, not one at a time.
Sizing payload — EVS-SWP and EVS-DWP
| Family | Axes | Payload range | Typical use |
|---|---|---|---|
| EVS-SWP | Single (rotation) | 200 / 500 / 1000 / 2000 kg classes | Flat parts, rotational seams, simple positioning |
| EVS-DWP | Dual (rotation + tilt) | 200 / 500 / 1000 / 2000 / 3000 kg | 3D parts, multi-face welding, complex orientation |
| EVS-DWP-U | Dual, U-frame structure | 3000 / 5000 kg | Heavy structural parts, large frames, ton-class castings |
Picking the class: take max part weight + fixture + 30% margin → land in the class above. For a 1500 kg part + 200 kg fixture (= 1700 kg), pick EVS-SWP-2000 or EVS-DWP-2000. The 30% margin covers dynamic loads (acceleration force) and future part variants.
Sizing turning torque — the spec most often missed
Torque scales with eccentricity × weight × angular acceleration. A 1000 kg part with 0.5 m eccentricity wanting 1 rad/s² acceleration needs 500 N·m + the static torque to hold against gravity offset. Spec sheets that quote payload but not torque are setting up a slow cell.
EVST positioner torque curves are published per axis per class. The selection rule: required torque ≤ continuous torque rating with 50% margin. If you can’t meet it, step up a class — even if payload is fine.
Sizing indexing accuracy — preserving the seam-tracking margin
The robot’s seam tracking (vision-based) typically holds ±0.5 mm at the joint. The positioner’s indexing error stacks on top: if the positioner indexes to ±1° on a 500 mm radius arc, the seam lands ~8.7 mm off where the tracking expects it — well outside the ±0.5 mm tracking window.
EVST positioners hold better than ±0.3° indexing accuracy (ISO 9283 repeatability test methodology). At 500 mm radius, that’s ~2.6 mm worst-case displacement — within typical seam-tracking compensation range. For tight-tolerance work (e.g., pressure vessel circumferential seams), spec better than ±0.1°.
Five layouts — pick by part type
| Layout | Best for | Why |
|---|---|---|
| Single-axis turntable (EVS-SWP) | Flat parts, rotational seams | One axis, simple, fast indexing |
| Dual-axis L-type (EVS-DWP) | 3D parts with multi-face welds | Rotation + tilt, every face downhill |
| Headstock-tailstock (EVS-DWP-HT) | Long shafts, pipes, tubes | Two coaxial chucks, span the part |
| H-type two-station (EVS-DWP-H) | High-throughput multi-orientation | Robot welds one side while operator loads other |
| Ferris wheel (EVS-DWP-F) | Dense small parts, multi-orientation batch | Multiple fixtures on rotating frame |
Two-station arc-on — the throughput multiplier
The single biggest cell throughput gain comes from two-station layouts. While the robot welds station A, the operator loads station B. The positioner rotates between them; arc time stays at ~80% of cycle instead of ~50%.
Cycle gain vs single-station: ~30–80% throughput lift on multi-face heavy parts. The exact gain depends on weld time vs load time ratio — when load time ≈ weld time, the gain approaches the theoretical 2×.
H-type, Ferris-wheel, and some L-type layouts all support two-station operation. EVST sizes the cell for two-station by default unless floor or safety constraints force single-station.
Pairing with a travel rail
For seams longer than the robot’s reach envelope (typically >2 m), pair the positioner with a W500–W800 travel rail. Both join the controller as coordinated external axes per ISO 9283. The positioner handles orientation; the rail handles length coverage.
Picking the rail class: match to robot payload + torch + cable package, then check against positioner static load. For a W650 rail (1000 kg payload) with an EVS-DWP-2000 positioner, the cell handles 2000 kg parts at the positioner while the robot stays within the rail’s 1000 kg envelope.
Standards the positioner runs under
- ISO 9283 — Robot performance criteria; positioner indexing repeatability tested under this standard’s methodology when serving as an external axis.
- ISO 10218 — Robot safety; positioner falls under the robot system safeguarded space.
- ISO 3834 — Welding quality; the WPS the positioner serves sits under one of the three classes.
- ISO 17663 — Heat treatment for the positioner’s structural metal parts (annealed).
- AWS D1.1 — Structural Welding Code: Steel.
FAQ
Why is turning torque a separate spec from payload? Because they answer different questions. Payload answers “can it hold the part.” Torque answers “can it move the part fast enough.” A positioner with adequate payload but inadequate torque indexes slowly and locks cell cycle time. EVST publishes both per axis.
What indexing accuracy do I actually need? Match it to your seam-tracking margin. Vision seam tracking at ±0.5 mm with a 500 mm radius part demands < ±0.3° indexing (worst-case 2.6 mm displacement). Tighter tolerance work (e.g., pressure vessels with ±0.1 mm seam) demands < ±0.1° indexing. EVST positioners cover both ranges.
Single-axis or dual-axis? Single-axis (turntable) if the part welds in one rotational plane only. Dual-axis (L-type, DWP) if any seam needs tilt to get to flat-gravity position. Most multi-face structural parts need dual-axis; most rotational symmetric parts (rings, gear blanks) need single-axis.
Headstock-tailstock vs dual-axis? Headstock-tailstock is the right answer for long axisymmetric parts (shafts, pipes, long tubes) where both ends need rotational support. Dual-axis L-type is right for 3D parts that need both rotation and tilt. They serve different geometries.
How much faster is two-station vs single-station? Typical 30–80% cell throughput lift, depending on load/weld time ratio. When load time equals weld time, the gain approaches 2×. EVST sizes two-station by default unless floor or safety constraints force single-station.
Can I add a positioner to an existing welding cell? Yes — most existing welding controllers accept external axes up to their architectural limit (typically 4–10 external axes). EVST runs a retrofit assessment that scopes axis count, safety integration and program migration before quoting.
Bringing it into your plant
Picking a welding positioner is picking three specs at once: payload, turning torque, indexing accuracy. Five layouts match part type, two-station doubles throughput, and pairing with W500–W800 rails extends reach. EVST scopes positioners with the Three-Spec Match method — payload sized for the part, torque for the cycle, indexing for the seam. See our guides to how to pick a robot travel rail, coordinated cell integration, and thick-plate multi-pass welding, or talk to EVST about sizing a positioner for your cell.
About the author — Sun Jie is the Lead Engineer of the EVST Welding Positioner Team, with 12 years of dedicated experience on welding positioner design — EVS-SWP single-axis, EVS-DWP dual-axis, EVS-DWP-U heavy U-frame and the headstock-tailstock and Ferris-wheel variants. He runs the Three-Spec Match method on every customer engagement and oversees the indexing-accuracy QC gate before shipment. Reviewed by EVST mechanical and quality engineering for technical accuracy; figures are typical achievable ranges, not guarantees, and are sized per project. Corrections and updates: see the Last Updated date.