NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/sys/cbm/2210

Internet Message Format | 1996-08-05 | 1.9 KB

Path: bcarh8ab.bnr.ca!mwandel
From: [e-mail] (Markus Wandel)
Newsgroups: comp.sys.cbm
Subject: Re: C= RESTORE
Date: 10 Feb 1996 02:32:45 GMT
Organization: Bell-Northern Research Ltd.
Message-ID: <4fh04d$[e-mail]>
References: <[e-mail]> <4ffsk2$[e-mail]>
Reply-To: [e-mail]
NNTP-Posting-Host: bcars758.bnr.ca

I just had a look at the schematics, and the RESTORE key is capacitively
coupled to the chip it drives.  What this means is that the pulse from the
key must have a fast rise time for the signal to be recognized.  And since
the keys have rubber contacts, the required fast rise time occurs if the key
is struck fairly hard, but not if it is tapped normally.

(OK, OK, before everybody jumps on me:  The RESTORE key generates a low-going
pulse.  But the technical term is still "rise time."  The actual circuit is
1M from RESTORE key to +5, then 51pF cap, then 470K to +5, then 556 trigger
input.  The 556 turns the short spike that makes it through the capacitor
back into a nicely timed NMI pulse.)

This is a feature.  The RESTORE key generates a non-maskable interrupt.
What this can do is screw up anything timing sensitive the machine may be
doing at the time, like a transfer on the serial I/O bus or RS232 interface.
It may crash the machine if the Kernel ROM is currently swapped out.  It
will restart a game cartridge image.  And so on.  You can't mask a NMI.
And with the key right next to the RETURN key, I guess they just wanted to
reduce the possibility of NMIs being generated by mistake.

I always thought this was an excellent feature.  RUN-STOP/RESTORE gets
out of all kinds of hung computer scenarios, and bashing RESTORE fairly
hard just kind of made sense.

--
Markus Wandel     Ottawa Ont. Canada     [phone]
[e-mail]  <-- NOT '[e-mail]' (that's for work only)

*** DISCLAIMER ***  Not speaking for or representing my employer in any way.