[comp.sys.amiga.tech] 2 meg Agnus in Amiga 2000

hawk@pnet01.cts.com (John Anderson) (08/25/90)

   I've been following the discussions (or lack there of) on hacking the 2000
a bit to get the 2 meg agnus to work in it.  Cost is of no importance so even
if the PLCC (or something like that) piggy back sockets are $120 each, that
doesn't matter.  Dave Haynie, if you have any info that I might need then
please let me know.  I have the schematics of the 2000 and the 1 meg agnus in
my B2000.  If I can get a copy of the 3000 schematics, will that be enough
info to convert the pins from the 1 meg agnus to the way they are supposed to
be on the 2 meg version?  A quick question (or two): are there the same number
of pins on the 1 meg agnus that are on the 2 meg agnus, and if yes then: what
happens to the lines that are converted?  If anybody else has any information
that can be of help then please let me know.  Thanks ahead of time.

uzun@pnet01.cts.com (Roger Uzun) (08/25/90)

I believe some signals that were internal to the old agnus have been brought
out or vice versa, so it was said that additional hardware would be required,
you cannot just redirect the pinout of the new agnus to sockets on the old
(2000) motherboard, I imagine you have to cut a trace on the motherboard
to get at some/insert some required signal.

-Roger

UUCP: {hplabs!hp-sdd ucsd nosc}!crash!pnet01!uzun
ARPA: crash!pnet01!uzun@nosc.mil
INET: uzun@pnet01.cts.com

ppessi@niksula.hut.fi (Pekka Pessi) (08/25/90)

>>>>> On 25 Aug 90 05:16:09 GMT, uzun@pnet01.cts.com (Roger Uzun) said:

> I believe some signals that were internal to the old agnus have been brought
> out or vice versa, so it was said that additional hardware would be required,
> you cannot just redirect the pinout of the new agnus to sockets on the old
> (2000) motherboard, I imagine you have to cut a trace on the motherboard
> to get at some/insert some required signal.

	I think that XCLK signal can be ignored if you do not use
	Genlock. Additional circuitry in A3000 consists of 2 PALs and
	3 FFs, but I think it can be reduced to a couple TTL:s if you
	do not want 32-bit interface to CHIP-memory.

--
ppessi@niksula.hut.fi !		M{ kaunehin oon
tel +358 0 455 4757    	!		ja tied{nkin sen
JMT 11E122;  02150 Espoo   !			ja siks' m{ Mua aina aattelen
							- Pekka Pessi

d6b@psuecl.bitnet (08/26/90)

In article <4093@crash.cts.com>, hawk@pnet01.cts.com (John Anderson) writes:
>    I've been following the discussions (or lack there of) on hacking the 2000
> a bit to get the 2 meg agnus to work in it.  Cost is of no importance so even
> if the PLCC (or something like that) piggy back sockets are $120 each, that
> doesn't matter.  Dave Haynie, if you have any info that I might need then
> please let me know.  I have the schematics of the 2000 and the 1 meg agnus in
> my B2000.  If I can get a copy of the 3000 schematics, will that be enough
> info to convert the pins from the 1 meg agnus to the way they are supposed to
> be on the 2 meg version?  A quick question (or two): are there the same number
> of pins on the 1 meg agnus that are on the 2 meg agnus, and if yes then: what
> happens to the lines that are converted?  If anybody else has any information
> that can be of help then please let me know.  Thanks ahead of time.

I don't have detailed information, but I've picked up some general info.
RAS1* no longer exists. A "MA9" (or something like that) has been
added. In other words, the "natural" configuration for the 2MB Agnus is
a single bank of 1M by 1 DRAMs (the 3000 has extra logic to control its
two banks -- needed to get 32 bits out at a time). Also, another address
line is conected (A20). That presents a problem: one more pin is used
than with the 1MB Agnus. As a person who knows nothing about chip design,
I'd say they should have ditched one of those redundant power/ground
lines, but I guess their layout requires those. To get the extra pin
(2 actually; I guess one is unused) they ditched the XCLK logic used
by Genlocks (specifically, XCLK and XCLKEN*). It's implemented by some
rather trivial external logic on the 3000 (and of course the 1000 as well).

Hope this helps a little...

-- Dan Babcock
Note my new addresses...
Mail ("forever"): P.O. Box 1532 / Southgate, MI  48195
Mail (school): Atherton Hall Room 63 / University Park, PA  16802
Internet: D6B@ECL.PSU.Edu
Bitnet:   D6B@PSUECL.Bitnet
          (capitalization is probably wrong)
Voice: (814)-862-2931