[comp.sys.att] 19200 baud on the Unix PC serial port

wtm@neoucom.UUCP (Bill Mayhew) (12/24/89)

Hi,

Happy holidays.


Now that we have that out of the way, yes indeed, the serial port
on the Unix PC and/or 7300 is capable of running at 19200 bits per
second.  Be aware that 19200 is not supported by AT&T, so if it
doesn't work for you....

If you have the basic uucp and cu that comes with the machine, try
specifying "exta" for the baud rate fields in L.sys and L-devices;
that should work.  "19200" is recognized directly by the Honey
Dan Ber Basic Networking Utilities package, if you use that for
uucp.

There could be got-chas in using 19200.  The expansion slots have
differening interrupt priorities.  I don't have the manuals, so I
can't elaborate on the subject.  You may have to try the EIA Combo
board in different slots to see which (if any) work better.  I
noticed that my uucp xferstats via tty000 decreased when I had a
Voice Power board in the middle slot and ran the voice daemon.  I
didn't need the voice daemon for my voice application, so just
bagged the daemon.

The trailblazer will yield about 11000-12000 bits per second when
used with the Unix PC, so it is definitely worth the while to
experiment some to seem if running the port at 19200 will work with
your given machine and sofware configuration.  I've benn using a
Trailblazer since December 1987 without any porblems, though your
results may vary.

The absolute limit on thoughput with the tty000 port set to 19200
is about 14000.  14000 seems to be the limit that characters can be
stuffed into the port at a sustained rate.  This doesn't
necessarily mean thta you can pull characters into the port at
that rate too.  I experimented with uucp e protocol and found that
I started to lose characters on input with the port running as fast
as it would go on big files.  Best to stick with the default g
protocol.


--Bill

flinton@eagle.wesleyan.edu (12/29/89)

In article <1861@neoucom.UUCP>, wtm@neoucom.UUCP (Bill Mayhew) writes:
> 
> If you have the basic uucp and cu that comes with the machine, try
> specifying "exta" for the baud rate fields in L.sys and L-devices;
>
Sorry -- should that  "exta"  be an  "extra" ?  Or is it right as is?
                                         ^         
---
Fred E.J. Linton  Wesleyan U. Math. Dept.  649 Sci. Tower  Middletown, CT 06457
---
ARPA/Internet:  FLINTON@eagle.Wesleyan.EDU           (preferred)
Bitnet:         FLINTON@WESLEYAN[.bitnet]     (also works)
on ATT-Mail:    !fejlinton                     ( ...!attmail!fejlinton )
Tel.:           + 1 203 776 2210 (home)  OR  + 1 203 347 9411 x2249 (work)

pat@rwing.UUCP (Pat Myrto) (12/30/89)

In article <5272@eagle.wesleyan.edu>, flinton@eagle.wesleyan.edu writes:
>
> Sorry -- should that  "exta"  be an  "extra" ?  Or is it right as is?

The flag `exta' is an abbreviation for "External A", and the spelling
"exta" is correct.  Note also, if getty is to be used at 19200 on a
port (assuming the /etc/getty supplied with the system) the baudrate
used in a gettydefs entry is "EXTA", instead of "B19200", the latter
giving an "Undefined:  B19200" error.  This is in spite of the fact
that the distributed gettydefs has an entry for 19200 using the B19200
flag.  As always, copy gettydefs to a work copy, edit on that, and do
a `getty -c' on that copy before making it the real file, to avoid
problems.

I hope this clears up some confusion...
-- 
pat@rwing                                       (Pat Myrto),  Seattle, WA
                     ...!uunet!pilchuck!rwing!pat
             ...!uw-beaver!sumax!polari!/
WISDOM:    "Travelling unarmed is like boating without a life jacket" 

wtm@neoucom.UUCP (Bill Mayhew) (12/31/89)

Actually, the token for 19200 bps is "EXTA".  I believe that comes
from the DRV-11 (or something similar) on the PDP-11 that supported
two _external_ clock rates, which were supposed to have been
supplied from a clock generator in the modem.  The DRV-11 is a
single line serial RS-232C or current loop interface.

If you look in /usr/include/sys/termio.h, you'll see both EXTA and
EXTB defined.  EXTB is not suppoted by the kernel, according to a
comment in termio.h; if EXTB worked, it would be for 38400 bps.


Bill

mdapoz@hybrid.UUCP (Mark Dapoz) (01/03/90)

In article <1865@neoucom.UUCP> wtm@neoucom.UUCP (Bill Mayhew) writes:
>If you look in /usr/include/sys/termio.h, you'll see both EXTA and
>EXTB defined.  EXTB is not suppoted by the kernel, according to a
>comment in termio.h; if EXTB worked, it would be for 38400 bps.

No, termio.h says B38400 is not supported, not that EXTB isn't supported.  I've
looked at the schematics for both the serial port on the motherboard and for
the combo card and both of them have circuitry there to allow the uart to
accept an external clock signal.  The transmit clock is fed from pin 15 on the
DB25 and the receive clock is on pin 17.  I was thinking of trying to generate
a clock signal for 38.4K baud which I could feed into these pins, but then I
realized that I don't know what clock frequency I need since I don't know what
constant the kernel used in the 8530 baud rate generator when set to EXTB.
Has anyone ever used these external clock pins before?  Anyone know what time
constant is used for EXTB?
-- 
  Mark Dapoz  (mdapoz@hybrid.UUCP)  ...uunet!mnetor!hybrid!mdapoz

I remind you that humans are only a tiny minority in this galaxy.
	   -- Spock, "The Apple," stardate 3715.6.