633 firmware 1.5 flash problems

On firmware 1.5, a write to user flash following a read from user flash has difficulties. This is reproducible under the packet debugger. If you send command 3, wait for its ACK, then send command 2, an ACK will for command 2 will never arrive.

The panel responds with 21 bytes, the first of which is garbage (0x2e), followed by 20 bytes that look like an ACK to command 3, except the CRC is invalid.

It takes several resends of command 3 before a valid ACK is received.

Also, while attempting to find a workaround, the panel went into a state (after writing to user flash) in which no commands would have any effect.

We have some CFA633 with v1.3, and they do not show this misbehavior. Is this a known bug in the 1.5 firmware?

I'll check into it.

I was able to reproduce the trouble on this end.

The currently shipping code does not have this trouble.

It does seem to save the data OK, then starts having trouble after the write. For instance, sending the write command, then disconecting and reconnecting power leaves the module operating OK, with the write intact.

It may be possible to send a reboot command (or several, until one is acknowledged) after the write, the reboot will restore the firmware to a stable state.