Previously the same byte was read for both the error code and the first byte of the checksum, which was definitely incorrect. The response is updated with an extra byte which (hopefully) provides a more accurate test case.
