Path: utzoo!mnetor!uunet!cbmvax!grr
From: grr@cbmvax.UUCP (George Robbins)
Newsgroups: comp.sys.amiga
Subject: Re: Nothing to do with Amiga and MIDI
Message-ID: <3070@cbmvax.UUCP>
Date: 3 Jan 88 10:13:44 GMT
References: <8712292114.AA22022@cory.Berkeley.EDU>
Reply-To: grr@cbmvax.UUCP (George Robbins)
Organization: Commodore Technology, West Chester, PA
Lines: 67

In article <8712292114.AA22022@cory.Berkeley.EDU> dillon@CORY.BERKELEY.EDU (Matt Dillon) writes:
> :In article <1325@sugar.UUCP> peter@sugar.UUCP (Peter da Silva) writes:
> :>First of all, conventional printers (a slow device) typically come with
> :>a very fast interface that they can't even begin to take advantage of, but
> :>restrictes them (theoretically) to within 6 feet of the computer.
> :
> :Which is that?  I've operated Centronics parallel interfaces over 50'
> :cables with no problem (although I admit this is stretching it a bit -
> :20' should be no problem whatsoever.)
> 
> 	A couple of things to note here:
> 
> 	(1) Centronics is syncronous, meaning that every byte is handshaked.
> 	    At 50' it will go quite a bit slower and much less reliably than
> 	    at 5'.  However, since printers are notoriously slow anyway one
> 	    usually doesn't see the difference.  A similar unshielded and
> 	    undriven serial cable would not be able to go quite as fast.
> 
> 	(2) Beyond popular belief, serial is NOT slower than parallel!  That's
> 	    right!  Consider that serial is just two lines (not including 
> 	    ground)... very easy to make into a twisted pair, shield, connect
> 	    to power drivers, or place on a carrier.  I've seen serial over
> 	    a cable TV channel at 56KBaud, and we weren't even trying!  Of
> 	    late we've had RS422 which utilizes balanced lines and can go even
> 	    faster than RS232.

EEK!  More confusion in the form of true, but marginally relevant facts.

You can have a "long" Centronics cable.  The length is basically limited
by the characteristics of the driving and receiving circuits, the cable
characteristics, the number of grounds connected and the hardware/software
timing.

Centronics is a parallel interface.  You put out the data, allow time (deskew)
for all the bits to get to the receiving end, and then send a strobe telling
the printer that the data is there.  The printer eats the data and sends back
busy/acknowledge signals.  This is all spec'ed to work over maybe a 10-15 foot
cable. 

It fails if:

a) the deskew delay isn't long enough for the data to have settled
b) the strobe or ack pulses are degraded by the cable characteristics
c) there is a grounding problem between the two machines
d) there is electrical noise or ringing on the data or strobe/ack lines

"Real Computers" offer special "long lines" interfaces with appropriate
timing, circuitry and cabling to avoid these problem.  Micro users get
the generic flavor and your milage may vary.

As for whether serial or parallel is faster, remember than the following
crude approximation holds:  characters/sec = lines/minute = bits-second/10.
For the typical serial printer (with a line buffer) anything over 2400 bps
is likely to be wasted.  For laser printers, especically in bit gulping
graphics modes, it's a different story, but remember that your computer
isn't a CRAY-7 and a 19.2Kbps serial line is probably quicker than your
computer can generate (although perhaps not dump from a file) real data.

Since the "centronics" interface is shoving 8 bits in parallel, you could
assume a 10 microsecond best case output loop, giving the equivalent of
a 1-megaBPS async serial line, but as pointed out above, this is pretty
much academic...

-- 
George Robbins - now working for,	uucp: {uunet|ihnp4|rutgers}!cbmvax!grr
but no way officially representing	arpa: cbmvax!grr@uunet.uu.net
Commodore, Engineering Department	fone: 215-431-9255 (only by moonlite)