Write command to aquarea fails

[tsybolek]

Hello,
First of all, thank you for this project, especially for the decrypted serial protocol, really great job.
I've been using an ESP32 board as an aquarea logger (connected to CN-CNT) for years, with my own code, and it works perfectly. That is, sending the magic packet and getting the reply from the heat pump is just fine.
However, when I try to write a command, like one of those examples from the decrypt page (setting DHW temperature, for instance), I just get a bunch of zero bytes as an answer, and the heat pump does not acknowledge the command, the setting remains the same.
Does anybody have an idea what is wrong?
Thanks a lot,
Oleksiy

[IgorYbema]

Should work as you described. The answer should be the same as a regular request (and first answer doesn't containt the change yet).
Nothing special to it. Or do you have the CZ-TAW1 running together also?

What is the byte string you tested?

[tsybolek]

No, I don't have CZ-TAW1.

[MiG-41]

Also , when HP send something after this command , means ,that command was ok . If command would be damaged ( i.e. CRC bad) , HP will not react ....
So question where you get this zeros, on sending to HP , or HP it self answes with zeros ...

[tsybolek]

Zeros come over the serial immediately after the command has been sent. The function for communicating with the pump is basically the same both for the query (0x71 header) and command (0xf1 header). Immediateny avter the message has been sent to the HP, the code listenes at the serial buffer for the data to come. In case of the query, everything works fine, the answer is normal (starting with 0x71 and containing correct HP data afterwards). In case of the command, I det only zeroes.
I will strip my code to the simplest version that still repordices the error and post it here, along with the debug output.

[tsybolek]

I got it working now. The zeroes were indeed coming not from the serial. HP was just not responding to the commands at all, and since the byte array for holding the HP data was filled with zeroes at the initialization, they were just not overwritten by the serial output. In the meantime, the HP accepts at least full-length commands incl. precalculated checksum (as on the decrypt page). Still not clear why the on-the-fly check-sum calculation failed (because it worked fine with the query message), but this is certainly an issue with my own code. Thanks once more for your suggestions!