Xref: utzoo comp.sys.mac.hardware:8327 comp.sys.mac.programmer:21359
Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!lll-winken!uwm.edu!ux1.cso.uiuc.edu!csrd.uiuc.edu!s28.csrd.uiuc.edu!beckmann
From: beckmann@s28.csrd.uiuc.edu (Carl J. Beckmann)
Newsgroups: comp.sys.mac.hardware,comp.sys.mac.programmer
Subject: Re: How fast can the serial ports go?
Keywords: maximum throughput
Message-ID: <1991Feb5.015757.29000@csrd.uiuc.edu>
Date: 5 Feb 91 01:57:57 GMT
References: <1991Feb2.062244.3235@nntp-server.caltech.edu>
Sender: news@csrd.uiuc.edu (news)
Organization: UIUC Center for Supercomputing Research and Development
Lines: 29

palmer@nntp-server.caltech.edu (David Palmer) writes:

>I'd like to build some hardware that stuffs data as fast as possible
>into the serial port of the Mac.

>I know the tricks used in MacRecorder:
>	1) External clock, with a clock rate equal to the bit rate
>	2) turn off interrupts
>	3) access the serial port hardware directly
>	4) tight software loop

>Given that, how fast can you run the serial ports on the various
>models of Mac?


The Z8530 serial chip ought to be able to go pretty fast in synchronous
mode.  Don't quote me because I don't have my Z8530 manual handy, but it'll
probably do at least 1 M bit/s.  And if you access the chip directly with
a very tight loop, the 68000 (020, etc.) ought to have no trouble keeping
up.  The REAL bottleneck is probably the serial drivers and receivers.
I don't know what their limits are, but "Inside MacIntosh" (volume II?)
has a diagram of the serial ports showing RC filtering on those ports.
You could take the RC time constant, multiply by 10, and invert to get a
ballpark estimate.

By the way, do you have any technical info on MacRecorder (i.e. how to
write your own software to access it directly)?  Thanks in advance.

-Carl                beckmann@csrd.uiuc.edu