[net.micro.pc] Running IBM serial boards at 19.2KB

topher@cyb-eng.UUCP (Topher Eliot) (04/19/85)

> The standard IBM serial adapter using the 8250 UART, as well as the
> serial ports on expansion boards such as AST SIXPACK can be programmed
> for 19200 and 38400 bits per second.

Huh?  I quote from page 1-238 of the I.B.M. PC2 Technical Reference manual:

	Note:  The maximum operating frequency of the baud generator is
	3.1MHz.  In no case should the data rate be greater than 9600 baud.

Do you know something IBM doesn't?

Cheers,
Topher Eliot                                           Cyb Systems, Austin, TX
         {gatech,ihnp4,nbires,seismo,ucb-vax}!ut-sally!cyb-eng!topher

robert@gitpyr.UUCP (Robert Viduya) (04/21/85)

> > The standard IBM serial adapter using the 8250 UART, as well as the
> > serial ports on expansion boards such as AST SIXPACK can be programmed
> > for 19200 and 38400 bits per second.
> 
> Huh?  I quote from page 1-238 of the I.B.M. PC2 Technical Reference manual:
> 
> 	Note:  The maximum operating frequency of the baud generator is
> 	3.1MHz.  In no case should the data rate be greater than 9600 baud.
> 

To bring a real-life application into the argument, I've written a comm-
unication program for the PC that can handle 19200 baud with no errors
or dropped characters.  This was using the standard IBM Async card.
Now, of course, it had to have XOn/XOff flow-control to work properly
(just can't write to that screen fast enough), but it was 19200.  You
also can't tell the difference between 9600 baud and 19200 baud.

			robert
-- 
Robert Viduya
Georgia Institute of Technology

...!{akgua,allegra,amd,hplabs,ihnp4,masscomp,ut-ngp}!gatech!gitpyr!robert
...!{rlgvax,sb1,uf-cgrl,unmvax,ut-sally}!gatech!gitpyr!robert

caf@omen.UUCP (Chuck Forsberg WA7KGX) (04/21/85)

>> The standard IBM serial adapter using the 8250 UART, as well as the
>> serial ports on expansion boards such as AST SIXPACK can be programmed
>> for 19200 and 38400 bits per second.
>
>Huh?  I quote from page 1-238 of the I.B.M. PC2 Technical Reference manual:
>
>	Note:  The maximum operating frequency of the baud generator is
>	3.1MHz.  In no case should the data rate be greater than 9600 baud.

Just talk to a user of Professional-YAM (Omen's Comminications program).
YAM works fine at 19200 bps provided there are no memory resident programs
that mess with interrupt latencey.

Compared to 9600 bps, 19k is noticeably crisper and file downloads (especially
when YMODEM protocol 1k packets are used) is much faster.
-- 
Chuck Forsberg WA7KGX	..!tektronix!reed!omen!caf
Omen Technology Inc 17505-V NW Sauvie IS RD Portland OR 97231
Voice: 503-621-3406	Modem: 503-621-3746

nigelh@uvicctr.UUCP (Nigel Horspool) (04/21/85)

> > The standard IBM serial adapter using the 8250 UART, as well as the
> > serial ports on expansion boards such as AST SIXPACK can be programmed
> > for 19200 and 38400 bits per second.
> 
> Huh?  I quote from page 1-238 of the I.B.M. PC2 Technical Reference manual:
> 
> 	Note:  The maximum operating frequency of the baud generator is
> 	3.1MHz.  In no case should the data rate be greater than 9600 baud.
> 
> Do you know something IBM doesn't?
> 

	I have coded my own VT100 terminal emulator program to run under
Coherent Unix on the PC/XT.  My program uses its own assembler-coded 
interrupt routines to handle I/O and, despite what the IBM manual says,
it operates perfectly at 19200 baud.  Possibly the emulator will work
at 38.4Kb too, but I don't have any equipment at home to test this
speed.  According to some (rough) calculations, the XT will have time to
execute about 300 instructions between two interrupts when serving
interrupts produced by a 38.4Kb I/O device.  A carefully coded interrupt
routine should easily be able to keep up without missing any interrupts.

			R. Nigel Horspool
			University of Victoria

			{uw-beaver, nrl-css} ! vax-populi ! nigelh
			vax-populi!nigelh@NRL-CSS.ARPA

tsc2597@acf4.UUCP (Sam Chin) (04/23/85)

<>
I don't know about 8250 UARTS, but with 8251 USARTS you have to worry about
the clock. If you use X16 mode and a 2 Mhz clock. The divisor which you have
to send to the counter (the 8251 has to be driven by an external clock) is
2.0 * 10E6/baud rate * 16 = 6.5
You unfortunately cannote send fractions to the counter and 6 or 7 don't
work. The solution for the 8251 is to use X1 mode in the counter but I don't
know how that would be done for the 8250. The point is that the clock does
matter. In the above example, if the clock was raised to 2.5 Mhz, you would
have a nice (almost) round 8.

                                              Sam Chin
                                              tsc2597.acf4@nyu
                                              allegra!cmcl2!acf4!tsc2597

tsc2597@acf4.UUCP (Sam Chin) (04/23/85)

<>

Another small note. Running above 19200 baud is probably possible but you
would have to do it synchronously.

                                               Sam Chin

vr0z05@unido.UUCP (04/24/85)

I have a filetransfer APPLE II+ <--> IBM PC/XT and it works asynchron
perfectly fine at 19.2 kBaud.

		   Uwe Hoch

cramer@kontron.UUCP (Clayton Cramer) (04/25/85)

> > The standard IBM serial adapter using the 8250 UART, as well as the
> > serial ports on expansion boards such as AST SIXPACK can be programmed
> > for 19200 and 38400 bits per second.
> 
> Huh?  I quote from page 1-238 of the I.B.M. PC2 Technical Reference manual:
> 
> 	Note:  The maximum operating frequency of the baud generator is
> 	3.1MHz.  In no case should the data rate be greater than 9600 baud.
> 
> Do you know something IBM doesn't?
> 
> Cheers,
> Topher Eliot                                           Cyb Systems, Austin, TX
>          {gatech,ihnp4,nbires,seismo,ucb-vax}!ut-sally!cyb-eng!topher

Probably.  I am using an XT running a VT100 terminal emulator at 19,200 baud
right now.

mjg@ecsvax.UUCP (Michael Gingell) (05/03/85)

Warning! Running your IBM PCs serial port at 19200 may be hazardous
to it's health!. I am just passing this on but I am reliably informed
that the reason IBM does not specify serial operation at 19.2 is that
it is under some circumstances possible to burn out the UART chip.
The reason may be that internal currents in MOS curcuits are often
directly related to switching frequency.

This is not to say you can't do it, in fact I have linked my PC to my
Data General One at 19200 baud and got away with it .. but for how long
that will last I don't know. Data General has a program for c
communicating to the PC which works up to 19200 baud but they reccomend
only using at 9600 for this very reason.

Mike Gingell,  ITT Raleigh, NC (919) 878 8100  ..decvax!mcnc!ecsvax!mjg

rap@oliven.UUCP (Robert A. Pease) (05/10/85)

.

> Warning! Running your IBM PCs serial port at 19200 may be hazardous
> to it's health!. I am just passing this on but I am reliably informed
> that the reason IBM does not specify serial operation at 19.2 is that
> it is under some circumstances possible to burn out the UART chip.
> The reason may be that internal currents in MOS curcuits are often
> directly related to switching frequency.
> 
> This is not to say you can't do it, in fact I have linked my PC to my
> Data General One at 19200 baud and got away with it .. but for how long
> that will last I don't know. Data General has a program for c
> communicating to the PC which works up to 19200 baud but they reccomend
> only using at 9600 for this very reason.
> 
> Mike Gingell,  ITT Raleigh, NC (919) 878 8100  ..decvax!mcnc!ecsvax!mjg

Out darned spot!  Point one; the 8250 is perfectly capable of running at
19.2K baud.  I quote from the NS data sheet for the INS8250-B;

	o Fully Programmable Serial-Interface Characteristics
	  --  5-, 6-, 7-, or 8-Bit Characters
	  --  Even, Odd, or No-Parity Bit Generation and Detection
	  --  1-, 1&1/2-, or 2-Stop Bit Generation
	  --  Baud Rate Generation (DC to 56k Baud)
				    ^^^^^^^^^^^^^^

Point  two;  MS-DOS  (and  therefore  PC-DOS)  runs  the  primary  and
secondary async ports !without! interupts.  This means that it is very
easy to lose characters!  Especially if running at higher baud  rates.
This  is  why  it  is recomended not to try serial port communications
above 9600 baud.  My recomendation is that if you want to  go  faster,
write  yourself  an  interupt  driven  RS-232  serial port driver that
accepts XON/XOFF protocols and use it.  Then go knock yourself out!
-- 

					Robert A. Pease
    {hplabs|zehntel|fortune|ios|tolerant|allegra|tymix}!oliveb!oliven!rap