While investigating this problem [url]https://forum.sierrawireless.com/t/uart-bug-in-openat-7-x/4764/1], we’ve been told by our local distributer that OpenAT sets CTS inactive when there is only one byte left in the USART receive buffer. Can anyone confirm this? It sounds very odd since most units communicating via USART have a send buffer and sometimes even a pipeline architecture which makes one byte notice almost impossible to handle. All other equipment seems to deassert CTS with som margin left in the buffer. And I can’t really se any point for OpenAT to wait until the last byte.
Indeed.
It is a poor implementation that relies upon CTS being noticed with such short notice!
awneil: It’s almost to poor to be true 
Can you confirm that it is in fact one byte?
V.43 specifically notes that the DTE may take a “considerable” time to recognise circuit 106 (CTS) going “off”
For a character-based link, “some thousand octets” would be excessive - but just one is clearly insufficient!
(no, I can’t confirm if it is actually 1).