Anker Prime 250W "Unexpected end of packet", sensors unavailable

[pkolbus]

Hi, recently purchased an Anker Prime 250W charger and attempting to integrate into Home Assistant. Ran across this library and https://github.com/flip-dots/HaSolixBLE which seemed the most promising. When I add to Home Assistant (using #18) i can see the device and sensors but all ports are showing unavailable. (For what it's worth, the discovery/connection is via one of several ESPHome Bluetooth proxies).

(For what it's worth, https://github.com/thomluther/ha-anker-solix also does not work).

I see you've written some great documentation for adding/debugging a device; intend to to dive in to help debug/fix this. If you have any suggestions - things to look at or try -- these would be welcomed.

2026-03-14 19:06:20.374 ERROR (MainThread) [SolixBLE.device] Unexpected end of packet! Data may be missing or invalid! Payload: '00a10101044859710a627e653606c6d0f100b4e2'
Traceback (most recent call last):
  File "/usr/local/lib/python3.14/site-packages/SolixBLE/device.py", line 389, in _parse_payload
    param_data = bytes([remaining_data.pop(0) for _ in range(0, param_len)])
                        ~~~~~~~~~~~~~~~~~~^^^
IndexError: pop from empty bytearray
2026-03-14 19:06:20.475 ERROR (MainThread) [SolixBLE.device] Unexpected end of packet! Data may be missing or invalid! Payload: '00a10102a202fd00a30144a40101a501028c2e88c598023c7d27d6c9edddf0af9e'
Traceback (most recent call last):
  File "/usr/local/lib/python3.14/site-packages/SolixBLE/device.py", line 389, in _parse_payload
    param_data = bytes([remaining_data.pop(0) for _ in range(0, param_len)])
                        ~~~~~~~~~~~~~~~~~~^^^
IndexError: pop from empty bytearray
2026-03-14 19:06:20.578 ERROR (MainThread) [SolixBLE.device] Unexpected end of packet! Data may be missing or invalid! Payload: '00a10103a2054553503332a307302e302e302e33a41041514c514a4231463532323030303832a5067ce913833daab78ca884812ed76cba4e013df23dac65'
Traceback (most recent call last):
  File "/usr/local/lib/python3.14/site-packages/SolixBLE/device.py", line 389, in _parse_payload
    param_data = bytes([remaining_data.pop(0) for _ in range(0, param_len)])
                        ~~~~~~~~~~~~~~~~~~^^^
IndexError: pop from empty bytearray
2026-03-14 19:06:20.679 ERROR (MainThread) [SolixBLE.device] Unexpected end of packet! Data may be missing or invalid! Payload: '002d4f48bf60a98f2ef64e7ec1fd12864c'
Traceback (most recent call last):
  File "/usr/local/lib/python3.14/site-packages/SolixBLE/device.py", line 389, in _parse_payload
    param_data = bytes([remaining_data.pop(0) for _ in range(0, param_len)])
                        ~~~~~~~~~~~~~~~~~~^^^
IndexError: pop from empty bytearray
2026-03-14 19:06:20.987 ERROR (MainThread) [SolixBLE.device] Unexpected end of packet! Data may be missing or invalid! Payload: '00a140845fba1b20b28172b243e9b37d0b0c706149deb48954ed2d78f3fe960472aa77ca218d1289c35a751941ab064ce9be64778ffc4c08f962d074c5309c6e790e9db76bffcf9ac81660b09313dc879ed0ec'
Traceback (most recent call last):
  File "/usr/local/lib/python3.14/site-packages/SolixBLE/device.py", line 389, in _parse_payload
    param_data = bytes([remaining_data.pop(0) for _ in range(0, param_len)])
                        ~~~~~~~~~~~~~~~~~~^^^
IndexError: pop from empty bytearray

[thomluther]

(For what it's worth, https://github.com/thomluther/ha-anker-solix also does not work).

What's the reason why this doesn't work through the cloud MQTT server? What is your model PN?

[pkolbus]

My charger has PN A2345. I'd been playing around with different configuration options and button entities, the closest I got was seeing entities for each port with a state of 'unavailable' and a message saying that the entity was no longer provided by the anker_solix integration. I've since done a reset of the charger to bring back the BLE communication.

[thomluther]

You would need to automate status requests or realtime triggers in HA, otherwise the device may not send regular MQTT messages and entities become stale or remain unavailable after reload.

[flip-dots]

@pkolbus could you send some debug logs, inlcuding the negotiation which happens when it attempts to connect to the device, those warnings arent super useful on their own, they just mean that they were not able to parse the end of the packet which can happen for any number of reasons, but happens a lot with this expertimental version as I haven't got the protocol fully figured out yet as these Anker Prime devices are different from the Solix ones.

The more logs the better, I have my suspicions about what the issue is but it would be good to confirm it.

[pkolbus]

@flip-dots here you go; let me know if there's anything else I can do (Unrelated warnings/errors are removed).

home-assistant_solix_ble_2026-03-15T13-02-33.702Z.trimmed..log

[flip-dots]

@pkolbus yep, looks like the sus thing I thought it was, give the latest push a try, it should at least you get further into the negotiation stages, there are actually 7 not 5 stages (there is an issue with this branch where it says negotiations are finished when they are not, I plan on fixing that later).

[pkolbus]

With the integration at commit cc2d6412815ef7001e8b27c37ae5fa9e3a816a81, looks like more progress but all port entities are still showing unavailable. Here's an updated log. I'm also not sure whether the connection is resuming correctly after a HA restart, but will take another look at that once SolixBLE can parse the messages properly.

home-assistant_solix_ble_2026-03-15T19-02-26.051Z.trimmed.log

[flip-dots]

@pkolbus I am not sure what to make of this. It sort of looks like the shared secret negotiation gives an incorrect secret, explaining why the earlier packets seem to be junk and fail to decode, but the last packet in the log is valid when decrypted using that shared secret, so it can't be that. I have pushed another update which should make it decrypt the payload even if it fails the authenticity check and improved the logging of the negotation which might help or at least give us some more clues.

Is anything plugged in? These devices only send updates when something changes so if nothing is plugged in it would explain why the values aren't populating. Also does anything happen after these log entries, does it just sit there or does it get kicked or do any other packets arrive?

[pkolbus]

@thomluther I went back to the ha-anker-solix integration again (v3.5.3), with a more controlled test and also with having the iOS app running intermittently during the test, and that did "sort of" work. Opened thomluther/ha-anker-solix#470 for the issue I'm running into. That said, I would like to get SolixBLE working for the reliability and privacy that local-only options offer. @flip-dots hopefully that helps.

@flip-dots five of the six USB ports have devices connected including my laptop, so the consumed power should be changing. I've waited for several minutes after the last entries in the log and didn't see anything else. I'll reset the charger and try your latest update; expect further news within the next few hours.

[pkolbus]

@flip-dots sensors are still unavailable... here's the latest log.

home-assistant_solix_ble_2026-03-15T23-14-41.364Z.log

[flip-dots]

@pkolbus Thanks, that does narrow things down a bit.

It looks like something is going wrong in the stage 6 negotiations as in response to sending the stage 6 message the device gives 09a1021e00 whereas it gives 00 with my Prime 160w charger. I am guessing 09a1021e00 is some sort of error, though it's hard to know what it means. Some other people were experimenting with the 160w charger in #9 and I am hoping that their results with this latest branch might reveal if this is some sort of difference between the 160w charger and 250w chargers or if there is something in the negotiations which is specific to my device which I have not picked out yet.

I do have some things you could do in the meantime if you want, this was made by one of the people playing with the 160w charger and if it at least negotiates successfully with your 250w charger it might tell us if its my implementation or a difference in protocol. I also pushed some changes which added even more logging to make the negotiations clearer and fixed the erroneous IndexError: pop from empty bytearray issue. Also when you are connecting to the device does its screen show a message saying connecting, if so does it hang or does it say connected after a while.

[pkolbus]

Hi @flip-dots,

https://github.com/Hyper-Beast/Anker_Prime_160W_WebBLE isn't working very well for me, I did get it to connect after serving it over https. It displays the MAC address and serial number but no detail about ports and plenty of errors. Here's the log from that, in case useful. This was attempted using the Intel Bluetooth adapter in my laptop rather than my ESPHome BT proxies. AnkerPrimeWebBle.txt

I'll get solix_ble commit f592cf6 installed and capture logs shortly.

[pkolbus]

@flip-dots , here are the logs from solix_ble; I cycled power on the charger at the same time I restarted Home Assistant so that both had the best chance of being clean. As for the display, there's no "connecting" message, when solix_ble is connected there's a Bluetooth icon in the top left corner of the display that replaces a grey "no-signal" icon. No difference from a phone connection as far as I can tell.

button.250w_prime_charger_mqtt_status_request is a ha-anker-solix entity; this integration is disabled for the test; those messages can be ignored but help show the logs were running for a couple of minutes.

home-assistant_solix_ble_2026-03-20T01-35-28.502Z.log

[flip-dots]

@pkolbus ok. I am still waiting to hear back from the other guys if it works with their 160w Prime Chargers. If they don't get back to me I might need packet captures and a dump of the apps data in order to figure out what the difference is, its kind of a huge pain in the arse but there is not a better way.

Capturing the bluetooth traffic is fairly simple, you basically take an ADB bugreport which usually includes a capture of all the Bluetooth traffic. The next part sucks, you have to either have a rooted phone in which case you can just use Frida otherwise you have to unpack the Anker apk, embed Frida, then re-package, re-sign, and re-install the modified Anker app. Then you use Frida to dump the shared preferences using a script, its something best avoided if possible but knowing what a successful handshake looks like is really useful.

[pkolbus]

@flip-dots Thanks for your persistence on this. Capturing a working BLE session makes sense and I'm happy to help; seems appropriate regardless since the 160w and 250w could somehow be different. Do you have a guide for capturing what you want that you can point me at? It's been about a decade since I last used ADB for much of anything.

I have a Lenovo tablet (leftover from some wallmount experiments) that looks like it's rootable; I'll get that going in the meantime, and set up Frida too.