Add ThinQ2 device handler for LG WT7300CW washer (T1789EFH_F)#77
Add ThinQ2 device handler for LG WT7300CW washer (T1789EFH_F)#77brystmar wants to merge 2 commits into
Conversation
anszom
left a comment
anszom
left a comment
There was a problem hiding this comment.
Thanks for the PR. Please see the attached review comments.
I would also ask you to add a unit test containing some data packets captured from your device for validation. See other devices' implementations for reference.
| ) | ||
| } | ||
|
|
||
| private processRecord(rec: Buffer) { |
There was a problem hiding this comment.
What about the remaining bytes? Take a look at other washers' implementations - I expect that many fields would carry over.
There was a problem hiding this comment.
I'm a PM by trade, so I decided to cut scope and get an MVP in place to validate the approach first. Plan was to fill in the other elements as I do more laundry cycles at home. Does that work for you, or do I need to decipher all the bytes first?
| if (this.lastPhase === 0 && phase !== 0) { | ||
| this.startTime = new Date().toISOString() | ||
| this.publishProperty('start_time', this.startTime) | ||
| } else if (this.lastPhase !== 0 && phase === 0) { | ||
| this.stopTime = new Date().toISOString() | ||
| this.publishProperty('stop_time', this.stopTime) | ||
| } | ||
| this.lastPhase = phase |
There was a problem hiding this comment.
I'm not really convinced that rethink should be doing this kind of processing. We are exposing raw data to HomeAssistant, and the user can do any processing they like on that side.
Also, If we wanted to do any post-processing, it should be implemented uniformly for all washers (and maybe other devices).
| remaining_time_min: { | ||
| platform: 'sensor', | ||
| unique_id: '$deviceid-remaining_time_min', | ||
| state_topic: '$this/remaining_time_min', | ||
| name: 'Remaining minutes', | ||
| device_class: 'duration', | ||
| unit_of_measurement: 'min', | ||
| }, | ||
| remaining_time_sec: { | ||
| platform: 'sensor', | ||
| unique_id: '$deviceid-remaining_time_sec', | ||
| state_topic: '$this/remaining_time_sec', | ||
| name: 'Remaining seconds', | ||
| device_class: 'duration', | ||
| unit_of_measurement: 's', | ||
| }, |
There was a problem hiding this comment.
This should be a single remaining_time field.
…s, add tests Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
brystmar
left a comment
brystmar
left a comment
There was a problem hiding this comment.
Removed the start/stop sensors, consolidated remaining_time fields, and added tests for LG WT7300CW (T1789EFH_F).
Removed the start_time and stop_time sensors along with the lastPhase tracking logic. These were derived from phase transitions rather than device-reported values, making them unreliable across reconnects.
Replaced remaining_time_min and remaining_time_sec with a single remaining_time field (minutes, device_class: duration), consistent with other device implementations.
I also fixed what appeared to be a bug in the original byte offsets: remaining_time_min was reading rec[3], which is always 0x00 in practice. The actual minutes field is rec[4]. I confirmed this by cross-referencing real captures against the ThinQ app: the running/paused bracketing sequence shows rec[4]=43 frozen across all three packets, and the heavy-duty capture's rec[4]=24 matches the corroborating 0xE2 packet's "~24 min remaining" reading.
Added 16 unit tests covering synthetic packets (Off, Wash main, unknown phase fallback) and 8 real validated captures from a live LG WT7300CW, including the full running > paused > resumed sequence confirming that remaining_time stays frozen during a pause. Also confirms that 0xE2 settings-echo, 0xD8 idle/ack, and 0x31 identity packets are silently ignored without errors.
2 participants
Adds support for the LG WT7300CW top-load washer (internal model
T1789EFH_F, deviceType 201, ThinQ2/LCW-007 Wi-Fi module).What this adds
A new device handler that parses the AA/BB-framed status packets (
0xECdual-record and0xEBsingle-record) to expose the following Home Assistant sensors:Bug Fix
This PR also fixes a bug in the original byte offsets:
remaining_time_minwas readingrec[3], which is always0x00in practice. I believe the actual minutes field isrec[4], confirmed by cross-referencing real captures against the ThinQ app. The running/paused bracketing sequence showsrec[4]=43frozen across all three packets, and the heavy-duty capture'srec[4]=24matches the corroborating0xE2packet's "~24 min remaining" reading.Notes
0xECpacket carries two back-to-back 27-byte records (current + previous); only the current (first) record is used.POWER_ONreconnect) — this is normal device behavior, not a network issue.