devices(powerline): all ten PoE injectors now state the total they can supply #2413

Merged
jaylouisw merged 1 commit from catalogue/1121-1122 into main 2026-11-21 11:39:26 +00:00
Owner

Refs #130 (does not close it - the UDW-PRO is still unpublished).

The defect

node scripts/audit-poe-budget.js names 411 models that source PoE. All ten entries carrying poe-injector were in its "silent" cohort, and not one of them was missing its ports: every PSE port already carried the vendor's per-port wattage. What was missing was the device total, so Device fell through to the model-name parser (parsePoeBudget('POE60U-1BT-2-R') finds nothing), computePoe() skipped the box, and show power inline answered:

budget not modelled on <dev>: the ports are reported, the power they could supply is not

This is the same field-level failure as the Dream Wall in this issue (#130) and the four increments since - a number a human can read and the engine cannot spend.

The figures, all vendor-stated

model total vendor line
phihong/POE60U-1BT-2-R 60 W "Output Power 60W @40C; 30W @50C" - data sheet
phihong/POE90U-1BT-N6-R 90 W "Output Power - 90W ... Power/Port - 90" - product page
microchip/PD-9001GR/AT/AC 30 W "Output power of 30W is guaranteed" - DS00003973B
microchip/PD-9501GCO 60 W "PoE output: 60W" - product page
microchip/PD-9004G/AC 120 W DERIVED: "up to 30W per port for all ports" x 4 ports - DS00003976B; Microchip prints no single total, and the row's description already said 120 W
altronix/NetWay8G 240 W "Total power: NetWay8G - 240W" - product page
tp-link/TL-PoE150S 15.4 W "Power Supply 15.4W (Max. 48VDC)" - POE150S page (tp-link.com redirects there)
ubiquiti/POE-24-24W 24 W "Output Voltage 24VDC @ 1.0A" - PoE Adapters DS
ubiquiti/POE-48-24W 24 W "Output Voltage 48VDC @ 0.5A" - same sheet
mikrotik/RBGPOE unknown "Output Voltage Same as input", "Max. Current 2A" - product page: a bare injector has no wattage until a PSU is attached, so it is marked budgetUnknown like mikrotik/CRS112-8P-4S rather than credited with 24 V x 1 A MikroTik never printed

Every quoted row and its reading date is recorded in the spec's own poe.note; poe.sourceUrl on each file is the vendor host that was read.

The gate

test/poe-injector-budget.test.js (new, 70 assertions):

  • every poe-injector that sources power declares poe.budgetW or poe.budgetUnknown - the class cannot regress into silence;
  • each total equals the published figure, cites a vendor host, and records the reading;
  • the engine spends it: show power inline prints Available:<N>(w), never "budget not modelled";
  • the total bounds its ports (max(port) <= total <= sum(port));
  • RBGPOE still answers "budget not modelled" on purpose, because that is the honest answer for a pass-through injector.

Gates run

  • node scripts/syntax-check.js - 527 files, all OK
  • node scripts/audit-port-types.js - 0 contradicted (2704 reported, unchanged)
  • node scripts/audit-poe-budget.js - silent 32 -> 22, unknown 23 -> 24 (RBGPOE), inferred 0
  • node test/hardware-invariants.test.js - 26 passed, 0 failed
  • node scripts/import-devices.js (dry run) - unchanged from main
  • node scripts/regen-registry-doc.js --check - catalogue pages current (2,352 models)
  • node scripts/run-tests.js --concurrency=6 - 447 suites, 15,195 assertions, 0 failed

UDW-PRO status (why #130 stays open)

Re-probed this run at 11:2x UTC: https://techspecs.ui.com/unifi/cloud-gateways/udw-pro -> 307, 1 byte (Location: /); https://store.ui.com/us/en/products/udw-pro -> 404. Ubiquiti still publishes no page, datasheet, store listing or FCC grant for that SKU, so its port/PoE census remains unmodellable without inventing it.

Refs #130 (does not close it - the UDW-PRO is still unpublished). ## The defect `node scripts/audit-poe-budget.js` names 411 models that source PoE. All **ten** entries carrying `poe-injector` were in its "silent" cohort, and not one of them was missing its *ports*: every PSE port already carried the vendor's per-port wattage. What was missing was the device total, so `Device` fell through to the model-name parser (`parsePoeBudget('POE60U-1BT-2-R')` finds nothing), `computePoe()` skipped the box, and `show power inline` answered: ``` budget not modelled on <dev>: the ports are reported, the power they could supply is not ``` This is the same field-level failure as the Dream Wall in this issue (#130) and the four increments since - a number a human can read and the engine cannot spend. ## The figures, all vendor-stated | model | total | vendor line | |---|---|---| | phihong/POE60U-1BT-2-R | 60 W | "Output Power 60W @40C; 30W @50C" - [data sheet](https://www.phihong.com/wp-content/uploads/POE60U-1BT-X-R.pdf) | | phihong/POE90U-1BT-N6-R | 90 W | "Output Power - 90W ... Power/Port - 90" - [product page](https://www.phihong.com/products/power-over-ethernet/poe90u-1bt-n6-r/) | | microchip/PD-9001GR/AT/AC | 30 W | "Output power of 30W is guaranteed" - [DS00003973B](https://ww1.microchip.com/downloads/en/DeviceDoc/PD-9001GR_Datasheet_00003973B.pdf) | | microchip/PD-9501GCO | 60 W | "PoE output: 60W" - [product page](https://www.microchip.com/en-us/product/PD-9501GCO) | | microchip/PD-9004G/AC | 120 W | DERIVED: "up to 30W per port for all ports" x 4 ports - [DS00003976B](https://ww1.microchip.com/downloads/en/DeviceDoc/PD-9004G_Datasheet_00003976B.pdf); Microchip prints no single total, and the row's description already said 120 W | | altronix/NetWay8G | 240 W | "Total power: NetWay8G - 240W" - [product page](https://www.altronix.com/products/NetWay8G) | | tp-link/TL-PoE150S | 15.4 W | "Power Supply 15.4W (Max. 48VDC)" - [POE150S page](https://www.omadanetworks.com/us/business-networking/omada-accessory-poe-adapter/poe150s/) (tp-link.com redirects there) | | ubiquiti/POE-24-24W | 24 W | "Output Voltage 24VDC @ 1.0A" - [PoE Adapters DS](https://dl.ui.com/datasheets/poe/PoE_Adapters_DS.pdf) | | ubiquiti/POE-48-24W | 24 W | "Output Voltage 48VDC @ 0.5A" - same sheet | | mikrotik/RBGPOE | unknown | "Output Voltage Same as input", "Max. Current 2A" - [product page](https://mikrotik.com/product/RBGPOE): a bare injector has no wattage until a PSU is attached, so it is marked `budgetUnknown` like mikrotik/CRS112-8P-4S rather than credited with 24 V x 1 A MikroTik never printed | Every quoted row and its reading date is recorded in the spec's own `poe.note`; `poe.sourceUrl` on each file is the vendor host that was read. ## The gate `test/poe-injector-budget.test.js` (new, 70 assertions): - every `poe-injector` that sources power declares `poe.budgetW` or `poe.budgetUnknown` - the class cannot regress into silence; - each total equals the published figure, cites a vendor host, and records the reading; - the engine spends it: `show power inline` prints `Available:<N>(w)`, never "budget not modelled"; - the total bounds its ports (`max(port) <= total <= sum(port)`); - RBGPOE still answers "budget not modelled" on purpose, because that is the honest answer for a pass-through injector. ## Gates run - `node scripts/syntax-check.js` - 527 files, all OK - `node scripts/audit-port-types.js` - 0 contradicted (2704 reported, unchanged) - `node scripts/audit-poe-budget.js` - silent 32 -> 22, unknown 23 -> 24 (RBGPOE), inferred 0 - `node test/hardware-invariants.test.js` - 26 passed, 0 failed - `node scripts/import-devices.js` (dry run) - unchanged from main - `node scripts/regen-registry-doc.js --check` - catalogue pages current (2,352 models) - `node scripts/run-tests.js --concurrency=6` - 447 suites, 15,195 assertions, 0 failed ## UDW-PRO status (why #130 stays open) Re-probed this run at 11:2x UTC: `https://techspecs.ui.com/unifi/cloud-gateways/udw-pro` -> **307, 1 byte** (`Location: /`); `https://store.ui.com/us/en/products/udw-pro` -> **404**. Ubiquiti still publishes no page, datasheet, store listing or FCC grant for that SKU, so its port/PoE census remains unmodellable without inventing it.
devices(powerline): ten injectors named their ports and not their power
All checks were successful
CI / Server smoke and Docker build (pull_request) Successful in 32s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 7s
CI / Test (Node 20) (pull_request) Successful in 1m2s
CI / Test (Node 22) (pull_request) Successful in 1m3s
CI / Test (Node 18) (pull_request) Successful in 1m4s
CI / Arm auto-merge (pull_request) Successful in 3s
76c44ab8ac
All ten `poe-injector` entries sourced PoE on ports that already carried the
vendor's per-port wattage, and none declared a device total: `Device` fell
through to the model-name parser, found nothing in "POE60U-1BT-2-R" or
"NetWay8G", and `show power inline` answered "budget not modelled" - "the
ports are reported, the power they could supply is not".

Each now carries the vendor's own total with the quoted row in `poe.note`:
POE60U-1BT-2-R 60 W, POE90U-1BT-N6-R 90 W, PD-9001GR/AT/AC 30 W,
PD-9501GCO 60 W, PD-9004G/AC 120 W (4 x 30 W "for all ports"), NetWay8G
240 W, TL-PoE150S 15.4 W, POE-24-24W 24 W, POE-48-24W 24 W.
mikrotik/RBGPOE is marked budgetUnknown: a bare pass-through injector has no
output wattage until a PSU is attached, so 24 V x 1 A would be a number
MikroTik never printed.

test/poe-injector-budget.test.js gates the class: every poe-injector that
sources power declares a total or marks it unknown, and each total is spent
by the engine (`show power inline` prints Available:<N>(w)).
forgejo-actions scheduled this pull request to auto merge when all checks succeed 2026-11-21 11:29:23 +00:00
jaylouisw deleted branch catalogue/1121-1122 2026-11-21 11:39:28 +00:00
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!2413
No description provided.