On Mon, 3 Jun 2019, David Ranch wrote:
>/(6) A RTTY, data or multiplexed emission using a specified digital code />
>/listed in ?97.309(a) of this part may be transmitted. The symbol rate
must not
/> >/exceed 56 kilobauds. A RTTY, data or multiplexed emission using an
unspecified /> >/digital code under the limitations listed in ?97.309(b) of this
part
also may /> >/be transmitted. The authorized bandwidth is 100 kHz. /
56 kilobauds is because the Telebit Trailblazer
existed; it used OFDM
and 6 "baud" carriers.
The Trailblazer was fast, but it wasn't *that* fast!
It actually was like 19200 baud async, internally some 18kbps sync.
This was fast compared to the usual V.22bis (2400 bps) modems of the day, but of
course later telephone modems appeared that could to 56k.
We ran a Unix system with UUCP at work in those days but we could not use modems that
fast
because the machine wasn't able to handle full 19200 baud async I/O as it could not
handle the interrupt load... those were the days.
(It was an NCR system with a separate processor for serial I/O, but IBM PC and clones
in those days also struggled at 19200, that is why the 16550 chip with hardware FIFO
was created...)
A couple of years later I bought a ZyXEL U-1496+ modem that also did 19200 over phone
lines, and was popular at some local ISPs. Like the Trailblazer it also used a 68000
processor and it cost well over $500 in local currency.
I agree that those modems are not going to work over radio. Even on telephone lines
they perform lengthy "training" sequences to measure the phase response of the
line,
and then assume it remains the same over the duration of the call.
Rob