I disassembled a discrete PAL NASA famiclone. Inside, there were these two chips:
CPU:
Markings on top:
Quote:
USC
2A03E
9136S (made week 36 year 1991)
919600
Markings on the bottom:
Quote:
SH 33
PPU:
Markings on top:
Quote:
USC
2A02E
9136S (made week 36 year 1991)
919610
Marking on bottom:
Quote:
SG 51
As you may see, they used a 2A03. Is this possible, or they just didn't care about making a different package for the PAL version?
I no longer own this thing, so I don't even know if it was really a PAL console (every TV in my house supports NTSC, PAL and SECAM, also I didn't have a original NES to compare the games' speed)
PS: A curiosity: this exact model cost about ~70USD and was sold even in large malls along with original NES consoles:
http://imageshack.us/photo/my-images/32/nasaabc.png/
I doubt they used a true 2A03, because only Nintendo had acess to them. Unless they corrupted sharp or something in the like but I doubt it's a possibility. However it is much more likely they just asked the manufacturer to write 2A03 on their own chip.
In theory a 2A03 would work with any frequencies, it would just make it's synchronisation with the PPU different to the usual "1/3" one used for NTSC (and also different to the "5/16" one used for PAL). I think only inner difference between the 2A03 and the 2A07 is the clock divider, the Square/Noise channel's length tables and the noise/DPCM pitch tables.
It's a fake chip, latest E revision chip i saw dated 86(87). 88 going with G revision. you can check if it is real "E" - pull up pin 30(usually grounded) will stop(HALT?) real "E" revision CPU
And what would happen if it's different than E?
Original "G" keep running, while original "E" stops. Didn't check original "H" yet and have no idea about "F" - never got them. Not sure about original revisions before "E" - earliest chip, i had, was dated beginning of 85 and had revision "E"
I'll add 'put a socket to my NES' on the todo list to test that fake 2A03.
On the original motherboard (which is very damaged) it's tied to ground
In famikon pin 30 grounded as well and function of this pin is not documented well. If pin 30 working as /SndDisable(disable sound ports and joypad ports, so 2A03 becomes 6502 without decimal mode), then why "E" stops?
In Nintendo's NES schematics, CPU pin 30 is labelled as REDY (for original front-loader) or TEST (for top-loader with rev G/H CPU).
mark_k wrote:
In Nintendo's NES schematics, CPU pin 30 is labelled as REDY (for original front-loader) or TEST (for top-loader with rev G/H CPU).
Do you have schematics for the Top-loader NES? Is pin 30 grounded on the top loader?
Yes it's grounded in both consoles, just the pin name is shown differently in the two schematics.
mark_k wrote:
Yes it's grounded in both consoles, just the pin name is shown differently in the two schematics.
Do you have a copy of the Top Loader Schematic? Or a link to one?
mark_k wrote:
In Nintendo's NES schematics, CPU pin 30 is labelled as REDY (for original front-loader) or TEST (for top-loader with rev G/H CPU).
Oh, that's interesting. Looks like rev E have this pin for DMA or BUSRQ, but still unclear about "TEST" function
For a rev G/H CPU, you could try lifting CPU pin 30 and pulling it high. (That is, connect pin 30 via a resistor to Vcc.)
It may be that earlier CPUs had an internal pull-up resistor on pin 30, and that was removed in later revisions.