Vendor Documents — citation register¶
Documents cited across this project, with the version and date printed in the document itself. Cite from here rather than by title alone: a title without a version ages silently, and FranklinWH revises these.
Per AP-14, a vendor statement is CONFIRMED evidence — but only against a stated revision. Behaviour has already been observed changing between revisions (see the Eth port naming below), so "the manual says" is not a durable citation on its own.
Register¶
| Key | Document | Version | Issued | Pages |
|---|---|---|---|---|
INST-1.2.07 |
FranklinWH System Installation Guide | 1.2.07 | May 09, 2026 | 89 |
COMM-AU-2.15.0 |
FranklinWH Commissioning Guide – AU & NZ (App Version 2.15.0) | 2.15.0 | June 12, 2026 | 61 |
COMM-2.8.0 |
FranklinWH Commissioning Guide (MAC 1 & aPower 2 & aPower S) | — | Dec 30, 2025 | 82 |
DS-AGATE-AU-V1.7 |
System Datasheet aGate X-01-AU & aPower X-02-AU | V1.7 | 2026-05-30 | 2 |
AHUB-V1.0 |
aHub Installation and Operations Manual | V1.0 (initial release) | 2026-01-22 | 37 |
FWH-SUPPORT-SC |
franklinwh.com — Support → Overview → Smart Circuits | — | retrieved 2026-09-18 | web |
SPAN-APPNOTE-FWH |
SPAN Tech Portal — App Note: Integration with FranklinWH | — | retrieved 2026-09-18 | — |
DS-SC-AU |
Smart Circuits Module Datasheet (AU & NZ) | — | 2026-03-25 | 2 |
Versions above are quoted from the documents' own title pages or revision
history, not inferred from filenames or PDF metadata. Where a document
prints no version, the column reads — rather than guessing.
Why revisions matter here¶
Two documents in this register contradict each other, and the contradiction is only resolvable because each is dated and scoped:
| Source | Hardware | Household cable → | Other port |
|---|---|---|---|
INST-1.2.07 p.62 |
(unqualified) | Eth1 | — |
COMM-2.8.0 p.7 |
aGate X 1.1 (AGT-R1V1-US) |
Eth2 | Eth1 (Debug) |
COMM-2.8.0 p.7 |
aGate X 1.3 / 1.3.1 | single ETH port | — |
"Eth1" is the internet port in one and the Debug port in the other. Citing
either without its revision and hardware scope produces a confident wrong
answer — which is how DEF-ETH-PORT-IDENTITY-UNCONFIRMED arose.
Citations in use¶
| Claim | Source |
|---|---|
AU/NZ supply is 230/240 VAC L/N/PE, 50 Hz |
DS-AGATE-AU-V1.7 |
| AU/NZ Smart Circuits: two circuits per aGate | DS-AGATE-AU-V1.7 |
aGate X-01-AU = AGT-R1V1-AU; aPower X-02-AU = APR-05K15V1-AU |
DS-AGATE-AU-V1.7 |
| aGate joins 2.4 GHz Wi-Fi only | INST-1.2.07 p.59 |
| Household cable to Eth1 | INST-1.2.07 p.62 |
| 4G is backup only | INST-1.2.07 p.59 |
aGate AP is AP_<last 9 of SN>, password <last 12 of SN> |
INST-1.2.07 p.59, COMM-AU-2.15.0 p.11 |
| Port naming differs by revision; Eth1 is Debug on aGate X 1.1 | COMM-2.8.0 p.7 |
| Wi-Fi is commissioned over the aGate's own hotspot, locally | COMM-AU-2.15.0 pp.10–13 |
| 4G connected by default; app offers Skip | COMM-AU-2.15.0 p.12 |
| Phone may drop the aGate connection after a successful Wi-Fi change | COMM-AU-2.15.0 p.12 |
| aHub: 4×240 V or 8×120 V circuits, programmable scheduling | AHUB-V1.0 p.10, p.28 |
| SPAN link is SunSpec Modbus TCP 502, aGate as client; SPAN Aux Comms/ETH-1 to aGate Ethernet; SPAN supplies the aGate's internet | SPAN-APPNOTE-FWH |
| Three Smart Circuits; SC1+SC2 merge when sharing a 2-pole switch | FWH-SUPPORT-SC |
| SOC Auto Cut-off sheds circuits during a grid outage — off-grid only | FWH-SUPPORT-SC |
| Overload cut-off disconnects circuits if load exceeds panel capacity | FWH-SUPPORT-SC |
| Schedules set a "cycle interval and execution time" | FWH-SUPPORT-SC |
Note on
SPAN-APPNOTE-FWH: the SPAN Tech Portal is a JavaScript-rendered Salesforce site and cannot be fetched programmatically; this entry was recorded from screenshots. It prints no version or date, so the register records only the retrieval date — an honest—rather than an invented revision.
Keeping this current¶
When a newer revision appears, add a row — do not overwrite one. A claim
cited against INST-1.2.07 remains true of that revision even after 1.3 ships;
what changes is whether it is still true of current hardware. Overwriting
destroys the ability to tell those apart, which is the whole reason the Eth port
contradiction was resolvable.