Misadventures in Genesis modding

[GiGaBiTe]

I've been wanting to try my hand at modding a Genesis for a good while now, and have pulled some 256kbit SRAMs from an early Pentium board awhile back, so I decided to try my hand at replacing the RAM on the Genny.

I've had this POS genesis console for awhile now that has a bad RF box and is more or less useless because it's darned near impossible to find that 9 pin DIN connector cable. This was my candidate for modding since I have two other genesis consoles that work properly. After buying a solder sucker, I started work on the long process.

I've had plenty of experience soldering before, but I must say, this was a pain in the ass. On the first (any only one I replaced so far), the RAM chip cracked the second I put the tip of the soldering iron on one of the pins to melt the solder. I was following the 3 second rule, so I know there wasn't too much heat applied.

Anyway, since the chip was damaged and probably dead, I used a flat head and wedged it under the chip and slowly pried it out and desoldered the remaining pins still stuck in the PCB.

This was more or less unwise because the thing exploded everywhere and left the pile you can see in the pic below (all that I could find of it.)



The un-exploded chip on the left is the new RAM from the Pentium board. By this time, I had already put the other chip on the board. After powering the board up and expecting failure, it actually worked as expected. Since I couldn't see any video output, I used headphones and listened for sound, which worked.

The workstation (yes, it's my bathroom, and the only place in the house with a decent surface to work on)





My ultimate goal is to upgrade the RAM from 64k to 128k (using a 74HC00 x 2) with 4 of the 15ns SRAMs. I know that games won't see it, and possibly malfunction, but there's been something I've been wanting to test for a good while now. I think I'm also going to order some DIPP sockets so I can easily remove the RAM and not take an hour per chip with the possibility of breaking it.

[jlf65]

There's space for 2MB of RAM in the address map, it's just aliased over the 64 KB. I'd love to boost mine to 2M... that would make the Genny much better for homebrew. The thing that makes the Genesis hard to work with is the lack of RAM. Even with a 32X, you only have 256 KB to work with.

[GiGaBiTe]

I'd love to upgrade it further, but I only have 6 of the SRAMs and you can only do multiples of 2, which obviously won't work.

[oompa loompa]

This is a picture of my SRAM mod . I only bumped up the SRAM size from the stock 64kb to 128kb. I think there were some some issues with putting bigger sized memory in the SRAM memory space, or else I would have maxed it out to 2mb , bu I really don't remember the issue i had

May 22, 2009 3:26:41 GMT -5 GiGaBiTe said:

I'd love to upgrade it further, but I only have 6 of the SRAMs and you can only do multiples of 2, which obviously won't work.

you can stack more chips to get more memory out of the thing . you just have to decode the address to turn on a specific chip

[GiGaBiTe]

I know how to solder stuff, but for electronics logic, I'm more than clueless. If you can make a schematic with the parts list, I can do it, otherwise it won't happen.

[jlf65]

I'd guess the main issue would be needing to find the higher address lines so you can decode 2 MB of space. Clearly, the lines to the existing chips won't have all the address lines. Probably the easiest place to find them is the cart port. The other issue is finding SRAM chips that size that are also 5V. The SRAM chips I've seen that are that size are all 3.3V these days.

[GiGaBiTe]

You can always step the 5v down to 3.3v, though I don't know how cleanly you could do it with resistors.

[jlf65]

That's not the problem - the problem is will the 3.3V chips have enough signal level to work with the rest of the system. If you need level shifters on the data lines, that would make the hack considerably more difficult.

[oompa loompa]

3.3v signals aren't suppose to work, but sometimes they do . Every time I have interfaced 3.3v signals to the Genesis, I have never had a problem. All the IC's on the Genesis are CMOS, and CMOS tells us that for a 5v part, bringing a signal high to 3.3v is in the "no-no" zone. If the Genesis were running at a higher clock rate, then it probably wouldn't work , but I guess at 8mhz, the Genesis takes its time to latch in the data bus.

I have to bring my Genesis over to the logic analyzer (probably would take a month for me to not be lazy haha), and run some sample test code to give you a valid answer on how to max out the RAM. I also have to see if I have the same model Genesis . I don't know how much the SRAM is mirrored out, so there will probably be some more glue logic, if I'm not able to use the SRAM enable signal already provided. I dunno, but there must have been a good reason for me to only use 128k of RAM instead of maxing the thing out because I've always had lots of RAM on my hands .

[jlf65]

Well, the MD-Pro flash cart uses Intel StrataFlash, which is 3.3V, so it at least works for roms.

[Tiido]

ToToTek stuff uses 3V chips that are 5V I/O tolerable...

A friend of mine said that you can make a 3V only chip work in 5V system using some 330ohm resistors on the signal lines... I am a bit sceptical though.

and MD1 is NMOS mostly AFAIK, MD2 is full CMOS design

[Tom Maneiro]

This reminds me to...

- build my own FlashROM Genesis cart, using old FlashROM BIOS chips, a salvaged cart board, and an old motherboard BIOS socket as the ROM programmer (Uniflash anyone?). No fancy wires, no messy programmers... just hot-swap

- fix the broken headphone output on my MK-1601 board (i b0rked the connector wire to the motherboard on the jack board side, so i will need to replace the entire jack, since, as everybody on this board knows, my soldering skills are WORSE THAN FAILURE )

- Make a better AV cable, to replace that frankenstein "hacked 32X AV patch cable" adaptor that i made in 2005...

- Cleanup the bastard and get rid of that super-rusty RF shielding...

- Nuke the RF box, since i don't want to know about that shit NEVER EVER.

[oompa loompa]

Current limiting resistors are used so that you don't fry a 3.3v part. Resistors are a very sloppy way to shift the voltage . An active level shifter is much better.

You're right, I think the MD1 has NMOS parts. The 68000 was brought up as an NMOS technology. I forgot what the part number is for the CMOS version, is it something like 68LC000? If the parts in the megadrive are NMOS, then there shouldn't be a problem with the different technology families . NMOS has something like a VIH of 1.8v or so. CMOS is around VIH = VCC*0.8 or even higher

[GiGaBiTe]

CMOS versions of the 68000 are MC68HC000 or HD68HC000 for the Hitachi part. Regular 68000s are just MC68000P8

The LC thing was used in the later 68040 to indicate the lack of a FPU.

[jlf65]

May 28, 2009 4:27:10 GMT -5 GiGaBiTe said:

CMOS versions of the 68000 are MC68HC000 or HD68HC000 for the Hitachi part. Regular 68000s are just MC68000P8

The LC thing was used in the later 68040 to indicate the lack of a FPU.

Motorola later made LC versions of most of their chips. I don't remember an LC 68000, though. I thought they only made LCs back to the 020. The A1200 uses an LC 68020.