Re: [Tomy Tutor ] My new website
> > The 9995 muxes the databus always. The 9995 is the indirect descendantof
> > the 9985 that was supposed to be in the 99/4 -- which instead got a 9900of
> > kludged to work similarly to the 9985. The 99/8 uses a special version
> > the 9995 that disables most of the on-chip RAM.muxed
> Explain this. Is there another "$8300 page" RAM out there in the 99/8?
> Is that page connected by the same 16-bit wide path as the 9900's, or
> in the same way as the rest of memory access? And what's most of the RAM?The off-the-shelf 9995 has 256 bytes of on-chip RAM at >F000-F0F9 and
>FFFA-FFFF that is on the 16-bit internal bus. The word at >FFFA is theon-chip decrementer and >FFFC-FFFF is the NMI (non-maskable interrupt, or
Load, vector). From what I understand, having never studied the 99/8 in
detail, it has the on-chip RAM at >F000-F0F9 disabled. The Myarc Geneve
9640 uses a standard 9995 and 8K memory pages. Because the 9995 does not
disable the external bus during writes, accessing a page mapped to
>E000-FFFF in the areas >F000-F0F9 and >FFFA-FFFF will mess with even bytesin the page. Internally a 16-bit operation is happenning, but only the even
data byte shows on the external bus.
With regard to the "$8300 page" RAM: This is the CPU PAD of the 99/4A and
the only directly accessible CPU RAM in a stock console. The standard 16K
RAM of the 99/4A is indirectly accessed via the VDP, the TMS9918A in North
American consoles. Much of the console ROM and GROM and peripheral ROM
(DSR's) is written such that the CPU PAD can be relocated.
E.g., the disk controllers typically report error codes at >8350. In the TI
floppy disk controller source code, that error code may have been written
with an instruction such as "MOV @ERR12,@>0050(R9)". The workspace register
R9 in the case of the 99/4A contains >8300. The 99/8 could have its CPU PAD