devices(peplink): the HD4 MBX could not power the eight PoE+ ports the vendor prints (refs #130) #2209

Merged
jaylouisw merged 1 commit from catalogue/1119-2052 into main 2026-11-19 21:12:49 +00:00
Owner

Refs #130 - does not close it. The UDW-PRO the issue asks for is still unpublished; this is the catalogue defect found on the routed-lab path the issue needs (a router/gateway node with NAT, DHCP, VPN and cellular WAN, which is what an HD4 MBX is).

What was wrong

devices/peplink/max-hd4-mbx-glte.json declared eight poeOut ports and no poe block, so computePoe() opened with if (!dev.poeBudget) continue; and skipped the whole box: eight 802.3at outlets that powered nothing, silently.

The same row was a copy of the non-MBX MAX HD4 datasheet (pepwave_max_hd4_datasheet.pdf, 200, 807,072 B, 2 pp), so it claimed 400 Mbps, ten copper ports (2 WAN + 8 LAN) and no budget. The MBX is a different chassis: 2.5 Gbps, eleven GE ports (3 WAN + 8 LAN), 55 W max consumption against the HD4's 38 W, and a 54V/3.34A ACW-630 adapter against the HD4's 12V/4.2A.

What the vendor prints

https://www.peplink.com/compare/tech-specs/hd4-mbx.pdf (200, application/pdf, 226,176 B, 6 pp), read 2026-11-19:

field value page
Ethernet WAN 3 / 2 (GE) 1
Ethernet LAN 8 / 9 (GE) 1
port configurations 3 WAN with 8 LAN, 2 WAN with 9 LAN 1
Router Throughput 2.5Gbps 1
PoE Input No 6
PoE Output Up to 8x 802.3at PoE+ 6
footnote 2x 54V DC input is needed for all 8x LAN ports to have 802.3at PoE. Plugging in 1x 54V DC input will result in 4x LAN ports having 802.3at PoE. 6
AC adapter DC Output 54V 3.34A [13] Compatible with ACW-630 6
consumption nominal 35, max 55W (w/o PoE) 6

The 5G sheet (tech-specs/hd4-mbx-5g.pdf, 200, 226,055 B, 6 pp) prints the same rows. The SKUs are real: the product page (200, 187,714 B) names MAX-HD4-MBX-GLTE-G-T and MAX-HD4-MBX-GLTE-S-T - GLTE is the CAT-20 option of this chassis.

The budget is derived, and the vendor's own footnote checks it

two 54V inputs : 2 x (54 V x 3.34 A) = 360.72 W  -  55 W  =  305.72 W  ->  eight 30 W ports (240 W) fit
one 54V input  :      (54 V x 3.34 A) = 180.36 W  -  55 W  =  125.36 W  ->  four 30 W ports (120 W) fit, five do not

The second line is Peplink's 1x 54V DC input will result in 4x LAN ports, reproduced exactly by the subtraction - which is why the first is trusted and recorded as poe.budgetW: 305.7, marked DERIVED in poe.note. It models the two-supply build, the only one in which all eight LAN ports source power. Which eight? The eight LAN ports of the 3 WAN with 8 LAN build, where outlet count and port count coincide, so no port has to be guessed; the 2 WAN with 9 LAN default is recorded in notes because the vendor does not say which of the nine loses power.

The siblings had the same hole

MAX-HD4-MBX-LTEA and MAX-HD4-MBX-5G print the same Up to 8x 802.3at PoE+ row and had no PoE ports at all. LTEA also cited download.peplink.com/resources/hd4.pdf, which now answers HTTP 403; it is re-pointed at the MBX sheet, whose printed rows match its existing notes exactly. All three now carry eight PoE+ LAN ports and the 305.7 W derived budget.

Verified

  • computePoe() now has an entry for the box; an AP One Flex on lan1 reports powered: true, used: 12.95, against budget: 305.7.
  • test/peplink-hd4-mbx-poe.test.js: 30 passed, 0 failed - pins the 11-port census, the eight PoE+ LAN ports, the derivation (including that one supply funds four ports and not five) and the siblings.
  • research/peplink-hd4-mbx-poe-census.md records the census, the wrong-chassis trap and the arithmetic.
  • research/incoming GLTE entry dropped so a re-import cannot restore the 400 Mbps copy (importer dry run: 1187 -> 1186 scanned, Would add: 1 unchanged, same 2 pre-existing rejections).
  • Gates: syntax 466 OK, audit-port-types clean, audit-poe-budget silent PSEs 102 -> 101 (409 PSE models), hardware invariants 26/0, spec-vs-description 3/0, duplicate-models 21/0, control bytes clean, doc citations clean, registry doc regenerated (2,355 models, 29,721 sockets, 971 shapes), run-tests 388 suites / 13,029 assertions / 0 failed.

Not touched

MAX-HD4-MBX (International) still cites tech-specs/hd4.pdf - the plain HD4 compare sheet - for 400 Mbps and 2-5 WAN / 5-8 LAN, which looks like the same copy-the-wrong-chassis defect, but no MBX-specific sheet prints those figures, so it is left as the open question rather than guessed. devices/peplink/max-hd4-lte.json is the genuine plain HD4 and stays on its own datasheet.

Refs #130 - does not close it. The UDW-PRO the issue asks for is still unpublished; this is the catalogue defect found on the routed-lab path the issue needs (a router/gateway node with NAT, DHCP, VPN and cellular WAN, which is what an HD4 MBX is). ## What was wrong `devices/peplink/max-hd4-mbx-glte.json` declared eight `poeOut` ports and no `poe` block, so `computePoe()` opened with `if (!dev.poeBudget) continue;` and skipped the whole box: eight 802.3at outlets that powered nothing, silently. The same row was a copy of the **non-MBX** MAX HD4 datasheet (`pepwave_max_hd4_datasheet.pdf`, 200, 807,072 B, 2 pp), so it claimed **400 Mbps**, **ten** copper ports (2 WAN + 8 LAN) and no budget. The MBX is a different chassis: **2.5 Gbps**, **eleven** GE ports (3 WAN + 8 LAN), **55 W** max consumption against the HD4's 38 W, and a 54V/3.34A ACW-630 adapter against the HD4's 12V/4.2A. ## What the vendor prints `https://www.peplink.com/compare/tech-specs/hd4-mbx.pdf` (200, `application/pdf`, 226,176 B, 6 pp), read 2026-11-19: | field | value | page | |---|---|---| | Ethernet WAN | `3 / 2 (GE)` | 1 | | Ethernet LAN | `8 / 9 (GE)` | 1 | | port configurations | `3 WAN with 8 LAN, 2 WAN with 9 LAN` | 1 | | Router Throughput | `2.5Gbps` | 1 | | PoE Input | `No` | 6 | | PoE Output | `Up to 8x 802.3at PoE+` | 6 | | footnote | `2x 54V DC input is needed for all 8x LAN ports to have 802.3at PoE. Plugging in 1x 54V DC input will result in 4x LAN ports having 802.3at PoE.` | 6 | | AC adapter | `DC Output 54V 3.34A [13] Compatible with ACW-630` | 6 | | consumption | nominal `35`, max `55W (w/o PoE)` | 6 | The 5G sheet (`tech-specs/hd4-mbx-5g.pdf`, 200, 226,055 B, 6 pp) prints the same rows. The SKUs are real: the product page (200, 187,714 B) names `MAX-HD4-MBX-GLTE-G-T` and `MAX-HD4-MBX-GLTE-S-T` - GLTE is the CAT-20 option of this chassis. ## The budget is derived, and the vendor's own footnote checks it ``` two 54V inputs : 2 x (54 V x 3.34 A) = 360.72 W - 55 W = 305.72 W -> eight 30 W ports (240 W) fit one 54V input : (54 V x 3.34 A) = 180.36 W - 55 W = 125.36 W -> four 30 W ports (120 W) fit, five do not ``` The second line is Peplink's `1x 54V DC input will result in 4x LAN ports`, reproduced exactly by the subtraction - which is why the first is trusted and recorded as `poe.budgetW: 305.7`, marked DERIVED in `poe.note`. It models the two-supply build, the only one in which all eight LAN ports source power. Which eight? The eight **LAN** ports of the `3 WAN with 8 LAN` build, where outlet count and port count coincide, so no port has to be guessed; the `2 WAN with 9 LAN` default is recorded in `notes` because the vendor does not say which of the nine loses power. ## The siblings had the same hole `MAX-HD4-MBX-LTEA` and `MAX-HD4-MBX-5G` print the same `Up to 8x 802.3at PoE+` row and had **no PoE ports at all**. LTEA also cited `download.peplink.com/resources/hd4.pdf`, which now answers **HTTP 403**; it is re-pointed at the MBX sheet, whose printed rows match its existing notes exactly. All three now carry eight PoE+ LAN ports and the 305.7 W derived budget. ## Verified - `computePoe()` now has an entry for the box; an AP One Flex on `lan1` reports `powered: true`, `used: 12.95`, against `budget: 305.7`. - `test/peplink-hd4-mbx-poe.test.js`: 30 passed, 0 failed - pins the 11-port census, the eight PoE+ LAN ports, the derivation (including that one supply funds four ports and not five) and the siblings. - `research/peplink-hd4-mbx-poe-census.md` records the census, the wrong-chassis trap and the arithmetic. - `research/incoming` GLTE entry dropped so a re-import cannot restore the 400 Mbps copy (importer dry run: 1187 -> 1186 scanned, `Would add: 1` unchanged, same 2 pre-existing rejections). - Gates: syntax 466 OK, `audit-port-types` clean, `audit-poe-budget` silent PSEs **102 -> 101** (409 PSE models), hardware invariants 26/0, spec-vs-description 3/0, duplicate-models 21/0, control bytes clean, doc citations clean, registry doc regenerated (2,355 models, 29,721 sockets, 971 shapes), `run-tests` 388 suites / 13,029 assertions / 0 failed. ## Not touched `MAX-HD4-MBX` (International) still cites `tech-specs/hd4.pdf` - the **plain** HD4 compare sheet - for 400 Mbps and 2-5 WAN / 5-8 LAN, which looks like the same copy-the-wrong-chassis defect, but no MBX-specific sheet prints those figures, so it is left as the open question rather than guessed. `devices/peplink/max-hd4-lte.json` is the genuine plain HD4 and stays on its own datasheet.
devices(peplink): the HD4 MBX could not power the eight PoE+ ports the vendor prints (refs #130)
All checks were successful
CI / Classify the change (pull_request) Successful in 15s
PR rebase / Rebase onto main and arm auto-merge (pull_request) Successful in 11s
CI / Server smoke test (pull_request) Successful in 11s
CI / Test (Node 20) (pull_request) Successful in 56s
CI / Test (Node 22) (pull_request) Successful in 57s
CI / Test (Node 18) (pull_request) Successful in 1m2s
CI / Docker build (pull_request) Successful in 20s
CI / Arm auto-merge (pull_request) Successful in 4s
f9f2bcb669
The three MAX HD4 MBX rows carried no usable PoE census. MAX-HD4-MBX-GLTE
declared eight poeOut ports and no poe block, so computePoe() skipped the box at
'if (!dev.poeBudget) continue;' - eight 802.3at outlets that powered nothing.

The same row was a copy of the NON-MBX MAX HD4 datasheet: 400 Mbps, ten copper
ports (2 WAN + 8 LAN), no budget. The MBX is a different chassis - 2.5 Gbps,
eleven GE ports (3 WAN + 8 LAN), 55 W max consumption, a 54V/3.34A ACW-630
adapter - so it understated throughput six-fold and was one LAN port short.

Budget is derived, not quoted: 2 x (54 V x 3.34 A) = 360.72 W, less 'Maximum
Power Consumption 55W (w/o PoE)' = 305.72 W for eight 30 W ports. The vendor's
own one-supply footnote ('1x 54V DC input will result in 4x LAN ports') is
reproduced exactly by the same subtraction: 180.36 - 55 = 125.36 W, which funds
four ports and not five.

MAX-HD4-MBX-LTEA and MAX-HD4-MBX-5G print the same PoE row and had no PoE ports
at all; LTEA also cited download.peplink.com/resources/hd4.pdf, now HTTP 403.

Sources: tech-specs/hd4-mbx.pdf (200, 226,176 B, 6 pp, pages 1 and 6),
tech-specs/hd4-mbx-5g.pdf (200, 226,055 B, 6 pp), product page 187,714 B naming
MAX-HD4-MBX-GLTE-G-T / -S-T. research/peplink-hd4-mbx-poe-census.md records the
census and the arithmetic; test/peplink-hd4-mbx-poe.test.js pins both.

The GLTE entry is dropped from research/incoming so a re-import cannot restore
the wrong-chassis copy.
forgejo-actions scheduled this pull request to auto merge when all checks succeed 2026-11-19 21:04:23 +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!2209
No description provided.