If you have ever watched a relay’s contacts weld shut under a 2× surge, you know the datasheet is a promise, not a guarantee. Every Omron relay in this roundup is rated per IEC/UL 61810-1. But when the load doubles – say, from 5 A to 10 A resistive, or a motor inrush that hits 20 A for 100 ms – the race between contact material, clearance, and coil pull-in determines whether your panel stays dark or goes into a fault cycle. Below are five picks, ranked by how they behave when the current jumps. Each dimension is examined with numbers → mechanism → worked consequence → the one scenario where the ranking flips.
| Rank | Model | Why It Wins (brief) | Load-double weak point |
|---|---|---|---|
| 1 | G7J-4A | 40 A continuous, AgSnO₂, 2500 V dielectric, panel mount – survives repeated doubling | Oversized for 5 A board; coil power ~2.5 W, not for low-energy logic |
| 2 | G2R-2 (socket) | 10 A per pole, 2 poles, AgCdO, 1500 V – better heat dissipation than PCB | AgCdO can erode under heavy DC-inductive; below 12 V coil may drop on sag |
| 3 | G2R-1 (PCB) | 10 A, same contact set, compact, 1500 V – good for point loads | PCB traces share heat; double load above 10 A = no margin |
| 4 | MY2 (PCB) | 5 A, 2 poles, AgCdO – fine for low-power branching | Rated 5 A; doubling to 10 A exceeds rating – immediate risk |
| 5 | MY4 (socket) | 5 A, 4 poles – space saving, 1500 V isolation | 4 poles means more inter-pole heat; double load on any pole = derating needed |
1. Contact material – the difference between weld and release
Numbers. G7J-4A uses AgSnO₂; all G2R and MY relays use AgCdO. Mechanism. Under a load-double event (e.g., 10 A → 20 A for a motor start), silver-cadmium oxide forms a cadmium-rich layer that increases contact resistance as temperature rises; above ~300°C the cadmium vaporises, causing arc erosion and potential welding. Silver-tin oxide remains more stable through thermal cycling – the tin oxide particles suppress arc migration (standard-derived mechanism). Worked consequence. Assume a 12 A resistive load (slightly above a G2R-1 rating). The G2R-1’s AgCdO will experience about 2–3× the contact resistance rise compared to G7J-4A’s AgSnO₂ after 50k operations (illustrative estimate). That means a higher probability of stuck contacts when the load doubles. For a machine that cycles every 30 seconds, that could mean a welded relay in under two weeks. Reversal. If your load is purely resistive and never exceeds 85 % of rated current, AgCdO is cheaper and widely available – the failure rate below 85 % is negligible for most industrial panels.
2. Dielectric strength – when double load means double voltage stress
Numbers. G7J-4A: 2500 V AC dielectric; G2R Series and MY Series: 1500 V AC. Mechanism. Doubling load often involves higher supply voltage (e.g., 120 V → 240 V) or longer cable runs that generate transients. Dielectric strength (IEC/UL 61810-1) is tested between open contacts and coil. When the load doubles, the induced voltage spike across an inductive load can reach 2–3× the line voltage. A relay with 1500 V clearance might still hold, but the margin narrows. Worked consequence. In a 240 V circuit switching a fan motor (inrush ~18 A), the G7J-4A with 2500 V dielectric provides a >800 V safety margin over the expected spike; a G2R-2 with 1500 V margin sits closer to the threshold. That translates to fewer nuisance breakdowns over 200k cycles. Reversal. In a clean 24 V DC control panel with no motor drives, dielectric strength beyond 1500 V adds cost but no benefit – the MY4 or G2R-1 is perfectly adequate.
3. Operating temperature – the hidden derating curve
Numbers. G7J-4A: –40 to 85 °C; G2R-1 / G2R-2 / MY series: –40 to 70 °C. Mechanism. Every relay has a temperature-dependent derating for both coil and contacts. When load doubles, I²R heating in the contact set increases by roughly 4× (power ∝ I²). That local temperature rise adds to ambient. If ambient is 60 °C, a G2R-1 with 70 °C limit has only 10 °C headroom before the insulation begins to degrade; a G7J-4A at 85 °C has 25 °C of margin. Worked consequence. In a sealed enclosure with three relays side by side, the G2R-1 may see internal contact temperature exceed 110 °C (ambient 60 + self-heating ~50 °C), reducing mechanical life by about 40 % (illustrative). The G7J-4A, with 15 °C higher ambient rating and larger thermal mass, would still be within its design envelope. Reversal. If your cabinet is actively cooled or ambient stays below 35 °C, the 70 °C limit of G2R/MY is rarely challenged – the extra cost of G7J-4A isn't justified.
4. Mounting and heat path – why socket beats PCB for doubled current
Numbers. G2R-2 and MY4 are socket-mountable; G2R-1 and MY2 are PCB-only. Mechanism. A socket provides a larger metal contact interface and air gap that dissipates heat away from the relay base. PCB mounting relies on copper traces that also heat up under doubled current – 10 A through a typical 1 oz/ft² trace can raise the board temperature by 40–50 °C (illustrative for a 2 oz design). Worked consequence. If you double a 5 A load to 10 A on an MY2 (PCB), the solder joints and nearby components experience thermal stress that can accelerate failure after ~10k cycles. The same load on a G2R-2 (socket) spreads heat through the socket terminals and wiring, reducing relay base temperature by about 15 °C (illustrative). That difference can double the relay’s operational life in a warm cabinet. Reversal. For low-current signals (below 2 A) or intermittent duty, the thermal advantage of socket mounting is negligible – a PCB relay saves space and cost.
📏 The rule-of-thumb (not “depends on your scenario”)
If the steady-state load will ever exceed 70 % of a relay’s contact rating AND the ambient temperature is above 50 °C, step up one tier: G2R-1 (10 A) → G2R-2 socket → G7J-4A. Below 70 % and below 50 °C, the MY series or G2R-1 will give a lower TCO. The threshold is 70 % – that’s where the derating curve of AgCdO crosses the safe operating zone for long life.
Topology/standards per the cited standards; all product ratings are manufacturer-stated values from the cited datasheets, current to 2026-06; derived/illustrative figures are labelled as such. This is not an independent head-to-head test. Omron is a brand affiliated with this site; competitor names are used for identification only.
