On Sat, 2009-01-31 at 12:35 -0600, CSM(r) Gary Huber wrote:
> I agree with the preference of using hardware flow control if possible. I
> don't remember where but somewhere I read (and learned by experience over 20
> years in my daytime job) that 9600 baud was the limit for software flow
> control on three wire circuits.... something about turn around time in the
> UART of the serial interface. But we digress as in this application there
> isn't enough data coming from the interface at a fast enough rate to fill a
> buffer and drop bits.
>
> Gary - AB9M
>
In software handshaking, it takes a full character time to send STOP! so
the chances of loosing data are high unless the device is smart enough
to yell HALT! with a couple character spaces in its character buffer
which in the old RS-232 chips was only one character deep. Today the
handiest UARTs at the computer end can recognize XON/XOFF and have 16 or
32 character buffers. But serial chips in the days of the OMNI V and VI
design, or microcontroller bit banging serial interfaces may not have
that sophistication. Hence hardware handshaking is more reliable because
its much faster and simpler to detect. Just check the CTS bit before
stuffing the next character into the send buffer and starting the send
clock. And in the radio, clear CTS the microsecond (plus cable
propagation time, about a nanosecond per foot) that the receive
buffer is full, which may well be after receiving only one character
until that controller has parsed that character.
73, Jerry, K0CQ
_______________________________________________
TenTec mailing list
TenTec@contesting.com
http://lists.contesting.com/mailman/listinfo/tentec
|