I’ve been trying to analyze how do certain parameters of serial communication (for instance parity, flow control etc.) affect the real transfer speed of RS-232. However, I’m not sure how to judge whether the real speed in the current situation is above or under expectancy.

Let’s say we have two measurements. One transfers a binary file of 0’s using YMODEM with a Baud Rate of 19 200 b/s and the other tranfers the same file using Kermit with Baud Rate 115 200 b/s. The first situation reached a real transfer speed of 17 720 b/s and the one with Kermit reached 420 000 b/s. My question is – if the real speed goes below or over the baud rate, does it confirm that speed really depends on the communication protocol for good or worse? Also what I mean by baud rate is the choice “Bits per second” in HyperTerminal emulation program.

What I want to make is a table full of “Over the standard” or “Under the standard” instead of speed values so that I can use a chi-squared test on the results to confirm their independence or dependence.

Thanks in advance.

I’ve already tried to add a cell to my table of measurements that writes “Over the standard” if the real speed goes over atleast 90 % of the Baud Rate (since speed has a tendency of being quite unstable sometimes) and in other case it writes “Under the standard”. However, using a chi-squared test on this approach actually proved to me that transfer speed shouldn’t depend on parity although I was expecting to prove that using parity actually decreases the real speed value.