A3000/4000 1MB ROM Mod

[SpeedGeek]

A3000/4000 1MB ROM MOD REV 10

OS3.9 BB2-BB4 has outgrown the standard 512KB ROM support provided
by Fat Gary. (Standard 512KB ROM's use $F80000-FFFFFF). After a
previous attempt to modify Fat Gary's address decoding to enable ROM's
@ $F00000-F7FFFF failed a new approach to this problem was required! The problem was solved with a new GAL chip which offers the following features:

- Enables 512KB of ROM @ $E00000-E7FFFF
- Generates ROMEN and STERM signals to run a bus cycle
independent of Fat Gary!
- ROM's only enabled on READ cycles. WRITE cycles to this address
space will be terminated via STERM for faster cycle termination
- Runs a 3 or 4 clock cycle (depending on Speed jumper enable) on
STERM output to support 100 ns or 150 ns ROM's
- Runs a 5 or 6 clock cycle (depending on Speed jumper enable) on
DSTERM output to support 200 ns ROM's or overclocked motherboards!
- NEW experimental version supports 4MB of ROM @ $01000000-013FFFFF

A schematic showing the GAL chip pinout is provided. The GAL chip
must be programmed with the .jed file provided. 21 jumpers must be
connected from the GAL chip to the motherboard or optional A3K and A4K PCBs can be used. A DIP24 or PLCC28 socket is recommended to allow easy programming changes to be made. Jumpers to the motherboard should be kept as short as possible. Note: The A3000 ROM's are not connected to A19! So, you will need a jumper from Fat Gary's A19 to pin 1 of the A3000 ROM's. Enjoy!

P.S. Following Amiga tradition the name "Romy" was chosen for the new GAL Chip! Your comments are welcome!

NOTE:
A4000T owners should see Post #2 for A4K Romy solutions.

[SpeedGeek]

** NEWS UPDATE **

Romy Rev. 10 has been released!

Here is some revision update info:

- New state machine logic replaces counter + flip-flop logic
- REG2 was eliminated and is now a free I/O pin
- STERM and DSTERM outputs are asserted for 2 clocks to provide more reliable operation on timing modified systems
- 4MB version address range moved to $01000000-$013FFFFF

** 2ND NEWS UPDATE **

The faulty A4000T previously mentioned finally got fixed and we spent some time working on A4K Romy solutions. As I suspected, it was in fact the NCR SCSI chip kludge causing these problems!

The work around for this A4000T design bug now requires the Romy Rev. 11 GAL update attached to this post and the new P652 update found here:

http://eab.abime.net/showthread.php?...05&#post778105

NOTE: Romy Rev. 11 asserts /STERM for only 1 clock and therefore Rev. 10 is still recommended and more reliable for A3000D/T and A4000D.