Skip to content

Commit d3ef93e

Browse files
0xAHAclaude
andcommitted
docs: correct the ShineWiLan entry - must be -X2, and it has a real limit
Two corrections to what I wrote an hour ago. It must be the -X2. The earlier ShineWiLan has no Modbus TCP server and cannot work with this integration at all, confirmed by a user who swapped modules. Anyone buying one or asking Growatt for it needs the suffix, so that is now a warning admonition rather than a passing mention. And 'known good' was too generous. My own analysis on #308 said the X2's local Modbus server is built for light polling rather than sustained access from a full integration, and that reporter saw repeated drops on a WIT -- the most register-hungry profile here. Recording that alongside the working reports from #309 and #336, with the mitigation (raise scan_interval before suspecting anything else) and the honest trade-off: a dedicated adapter for fast polling, the X2 if you want one device doing both local and cloud. Also states the genuine advantage properly -- it serves Modbus locally while keeping the cloud connection, so users are not choosing between Home Assistant and the ShinePhone app. Co-Authored-By: Claude <noreply@anthropic.com>
1 parent 8c14d11 commit d3ef93e

1 file changed

Lines changed: 11 additions & 2 deletions

File tree

docs/troubleshooting/rs485-gateways.md

Lines changed: 11 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -48,9 +48,18 @@ Set the work mode so it performs **Modbus TCP to RTU** conversion rather than pl
4848

4949
One EW11A report ([#309](https://github.com/0xAHA/Growatt_ModbusTCP/issues/309)) showed all entities reading zero, but the same symptom followed the reporter onto a Waveshare adapter, so the gateway was not the cause. They resolved it with a Growatt WiLan-X2.
5050

51-
### ✅ Growatt WiLan-X2 — known good
51+
### ✅ Growatt ShineWiLan-X2 — works, within limits
5252

53-
Growatt's own dongle, which exposes Modbus directly. Reported working in [#309](https://github.com/0xAHA/Growatt_ModbusTCP/issues/309) after two third-party adapters had been ruled out. It also buffers roughly a month of data and re-synchronises after a power cut, which no generic serial server does.
53+
Growatt's own dongle. It exposes a **local Modbus TCP server on port 502 while keeping its cloud connection**, so you get local data in Home Assistant and the ShinePhone app at the same time — no need to choose. It also buffers data and re-synchronises after a power cut, which no generic serial server does.
54+
55+
!!! warning "It must be the **-X2**"
56+
The earlier ShineWiLan has **no Modbus TCP server** and cannot be used with this integration at all. Confirmed by a user who swapped modules. If you are buying or asking Growatt for one, the `-X2` suffix is the whole difference.
57+
58+
**The one real limitation:** that local Modbus server is built for light polling, not sustained access from a full integration. A WIT owner saw the connection drop repeatedly under normal polling ([#308](https://github.com/0xAHA/Growatt_ModbusTCP/issues/308)), and WIT is the most register-hungry profile here.
59+
60+
If you see intermittent drops on an X2, raise **scan_interval** before suspecting anything else — 60 s or more, and longer again on WIT. A dedicated RS485 adapter is the answer if you want fast polling; the X2 is the answer if you want one device doing both jobs.
61+
62+
Working setups reported in [#309](https://github.com/0xAHA/Growatt_ModbusTCP/issues/309) (after two third-party adapters were ruled out) and [#336](https://github.com/0xAHA/Growatt_ModbusTCP/issues/336) on MOD/MID XH.
5463

5564
### ⚠️ PUSR / ShineWiFi-class serial bridges — replay stale frames
5665

0 commit comments

Comments
 (0)