admin
6a1f715a87
gates / gates (push) Successful in 6m43s
scripts/bench/plugmeter.py prints watts/mA/V/kWh as CSV from the Tuya plug on the LAN (default) or from Home Assistant on demo-hp's guest. 10-minute side by side and a 3-minute control in documentation/audits/plugmeter-2026-10-13/: Home Assistant alone did not move for 3 minutes, so the LAN reader is the measurement source. No key, device id or token in the repo; secrets scan 0 hits with a positive control.
69 lines
4.5 KiB
Markdown
69 lines
4.5 KiB
Markdown
# Reading the N100's metered plug from DooPlex — 2026-10-13
|
||
|
||
Two readers for the Tuya Wi-Fi plug that powers the N100 (`demo-felhom`): tinytuya on the LAN, and Home Assistant
|
||
installed from the Felhom catalog on demo-hp's guest 9201. Tool: `scripts/bench/plugmeter.py`
|
||
(`scripts/bench/README.md`). No key, device id or token is in this folder; they live only in DooPlex
|
||
`~/.config/felhom-bench/` (700/600).
|
||
|
||
## Setup facts
|
||
|
||
- tinytuya **1.20.0** in `~/.venvs/plugmeter` (user venv; no sudo, no system package, no service).
|
||
- Wizard: region `eu`, run in `~/.config/felhom-bench/` with its output sent to a 600 file. It found **1 device**,
|
||
„Smart Wi-Fi Power Meter Socket", category `cz`.
|
||
- `tinytuya scan`: LAN `192.168.0.115`, **protocol 3.4**. No fixed router address yet.
|
||
- Home Assistant: deployed through the dashboard endpoint (`POST /api/stacks/home-assistant/deploy`,
|
||
`{"DOMAIN":"enkisfelhom.hu","SUBDOMAIN":"ha"}`) at 14:33:34Z → **202**, container healthy at 14:35:46Z
|
||
(image pull included). Before the first setup the public address answered **401** (the setup gate closed, as
|
||
designed). The operator passed the gate with the dashboard sign-in, made the owner, added the Tuya
|
||
integration by QR, and made a non-admin, local-access-only user `felhom-bench` with a long-lived token.
|
||
|
||
## Data points (read from the device, checked against Home Assistant's Tuya reading)
|
||
|
||
| DP | code | unit | scale | CSV column |
|
||
|---|---|---|---|---|
|
||
| 19 | `cur_power` | W | ×10 | `watts` |
|
||
| 18 | `cur_current` | mA | ×1 | `milliamps` |
|
||
| 20 | `cur_voltage` | V | ×10 | `volts` |
|
||
| 17 | `add_ele` | kWh | ×1000 | `kwh_total` — energy since the plug last reported to the cloud, NOT lifetime |
|
||
|
||
Check: at 14:44–14:45Z tinytuya read 6.6–7.0 W, 59–60 mA, 229.9 V; Home Assistant (Tuya cloud — the same
|
||
source the Smart Life app shows) read „7 W, 0.06 A, 230 V" on the operator's screen and 6.6 W through the API.
|
||
Match within 0.5 W. Home Assistant's entities: `sensor.smart_wi_fi_power_meter_socket_{power,current,voltage,total_energy}`.
|
||
|
||
A refresh is required: plain `status()` repeated one cached reading (3 identical rows at 2 s); `updatedps([18,19,20])`
|
||
then `status()` gives a new value each time — the plug's previous measurement, so a row lags about one interval.
|
||
|
||
## Side by side — 10 minutes, N100 at rest, `--interval 2` (14:45:17–14:55:17Z)
|
||
|
||
| | rows | mean W | median W | spread (sd) | min–max W | read gap median / max | gaps > 3 s | failed reads |
|
||
|---|---|---|---|---|---|---|---|---|
|
||
| tuya (LAN) | 215 | 8.36 | 7.20 | 2.67 | 6.1–19.1 | 2.0 s / 8.0 s | 56 | 0 |
|
||
| ha (cloud) | 301 | 8.40 | 7.20 | 2.74 | 6.1–19.1 | 2.0 s / 2.1 s | 0 | 0 |
|
||
|
||
`side-by-side-tuya.csv`, `side-by-side-ha.csv`. The two agree almost exactly — **but not independently**:
|
||
|
||
**Control (`ha-alone-3min.csv`, 14:55:44–14:58:44Z):** with the tuya reader stopped, Home Assistant's power value
|
||
**did not change once in 3 minutes** (6.5 W throughout). When the tuya reader ran again for 20 s, Home Assistant
|
||
moved to the last value it pushed (7.9 W) and stayed there. So the plug reports to the cloud only when asked; Home
|
||
Assistant's 2-second freshness in the side-by-side was the tuya reader's doing. Earlier, at 14:44:48Z, Home
|
||
Assistant still showed 18.7 W from 14:44:21Z while the plug read 6.7 W.
|
||
|
||
**Recommendation: `--source tuya` for every measurement; `--source ha` as the backup and for history.**
|
||
Each tuya gap > 3 s is a slow read (up to 8 s), not a lost one.
|
||
|
||
## Home Assistant from our catalog — product notes
|
||
|
||
- Install, gate and first setup worked with no defect found. Deploy-to-healthy ~2 min 12 s.
|
||
- A non-admin Home Assistant user **can still switch the plug** (`switch.smart_wi_fi_power_meter_socket_socket_1`
|
||
is visible to it). Home Assistant has no read-only role. So the `felhom-bench` token can power off the N100;
|
||
plugmeter only reads states. Home Assistant's design, not ours — no register row.
|
||
- A household reading a Tuya meter plug only through Home Assistant's cloud integration gets a value that can sit
|
||
still for minutes. Tuya's behaviour, not ours — no register row.
|
||
- The `felhom-bench` user is local-access-only, so the public (Cloudflare) path is refused for it; plugmeter
|
||
connects to the guest's LAN address `192.168.0.155` and checks the certificate by name.
|
||
|
||
## Teardown
|
||
|
||
Machine: nothing provisioned on the N100. Host: none. Guest 9201: **Home Assistant stays installed on purpose**
|
||
(the plug's second reader, `operations/nodes.md`). Hub: nothing. DooPlex: the user venv and `~/.config/felhom-bench/` stay.
|