Direct Answer

At the proposal stage you do not know how many gateways one switch cabinet needs — estimate too few and you rework, estimate too many and you waste. Estimate as "per-device data-point count × device count + 20% margin": a multi-parameter monitor is about 20–40 points per unit, so 2000 points carries about 30 multi-parameter devices or 50–80 small devices; also verify against the RS485 bus node count (≤32 nodes) and distance. The ESX-0223-GR and CW-C1/C2/C3 plan single-gateway capacity under the dual constraints of data points and device count.

A scene: a proposal needed the gateway count for a switch room — the commercial team estimated 2 gateways; the site inventory showed 45 monitors at 30 points each — 2 gateways were overloaded, requiring rework and another cabinet; estimating by "points × devices + 20%" from the start would have made 3 the right number. The one-line formula for gateway count: count points first, then count devices, and take the higher of the two constraints.

1. The Per-Device Point Table: Know "How Many Points One Device Has"

Meter (basic metering) — about 10–15 points per unit: voltage, current, power, energy, power factor and other base quantities; Multi-parameter monitor — about 20–40 points per unit: electrical parameters + switch states + temperature + harmonics, depending on configuration; Discrete (DI/DO) collection — about 8–16 points per unit: count by DI/DO channel, one channel is one point; Estimation motto — "meter 15, monitor 30, discrete by channel count" — apply the base points by device type first, then fine-tune by actual configuration.

In one sentence: points are "how many quantities each device has" — a meter about 15, a monitor about 30, discretes by channel — apply base points first, then fine-tune by configuration.

A point-table example: 10 meters plus 5 monitors in a cabinet — meters at 15 points, monitors at 30: 10×15 + 5×30 = 300 points — plus 20 discrete channels = 320 points — well within one gateway's 2000-point capacity. Point estimation is "apply the base, then fine-tune" — device type sets the base, the on-site configuration adjusts the deviation.

2. The 30-Device / 2000-Point Dual Constraint: Which Hits First

2000 points is the "point constraint" — the data-point ceiling: beyond it, cache, polling and uplink bandwidth all strain; 30 devices is the "device-count constraint" — the conventional device ceiling per gateway: too many devices lengthen polling cycles and exceed RS485 bus nodes; Which hits first — few-device-many-point scenarios hit the point constraint first (many monitors); many-device-few-point scenarios hit the device constraint first (many meters) — compute both constraints; How to judge — calculate the gateways needed by "point capacity" and by "device capacity" separately, and take the larger.

In one sentence: 2000 points and 30 devices are two independent constraints — points exceed, add gateways by points; devices exceed, add gateways by device count — the larger of the two is the real need.

A first-hit example: 60 meters at 15 points each in one cabinet: 900 points (under 2000), but 60 devices (over the 30 ceiling) — by device count you need 2 gateways; another cabinet: 20 monitors at 40 points each — 800 points (under), 20 devices (under) — 1 gateway is enough. Compute the two constraints separately — counting only points misses the device ceiling; counting only devices misses the point ceiling.

3. RS485 Bus Nodes and Distance: The Third Check

Node limit — a single RS485 bus conventionally holds ≤32 nodes (standard load) — more devices need grouping across multiple buses; Distance limit — RS485 conventionally runs 1200 m (baud-dependent) — fine inside a switch room, re-check across buildings; How to check — after the gateway count is calculated, verify "how many buses per gateway, how many nodes per bus, how far the runs are" — node overruns need more gateways or grouping; Practical reminder — "2000 points fits" does not mean "RS485 can hold it" — node count and distance are the third constraint; check it at the proposal stage.

In one sentence: after counting points and devices, verify RS485 — ≤32 nodes per bus, 1200 m — nodes over, add gateways or group; distance over, add repeaters.

A verification example: a project counted 1 gateway as sufficient by points (1200 points), but 45 devices needed hanging — a single RS485 bus exceeded 32 nodes — after review, split into 2 buses (22 each) or added a gateway by zone — reviewing at the proposal stage avoided site rework. Points fitting ≠ bus holding — the RS485 node check is the most commonly missed step at the proposal stage.

4. Why the 20% Margin: Room for the Future

Site variables — the points estimated at the proposal stage are usually too few: actual configurations are richer, devices get added later — a margin is tolerance; Polling and cache — headroom keeps polling cycles and cache relaxed: a gateway running at full load has higher latency and loss risk; Engineering convention — size gateways as "need + 20%": a 1000-point need is planned at 1200 — 20% is the industry norm; The margin's boundary — 20% is not "the more the better": 50% means buying devices you do not use — 20% is the balance of "enough and not wasteful".

In one sentence: the 20% margin is "room for site variables" — under-counted points have a fallback, polling and cache stay relaxed — 20% is convention, more is waste.

A margin example: a project estimated 900 points — planned at 900 × 1.2 = 1080 points, one gateway (2000) sufficed; later the client added 6 monitors (180 points) — 1260 total still fits. The 20% margin lets "later additions" avoid another gateway — that 20% kept at planning time is the cushion for site changes.

5. When Over the Limit: Gateway Cascading and Zoning — Three Expansion Moves

Add gateways — the simplest: each gateway manages a portion — zone by area or by device type — gateways do not interfere; Group the buses — one gateway with multiple RS485 buses: ≤32 nodes each — group buses when nodes exceed; Cascade — multiple gateways converge upstream to a switch/platform: cascading solves "one gateway cannot hold it" — note unified management on the platform side.

In one sentence: three expansion moves — add gateways by zone, group the buses, converge by cascade — a clear path from "one gateway cannot hold it" to "many gateways working together".

An expansion example: 120 devices and 3500 points on a campus — one gateway cannot hold it: split by zone into 2 gateways (Zone A: 60 devices, 1800 points; Zone B: 60 devices, 1700 points) — each grouped under 32 nodes per bus, converging upstream for unified platform management. Over the limit without rework: zone + group + converge — capacity doubles without redesign.

6. Common Misconceptions: The Three Most Frequent Gateway-Count Errors

Misconception one: count points only, not devices — points fit, so it feels sufficient, while the device count actually exceeded — only half of the dual constraint computed; Misconception two: count capacity only, skip the bus check — 2000 points fit, but RS485 cannot hold 32+ nodes — capacity passes, the bus fails; Misconception three: treat the margin as optional — plan exactly at the need, then one added device means rework — the margin is not optional, it is mandatory.

In one sentence: three misconceptions map to three constraints — points, device count and bus check; skip one and you rework once more.

A misconception example: a project planned 1 gateway at 1500 points — but 45 devices exceeded the 32-node single bus — at acceptance, polling was slow and the bus unstable — the review: the device-count constraint and bus check were skipped. The three constraints for gateway estimation are not a multiple-choice — they are all mandatory: points, device count and bus.

7. Back to FEXLINK: How Point Estimation Lands

To land the capacity logic on products: the ESX-0223-GR and CW-C1/C2/C3 plan single-gateway conventional capacity at 30 devices / 2000 points, supporting multiple RS485 bus groups and cascade expansion — at the proposal stage, size the gateway count in three steps: "points × devices + 20% margin + RS485 check". Three steps for the gateway count: ① count points (apply the base by device type) → ② count devices (take the higher dual constraint) → ③ verify RS485 (nodes + distance) — three steps done, the gateway count is unique.

One summary: single-gateway capacity = point constraint + device-count constraint + RS485 check — estimate as "per-device points × device count + 20% margin", take the higher of the two constraints, verify the bus nodes — under-estimating reworks and over-estimating wastes, both from a missing constraint. Apply the base points, compute the dual constraints, verify the bus — three steps, and the gateway count cannot go wrong.

One reminder: the estimation formula is the "map at the proposal stage" — the site follows the device list: point configurations, bus topology and later expansion are all verified on site — estimate by formula at the proposal stage, verify by list at the deployment stage. Estimation is the start, not the end — formula for the proposal, list for the site; two rounds of cross-checking make the gateway count accurate.

Scope and Limitations

  • The conclusions in this article are intended for LV distribution and electrical-safety sites. Criteria are compiled from public codes and FEXLINK product documentation, and product parameters follow the latest selection handbook and product knowledge base.
  • Quantitative criteria (protection zone, discharge and residual voltage, earth resistance, thresholds and grading) hold under the preconditions stated in this article. Where the site voltage level, earthing arrangement, environmental category or load profile differs, the figures must be recalculated. Specific configurations involving ESX-0223-GR、CW-C1/C2/C3 should be verified against actual circuit parameters.
  • Specific product models and parameters are subject to the latest selection handbook and product knowledge base; this article does not constitute selection or procurement advice.
  • Standard numbers cited here refer to the currently effective editions; when a standard is revised, the related quantitative conclusions must be re-checked.
  • Deployment and implementation for specific projects must be carried out by qualified engineers based on site survey and design drawings.

Frequently Asked Questions

30 Devices, 2000 Points per Gateway

At the proposal stage you do not know how many gateways one switch cabinet needs — estimate too few and you rework, estimate too many and you waste