Path: utzoo!mnetor!uunet!tektronix!tekgen!tekigm2!phils
From: phils@tekigm2.TEK.COM (Philip E Staub)
Newsgroups: comp.sys.amiga
Subject: Re: hardware question on DTACK'
Message-ID: <2487@tekigm2.TEK.COM>
Date: 6 Feb 88 01:09:48 GMT
References: <1107@louie.udel.EDU>
Reply-To: phils@tekigm2.UUCP (Philip E Staub)
Organization: Tektronix, Inc., Beaverton, OR.
Lines: 33

In article <1107@louie.udel.EDU> rminnich@udel.EDU (Ron Minnich) writes:
>I am trying to add some memory to an A1000. For several reasons
>the hardware is not asserting DTACK. This brings up a question:
>something else is! Is there a watchdog timer in the custom chips that
>automatically asserts DTACK after x cycles? I guess there would have to
>be, else if you addressed non-existent memory the thing would hang. 
>Can anyone fill me in? The hardware manual is not helpful in this case.
>I am hoping to hear that DTACK is not just asserted always by the hardware 
>as this memory cycles in 675 ns (i.e. slowly) so i really need DTACK
>to work right.
>
>-- 
>ron (rminnich@udel.edu)

Ah, but the hardware manual *is* helpful. You just have to know what to look
for. I don't know the exact page, but what happens is that that Amiga is
*not* expecting to have DTACK asserted by any expansion device. DTACK is
generated internally, unless you sit on a busy line (I think it's XBUSY, or
something like that, no manual handy) to say you want to stretch the bus
cycle. There is also a spec as to how soon after the bus cycle begins that
you have to assert the busy line. This looks (at least to me) like shades of
Intel processors(!), which assume that bus cycles have completed
successfully unless your hardware sits on a bus error or wait line to say
that it failed or needs to be stretched.


Hope this helps.
Phil
-- 
------------------------------------------------------------------------------
Phil Staub                     "I do NOT approve. I merely said I UNDERSTAND."
tektronix!tekigm2!phils                                              - Spock
phils@tekigm2.TEK.COM