Zyxel GS1900: the seven missing members, and a PoE budget that does not follow port count #1665

Merged
jaylouisw merged 1 commit from catalogue/1116-1255 into main 2026-11-16 13:08:07 +00:00
Owner

The catalogue held the Zyxel GS1900-8, -24 and -48 as three bare port lists: no source, no fabric, no PoE, no notes. Zyxel's own specification table covers ten members, so this adds the seven that were missing and re-sources the three that were there.

Sources. Datasheet Datasheet GS1900 Series dated 16/11/22 (https://static.us.zyxel.com/pdf/2024/switch/ds_GS1900-10HP.pdf) for every port count, PoE budget, switching capacity, forwarding rate, power draw, dimension and ordering/revision row; the specification table at https://www.zyxel.com/global/en/products/switch/gs1900-series/specification as the second surface for the same numbers; the GS1900 Series User's Guide V2.90 Ed2 (https://download.zyxel.com/GS1900-48HPv2/user_guide/GS1900-48HPv2_V2.90_Ed2.pdf) for the hardware comparison tables, the PoE port numbering of the 8HP and the management feature list; and https://www.zyxel.com/uk/en-gb/support/download?model=gs1900-24hpv2 for the current firmware release.

Added (7): GS1900-8HP, GS1900-10HP, GS1900-16, GS1900-24E, GS1900-24EP, GS1900-24HPv2, GS1900-48HPv2. Corrected (3): GS1900-8, GS1900-24, GS1900-48 - same ports, now with connectors, speed caps, fabric, source URL and notes. Every field carries the URL it came from in sourceUrl/poe.sourceUrl, and the notes state which document each fact came from.

What the numbers actually say.

  • Port count is not copper count. GS1900-24 and -48 are 26 and 50 total ports (24+2 SFP, 48+2 SFP); GS1900-24E is 24 with no SFP; GS1900-24EP is 24 with no SFP; GS1900-10HP is 10 = 8 RJ45 + 2 SFP.
  • PoE budget does not follow port count. 8HP and 10HP share 77 W across 8 PSE ports. 24EP has 24 ports but only 12 PSE ports against 130 W. 48HPv2 has twice the copper of 24HPv2 and the same 170 W, so the 48-port chassis powers exactly as many PDs as the 24-port one. 24HPv2 declares 24 x 30 W = 720 W of port capability against a 170 W supply: at most five ports run flat out.
  • Nobody in this line has a console. No console port, no CLI, in any of the ten: web GUI, SNMP v1/v2/v3, RMON 1/2/3/9, syslog, LLDP/LLDP-MED, 802.1x, RADIUS, TACACS+. No spec here declares cli, and that is deliberate.
  • Firmware. 2.90(ABTP.3)C0 released 2026-10-22 is current for the GS1900-24HPv2; build codes are per model (ABTP here) so the string does not transfer verbatim to the other nine, and the specs say so rather than claiming one version for the family. The hardware revision and part number of the current build are recorded per model from the datasheet's ordering table (e.g. GS1900-8HP is B2, EU0103F/US0103F, with B1 part numbers still listed).

One deliberate inference, stated in the spec. The datasheet gives the PSE group as a count - "12 x GbE PoE RJ-45 ports, 12 x GbE RJ-45 ports" for the 24EP, "24 x GbE PoE RJ-45 ports, 24 x GbE RJ-45 ports" for the 48HPv2 - and neither the datasheet text layer nor the series guide names the port numbers. So ports 1-12 and 1-24 are modelled as the PoE groups in the model list's left-to-right order, and poe.note says that this numbering is inferred while the counts, the standard and the budget are cited. Nothing else in either spec is inferred.

Gates. scripts/syntax-check.js 349 files OK; scripts/audit-port-types.js clean ("No port contradicts its declared type"); test/hardware-invariants.test.js 26 passed, 0 failed (PoE claims internally consistent, no phantom budgets); test/importer-roundtrip.test.js 26 passed; test/import-devices-capabilities.test.js 34 passed; test/poe-agreement.test.js 16 passed; test/poe-inference.test.js 17 passed; test/catalogue-hygiene.test.js 25 passed; test/doc-catalog-counts.test.js 24 checks passed after regeneration; test/faceplate-catalogue.test.js 14 passed - it now prints 2,210 models / 25,016 sockets / 867 distinct port shapes, and docs/CONNECTORS.md, README.md, docs/DEVICE_REGISTRY.md, docs/wiki/Home.md and scripts/build-wiki.js carry those regenerated numbers. The batch was imported through scripts/import-devices.js --upgrade with PF_INCOMING narrowed to this one research file, so no other vendor's pending research was picked up.

Unverifiable / not modelled. Which port numbers are the PSE group on 24EP and 48HPv2 (as above). Layer 3 is absent from the line (the GS2220 is the L3+ sibling), so no spec claims l3-forward. PoE class per port beyond 802.3at/30 W is not published per model. The GS1900-24HP (no suffix) is a different chassis and is left for its own entry.

The catalogue held the Zyxel GS1900-8, -24 and -48 as three bare port lists: no source, no fabric, no PoE, no notes. Zyxel's own specification table covers ten members, so this adds the seven that were missing and re-sources the three that were there. **Sources.** Datasheet *Datasheet GS1900 Series* dated 16/11/22 (`https://static.us.zyxel.com/pdf/2024/switch/ds_GS1900-10HP.pdf`) for every port count, PoE budget, switching capacity, forwarding rate, power draw, dimension and ordering/revision row; the specification table at `https://www.zyxel.com/global/en/products/switch/gs1900-series/specification` as the second surface for the same numbers; the GS1900 Series User's Guide V2.90 Ed2 (`https://download.zyxel.com/GS1900-48HPv2/user_guide/GS1900-48HPv2_V2.90_Ed2.pdf`) for the hardware comparison tables, the PoE port numbering of the 8HP and the management feature list; and `https://www.zyxel.com/uk/en-gb/support/download?model=gs1900-24hpv2` for the current firmware release. **Added (7):** GS1900-8HP, GS1900-10HP, GS1900-16, GS1900-24E, GS1900-24EP, GS1900-24HPv2, GS1900-48HPv2. **Corrected (3):** GS1900-8, GS1900-24, GS1900-48 - same ports, now with connectors, speed caps, fabric, source URL and notes. Every field carries the URL it came from in `sourceUrl`/`poe.sourceUrl`, and the notes state which document each fact came from. **What the numbers actually say.** - *Port count is not copper count.* GS1900-24 and -48 are 26 and 50 total ports (24+2 SFP, 48+2 SFP); GS1900-24E is 24 with **no** SFP; GS1900-24EP is 24 with no SFP; GS1900-10HP is 10 = 8 RJ45 + 2 SFP. - *PoE budget does not follow port count.* 8HP and 10HP share 77 W across 8 PSE ports. 24EP has 24 ports but only **12** PSE ports against 130 W. 48HPv2 has **twice the copper** of 24HPv2 and the **same 170 W**, so the 48-port chassis powers exactly as many PDs as the 24-port one. 24HPv2 declares 24 x 30 W = 720 W of port capability against a 170 W supply: at most five ports run flat out. - *Nobody in this line has a console.* No console port, no CLI, in any of the ten: web GUI, SNMP v1/v2/v3, RMON 1/2/3/9, syslog, LLDP/LLDP-MED, 802.1x, RADIUS, TACACS+. No spec here declares `cli`, and that is deliberate. - *Firmware.* 2.90(ABTP.3)C0 released 2026-10-22 is current for the GS1900-24HPv2; build codes are per model (ABTP here) so the string does not transfer verbatim to the other nine, and the specs say so rather than claiming one version for the family. The hardware revision and part number of the current build are recorded per model from the datasheet's ordering table (e.g. GS1900-8HP is B2, EU0103F/US0103F, with B1 part numbers still listed). **One deliberate inference, stated in the spec.** The datasheet gives the PSE group as a *count* - "12 x GbE PoE RJ-45 ports, 12 x GbE RJ-45 ports" for the 24EP, "24 x GbE PoE RJ-45 ports, 24 x GbE RJ-45 ports" for the 48HPv2 - and neither the datasheet text layer nor the series guide names the port numbers. So ports 1-12 and 1-24 are modelled as the PoE groups in the model list's left-to-right order, and `poe.note` says that this numbering is inferred while the counts, the standard and the budget are cited. Nothing else in either spec is inferred. **Gates.** `scripts/syntax-check.js` 349 files OK; `scripts/audit-port-types.js` clean ("No port contradicts its declared type"); `test/hardware-invariants.test.js` 26 passed, 0 failed (PoE claims internally consistent, no phantom budgets); `test/importer-roundtrip.test.js` 26 passed; `test/import-devices-capabilities.test.js` 34 passed; `test/poe-agreement.test.js` 16 passed; `test/poe-inference.test.js` 17 passed; `test/catalogue-hygiene.test.js` 25 passed; `test/doc-catalog-counts.test.js` 24 checks passed after regeneration; `test/faceplate-catalogue.test.js` 14 passed - it now prints 2,210 models / 25,016 sockets / 867 distinct port shapes, and `docs/CONNECTORS.md`, `README.md`, `docs/DEVICE_REGISTRY.md`, `docs/wiki/Home.md` and `scripts/build-wiki.js` carry those regenerated numbers. The batch was imported through `scripts/import-devices.js --upgrade` with `PF_INCOMING` narrowed to this one research file, so no other vendor's pending research was picked up. **Unverifiable / not modelled.** Which port numbers are the PSE group on 24EP and 48HPv2 (as above). Layer 3 is absent from the line (the GS2220 is the L3+ sibling), so no spec claims `l3-forward`. PoE class per port beyond 802.3at/30 W is not published per model. The GS1900-24HP (no suffix) is a different chassis and is left for its own entry.
Zyxel GS1900: the seven missing members, and a PoE budget that does not follow port count
All checks were successful
CI / Classify the change (pull_request) Successful in 13s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 18s
CI / Server smoke test (pull_request) Successful in 24s
CI / Docker build (pull_request) Successful in 33s
CI / Test (Node 18) (pull_request) Successful in 1m35s
CI / Test (Node 20) (pull_request) Successful in 1m37s
CI / Test (Node 22) (pull_request) Successful in 1m46s
1bfbdebd64
The catalogue had GS1900-8/24/48 as three bare port lists with no source, no
fabric and no PoE. The vendor spec table covers ten members; this adds the
seven that were missing and re-sources the three that were there.

Facts now recorded per model (datasheet dated 16/11/22, the current 2.90
series): total port count vs copper count (GS1900-24 is 26 ports, 24E is 24),
which ports carry PoE (8HP/10HP/24HPv2 all copper; 24EP 12 of 24; 48HPv2 24
of 48), 802.3at at 30 W/port, budgets 77/77/130/170/170 W, switching capacity
16-100 Gbps, forwarding rate, 525 K buffer, 8 K MAC, 9 K jumbo, max power
draw, noise, and the hardware revision + part numbers of each current build.

Two of them are worth the entry on their own: the 48HPv2 spreads the same
170 W budget over twice the copper as the 24HPv2, so it powers no more PDs;
and the 24EP has 24 ports of which only 12 are PSE ports against a 130 W
budget. No member of the line has a console port or a CLI - management is
web GUI, SNMP v1/v2/v3, RMON, syslog, LLDP, 802.1x, RADIUS, TACACS+.

Also: the series firmware release (2.90(ABTP.3)C0, 2026-10-22, per-model
build codes) and the one field the vendor does not publish - which port
numbers form the PSE group on the 24EP and 48HPv2 - stated as an inference
in the spec rather than quietly modelled.

Catalogue counts in README/CONNECTORS/registry/Home regenerated:
2,203 -> 2,210 models, 24,858 -> 25,016 sockets.
jaylouisw scheduled this pull request to auto merge when all checks succeed 2026-11-16 13:07:55 +00:00
jaylouisw force-pushed catalogue/1116-1255 from 1bfbdebd64
All checks were successful
CI / Classify the change (pull_request) Successful in 13s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 18s
CI / Server smoke test (pull_request) Successful in 24s
CI / Docker build (pull_request) Successful in 33s
CI / Test (Node 18) (pull_request) Successful in 1m35s
CI / Test (Node 20) (pull_request) Successful in 1m37s
CI / Test (Node 22) (pull_request) Successful in 1m46s
to ca94d03991
All checks were successful
CI / Classify the change (pull_request) Successful in 13s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 12s
CI / Test (Node 18) (pull_request) Has been skipped
CI / Test (Node 20) (pull_request) Has been skipped
CI / Test (Node 22) (pull_request) Has been skipped
CI / Server smoke test (pull_request) Has been skipped
CI / Docker build (pull_request) Has been skipped
2026-11-16 13:07:58 +00:00
Compare
Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
jaylouisw/packetforge!1665
No description provided.