For spot, projection, and robotic-arc welding cells, the relay that sits between the PLC enable output and the welding contactor sees a transformer magnetising inrush of 80-200A peak for 50-100ms, which is 10-50x the steady-state coil current [S3]. A standard 10A ice-cube relay (for example, the Omron MY2N at 5A NO/3A NC, 250VAC) is not designed for that inrush, and the typical end-state is contact pitting, material transfer, and a fail-ON condition that keeps the welding gun energised [S3].
For welding duty the spec floor is a force-guided (captive-contact) safety relay built to EN 50205 with a minimum 1mm NC gap when the NO contacts are closed, an AgSnO₂ or similar high-inrush contact set, and a continuous rating that comfortably covers the contactor coil plus any auxiliary load. Where the inrush is still too aggressive for electromechanical contacts, a zero-cross or instant-on SSR (such as the Omron G3NA series, 5-60VDC load, 5-40A, 15ms ON / 30ms OFF max) mounted on an aluminium DIN rail with thermal compound keeps the junction below 90°C [S3]. See the broader safety relay reference for the terminology and standards that govern these choices.
Why standard ice-cube relays fail on welding contactors
Ice-cube and octal plug-in relays are general-purpose devices: their 5A/10A ratings assume resistive or lightly inductive loads, and they have no arc-quenching for DC contactor coils [S3]. When one of these relays is wired to a welding transformer primary contactor, three failure mechanisms stack up.
First, magnetising inrush. A spot-weld transformer draws 10-50x normal current at switch-on, so a 10A resistive rating collapses within the first half-cycle; AgNi contacts micro-weld and then tear, leaving a permanent spot weld [S3]. Second, DC arc suppression. Welding cells often use 24VDC contactor coils; without a snubber or polarity-specific arc chamber, the plasma arc between separating contacts prevents the relay from ever reaching its full open gap [S3]. Third, position sensitivity: some contactor-auxiliary relays are specified to operate only in a defined mounting orientation, and a panel-mounted retrofit in a welding cabinet can leave the contact wipe below the minimum needed for positive separation [S3].
Force-guided (captive-contact) relay as the default choice
A force-guided relay, also called a captive-contact, positively-guided, or mirror-contact relay, uses a mechanical bridge or slider so that NO and NC contact sets cannot be in the same state at the same time [S4]. The defining spec is the EN 50205 forced-guidance requirement: when the NO set is closed, the NC set must show a minimum 1mm contact gap, which gives the safety controller a hard, detectable fault if the NO side welds shut [S4].
For welding duty, look for: AgSnO₂ contact material rated for motor-start / capacitive inrush; a continuous thermal current of at least 6A NO and 4A NC; dielectric withstand of 2500VAC coil-to-contact; and a mechanical life of 1×10⁶ cycles, with electrical endurance quoted separately at the actual inrush profile you will run. The Fanhar FH12, for example, is published at 40A max switching, 7500VA max switching capacity, 2500VAC coil-to-contact dielectric, 1000MΩ insulation at 500VDC, 1×10⁶ mechanical operations, and 2×10⁴ electrical operations at 40A 250VAC, with Class F (per UL) insulation and UL/cUL, TÜV, and CQC approvals [S2]. The trade-off is electrical life: at 40A the FH12 is rated for 2×10⁴ cycles, which is fine for spot welding at 6-20 welds/min but not for a 200 welds/min production line, where an SSR pair becomes the more honest choice [S2][S3].
Sizing the contact rating against welding inrush
The sizing rule is conservative, not optimistic. Measure or look up the transformer inrush (typical range 80-200A peak, 50-100ms) and pick a relay whose published inrush rating (I²t or peak, not steady-state A) clears that with at least 2x headroom, plus a snubber of 0.1µF + 100Ω across the contacts to bleed the inductive kick [S3]. For a 24VDC control relay driving a contactor coil, also add a flyback diode or varistor across the coil; without it, the back-EMF at drop-out will eat the contact tips within weeks.
Use the matrix below to triage by transformer size and duty cycle. It is a starting envelope, not a substitute for a thermal calculation: any SSR line must still be heat-sunk to keep Tj below 90°C at Imax, and that almost always means an aluminium DIN rail plus thermal compound [S3].
Decision table: safety-relay topology vs welding duty profile
1. Spot welding, < 20 welds/min, transformer inrush 80-150A peak, 50-100ms: Force-guided relay, AgSnO₂ contacts, ≥6A NO continuous, EN 50205 compliant. Example envelope: 25A NO at 240VAC class (Omron G7J-2C1B territory) [S3].
2. Spot welding, 20-100 welds/min, inrush 150-200A peak: Force-guided relay in the main path plus an SSR pre-contactor for the high-cycle leg, or a dual-relay cross-check (K1 and K2 in series, with NC feedback to the PLC) for fault detection on every cycle [S3].
3. Robotic MIG/MAG or high-cycle projection welding, > 100 welds/min, continuous contactor chatter: Solid-state relay pair (zero-cross for AC line, instant-on for DC) on a heatsinked DIN rail, with the welding contactor's mechanical contacts only as a safety-rated e-stop isolator. SSRs are not a safety device on their own, so the e-stop path must still go through a force-guided relay to EN ISO 13849-1 PLc or higher [S3][S4].
Wiring, monitoring, and standards mapping
For a compliant welding-cell safety chain, the NC contact of the force-guided relay feeds the safety controller's feedback input, and the NO contact carries the welding contactor coil [S4]. The PLC or safety relay then runs the classic cross-check: it de-energises the relay, waits for the NC to close and the NO to open, and only enables the next weld cycle if both transitions are confirmed. The LED monitoring variant, with one LED anti-parallel to AC contacts (or a single LED to DC) and both LED states wired back to PLC inputs, is a low-cost way to get per-cycle diagnostics on a non-safety-rated relay pair [S3].
Standards to keep on the print: EN 50205 for the relay's mechanically-linked contact behaviour, EN ISO 13849-1 for the performance level of the e-stop and guard-door chain, and IEC 61508 if the welding cell sits inside a functional-safety instrumented function [S4]. For explosion-risk welding of coated components or in paint shops, the relay enclosure, not the relay itself, is what carries the ATEX/IECEx zoning, and the safety barrier and fire safety pages cover the boundary cases where a welding cell sits adjacent to a hazardous area. Welding-specific PPE, fume extraction, and arc-flash boundary sit outside the relay scope but are covered in the welding cutting tool reference, which pairs naturally with the relay spec.
Common failure modes and what to check first
If a welding cell starts dropping guns or firing uncommanded welds, walk this list before swapping the relay: (1) measure the actual transformer inrush with a clamped current probe, not the nameplate; (2) confirm the contactor coil voltage matches the relay's rated load (24VDC vs 230VAC changes the arc energy by an order of magnitude); (3) inspect for position-sensitive relay mounting that has been rotated during a cabinet clean; (4) check the snubber and flyback components, since a failed 0.1µF cap or open diode turns a healthy relay into a sacrificial one within a shift [S3].
Temperature-regulator relays are a separate, recurring failure source in resistance-welding controllers: a wrong temperature-class choice leads to contact welding in the heater circuit, short cycling, and compressor abuse, which is the same fail-ON signature but on the thermal side rather than the welding side [S1]. For a side-by-side of safety relay selection in process plants (SIL, ATEX, E-stop wiring) versus welding cells, the Safety relay selection for chemical plants: SIL, ATEX, and E-Stop wiring criteria article is the closest reference and the machine safety page covers the wider PL/SIL framework that wraps any of these choices.
Trackable signals to watch over the next two quarters: EN 50205 revision activity (forced-guidance gap and weld-detection criteria), any IEC 60947-4-1 update touching contactor-relay inrush ratings, and OEM-side migration notes from 24VDC pilot logic to safety-rated fieldbus (CIP Safety, PROFIsafe) that would let the force-guided relay move into the I/O slice rather than the panel.