Suggestion (#1) for a small change in line 205 ('section' #PINGRESP in function wait_msg):

return res[0] instead of return None

This allows you to check in your own consumer code that called check_msg() or wait_msg(), that the ping response came by and was valid, e.g.: if(oReply == 0xD0): # valid ping response

And suggestion (#2) for another small change in line 122 (section 'ping') and line 118 (section 'disconnect'):

add line self.sock.setblocking(False) in front of self.sock.write(b"...")

This prevents lóóóng (default 30 secs?) lock-ups when the connection is lost (or not ready yet).
I haven't experienced a down-side yet of this non-blocking write (with and without a working connection).

P.s. I guess the problem is only valid when using SSL, as the (plain) socket is otherwise already initiated with specified timeout at 'connect' and may not stall that long. Once wrapped in an SSL socket though, the original timeout seems either lost or ineffective. Attempting to use settimeout(t) instead of setblocking(False) here before writing to the socket, is currently not possible with SSL sockets in micropython (see this request)

Thanks!