Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxn!ihnp4!houxm!hjuxa!petsd!peora!jer
From: jer@peora.UUCP (J. Eric Roskos)
Newsgroups: net.micro.mac
Subject: Re: Delphi Mac Digest V2 #20
Message-ID: <2193@peora.UUCP>
Date: Thu, 29-May-86 08:21:15 EDT
Article-I.D.: peora.2193
Posted: Thu May 29 08:21:15 1986
Date-Received: Fri, 30-May-86 20:27:11 EDT
References: <5037@topaz.RUTGERS.EDU>
Organization: Concurrent Computer Corporation, Orlando, Fl
Lines: 48
Keywords: Delphi

> To: mazur@harvard.HARVARD.EDU (Eric Mazur)
> Subject: Interrupt switch problems on the MacPlus
>
> The freeze upon pressing the Interrupt button will most likely occur when
> ANY powered-on modem is connected; I'd guess the more precise condition is
> any device asserting the CTS signal (HSKi pin).  The cause is a ROM bug.
>
> ------------------------------
>
> From: LOFTUSBECKER (8391)
> Subject: RE: re: Interrupt switch problems on the (Re: Msg 8389)
> Date: 22-MAY 06:43 MUGS Online
>
>         But only "most likely".  I just hit interrupt after reading your
> message, got the debugger, pressed G, and was back on -- with full mouse
> control.  Using a USRobotics Password modem.


This is *not* a feature of the USR Password... it's because the modem was
still online when you pressed the interrupt switch.  If carrier is
present, there should be no problem... it's only if a modem is on and
*negating* the CTS signal (not asserting it).  When no modem is connected,
or when it's off, the pin "floats" to CTS.  When a modem is on but no
carrier is present, it is driven to /CTS, causing the "hardware flow
control" in the serial driver to wait for CTS (which it is supposed to do;
it is not a "ROM bug" at all).

The exact same situation happens if you open the serial port on a plain
512K Mac with the old ROMs and then try to write to it, without disabling
hardware flow control.  Maybe the debugger should have assumed no flow
control for its initial state.

[As a side note, actually the CTS pin on the external connector has to be
driven negative before any change occurs on the output of the line
receiver... back when I was designing my MIDI interface for use with DMCS
(described a month ago in net.music.synth) it took me a good while to
figure that out.  Of course, that follows when you think about all the
above, but when you are using the pin to clock the UART (its other
function) it's easy to forget... so if you use the pin for clocking, the
clock pulses have to go negative with respect to the ground pin for them
to do anything.]
-- 
E. Roskos

"The ugliest tag we've ever put on a car"... "not only was the Governor em-
barassed, he was flabbergasted"... "I don't want to put something that looks
this ugly on my car." -- Florida legislators commenting on the first sample
of the license tag they selected after four months of debate.