SNES Nonvolatile SRAM board

Includes but not limited to: SNES, Genesis, Sega CD, PlayStation 1, Nintendo 64, Dreamcast, Game Gear and I guess the Virtual Boy.

Moderator:Moderators

qwertymodo
Posts:122
Joined:Sat Apr 07, 2012 10:54 am
360 GamerTag:qwertymodo
Steam ID:qwertymodo
SNES Nonvolatile SRAM board

Post by qwertymodo » Sat Jun 09, 2012 3:47 pm

After seeing the mods for nonvolatile Game Boy and N64 Controller Pak saves, I figured why not SNES? The SNES has plenty of good RPG's that could benefit from it. So I went with Ramtron Fe-RAM, same as before, and traced out this board. Thankfully, Nintendo was nice with their SRAM, unlike their mask ROMs, and used standard pinouts, so this board works for both the 64k and 256k chips. I'd have to play around a bit to get 16k working, plus Ramtron doesn't sell 16k chips, so you'd have to use a 64k chip, which would mean I'd want to add bank switching for 4x saving... which I currently don't care enough to do, since all the games I wish to mod are all 64k. Still waiting to get the boards back from the fab house, so I'll post back with results when I get them, but I don't see why they wouldn't work. If it works, I'll post gerber files, or perhaps just sell the extra boards (I have to order in multiples of 10). Here's what the board looks like:
Spoiler:
Image

User avatar
evilteddy
Portablizer
Posts:423
Joined:Tue Mar 25, 2008 2:11 am
360 GamerTag:Kirren of Smeg
Steam ID:kizzinator
Location:Newcastle, Australia

Re: SNES Nonvolatile SRAM board

Post by evilteddy » Sat Jun 09, 2012 10:27 pm

Awesome idea, and I love that you're getting a PCB made to keep it all clean and pretty looking. I'm sure they'll be people willing to buy boards off you, especially if you do the SMD work for them. :lol:

A couple of small tips for future PCBs: it's generally good practice to have a ground plane filling the unused parts of the board on both layers to help reduce noise and cross talk (and also helps with power supply which is clearly not an issue here). If you can avoid putting vias in pads then it'll make your life very slightly easier (but it tends to be more of a problem when you're not hand soldering).

Anyway, looks like a cool project. Keep us posted.

qwertymodo
Posts:122
Joined:Sat Apr 07, 2012 10:54 am
360 GamerTag:qwertymodo
Steam ID:qwertymodo

Re: SNES Nonvolatile SRAM board

Post by qwertymodo » Mon Jun 11, 2012 12:31 am

Thanks for the tips. I actually knew I should have added a ground plane, just didn't think to do it. I've updated the board in case I have any future use for it (not really sure how many of these I'll need). I moved the vias off the pads as well.
Spoiler:
Image

User avatar
evilteddy
Portablizer
Posts:423
Joined:Tue Mar 25, 2008 2:11 am
360 GamerTag:Kirren of Smeg
Steam ID:kizzinator
Location:Newcastle, Australia

Re: SNES Nonvolatile SRAM board

Post by evilteddy » Mon Jun 11, 2012 1:54 am

Very pretty

Haunted360
Posts:1000
Joined:Sat Jan 30, 2010 12:22 am
PSN Username:Haunted360
360 GamerTag:Haunted 360
Location:Australia
Contact:

Re: SNES Nonvolatile SRAM board

Post by Haunted360 » Mon Jun 11, 2012 3:09 am

Very, very awesome. Excellent work. I am impressed!

qwertymodo
Posts:122
Joined:Sat Apr 07, 2012 10:54 am
360 GamerTag:qwertymodo
Steam ID:qwertymodo

Re: SNES Nonvolatile SRAM board

Post by qwertymodo » Thu Jun 14, 2012 1:48 pm

Got the boards in today. Haven't had a chance to solder one up yet, but I'll post back with results. The only thing I can see right off the bat might be an issue is the height of the board + chip being too high to close the cart properly, but if that ends up being an issue, I'll just have to find pins with thinner spacers than these to give me a lower profile: https://www.jameco.com/webapp/wcs/store ... 1_68339_-1 The other, more extreme solution would be to redesign the board with the SMD chip on the BOTTOM of the board, but before I consider going that route, I need to find out whether or not it's even necessary. We'll see how it goes.

qwertymodo
Posts:122
Joined:Sat Apr 07, 2012 10:54 am
360 GamerTag:qwertymodo
Steam ID:qwertymodo

Re: SNES Nonvolatile SRAM board

Post by qwertymodo » Fri Jun 15, 2012 9:55 am

Alright, got the boards, got one soldered up. It's puuurrrtty :) Still haven't been able to test it. I tried and it didn't work, but my socketed donor carts are giving me nothing but grief. I'll have to pull the sockets and solder it in direct and see if it works. It *does* stick up higher than the rest of the chips, and so doesn't quite let the board sit flat in the cart shell. I'll have to see whether or not that affects the ability to close the carts properly, but right now, like I said, I have sockets in my donor cart, which already makes it impossible to put them back in a shell, so I'll put it in one of the other old sports games I bought to play around with and I'll let you know how it fits. Then, if they work (and I don't see why they shouldn't) I'll see about offering the extras in the Trading Post.
Spoiler:
Image

qwertymodo
Posts:122
Joined:Sat Apr 07, 2012 10:54 am
360 GamerTag:qwertymodo
Steam ID:qwertymodo

Re: SNES Nonvolatile SRAM board

Post by qwertymodo » Sat Jun 16, 2012 10:58 am

Ok, so it doesn't work :( I tested all the traces for continuity and checked my soldering for bridges, so now I'm at the point of thinking perhaps the chip just isn't compatible. I'm poring over the relevant datasheets, but so far coming up empty. If anybody with more insight into such things would be interested in taking a look, here's the docs:

Sharp LH5268AF-10YLL (original SNES SRAM): http://pdf1.alldatasheet.com/datasheet- ... 5268A.html
Ramtron 16W08 (nonvolatile replacement): http://ramtron.com/files/datasheets/FM16W08_ds.pdf

qwertymodo
Posts:122
Joined:Sat Apr 07, 2012 10:54 am
360 GamerTag:qwertymodo
Steam ID:qwertymodo

Re: SNES Nonvolatile SRAM board

Post by qwertymodo » Sat Jun 16, 2012 11:09 am

Ok. I've found one thing that might be an issue. Does the SNES pulse /CE on the cart or does it just pull it to GND? The FeRAM latches addresses on the falling edge of /CE, so if the SNES is just pulling it to GND, then that would explain it.

qwertymodo
Posts:122
Joined:Sat Apr 07, 2012 10:54 am
360 GamerTag:qwertymodo
Steam ID:qwertymodo

Re: SNES Nonvolatile SRAM board

Post by qwertymodo » Sat Jun 16, 2012 2:02 pm

Bummer, the FM28V020 (with a 5V->3.3V LDO regulator and A13/A14 pulled to GND) doesn't work either. It has its own strangeness. The FM16W08 has no weird symptoms, it just doesn't actually save. You save, it says it succeeded, you go to load and there's nothing. The FM28V020, on the other hand, causes weird gray wavy lines across the screen (I can still hear the audio playing), and the MAD-1 gets REALLY hot (it almost burned me when I touched it)... so, something weird is going on there. Again, if anybody else is willing to take a look, here's the datasheet for the FM28V020: http://ramtron.com/files/datasheets/FM28V020_ds.pdf

Edit: *DOH... I can't believe I totally just did that... I wired the voltage regulator BACKWARDS >.< I always get confused with the Vdd/Vss notation, rather than Vcc/GND :(

Edit 2: Nope, still no luck :(

User avatar
evilteddy
Portablizer
Posts:423
Joined:Tue Mar 25, 2008 2:11 am
360 GamerTag:Kirren of Smeg
Steam ID:kizzinator
Location:Newcastle, Australia

Re: SNES Nonvolatile SRAM board

Post by evilteddy » Sat Jun 16, 2012 7:58 pm

Its possible that you could have killed your chips when you had them in reverse. Other than that I can't see any problems other than there possibly being a problem on the Chip enable lines. If you have a logic analyser or a scope then connect those up and you'll be able to see what's up.

qwertymodo
Posts:122
Joined:Sat Apr 07, 2012 10:54 am
360 GamerTag:qwertymodo
Steam ID:qwertymodo

Re: SNES Nonvolatile SRAM board

Post by qwertymodo » Sat Jun 16, 2012 8:34 pm

Unfortunately, my only access to a logic analyzer or oscilloscope was in the university EE labs, and school just got out for the year. I considered the fact that I might have fried the chip, so I am willing to try once more (I have 3 samples of the FM28V020's), but at least I know the rest of the SNES cart is good, I threw the original SRAM back in it and it worked just fine.

qwertymodo
Posts:122
Joined:Sat Apr 07, 2012 10:54 am
360 GamerTag:qwertymodo
Steam ID:qwertymodo

Re: SNES Nonvolatile SRAM board

Post by qwertymodo » Mon Jun 18, 2012 5:57 pm

No luck, I've tried twice more with my two remaining samples, once with A13/A14 lifted and grounded and once with them grounded without lifting them. The one where I didn't lift A13/A14 won't even boot when attached to the cart (but the cart boots just fine without any SRAM at all), and the one with A13/A14 lifted boots fine but still won't save :(

qwertymodo
Posts:122
Joined:Sat Apr 07, 2012 10:54 am
360 GamerTag:qwertymodo
Steam ID:qwertymodo

Re: SNES Nonvolatile SRAM board

Post by qwertymodo » Thu Jun 28, 2012 6:33 pm

Ok, I'm just ordered some level shifters, so I'll try it one more time with the FM28V020 properly driven (LDO regulator + level shifters on all logic lines). If this doesn't work, then I'm calling it a bust unless anyone else has any ideas. Especially after the amount of pain I went through to hand route all the extra traces to accommodate the regulator and level shifters:
Spoiler:
Image

qwertymodo
Posts:122
Joined:Sat Apr 07, 2012 10:54 am
360 GamerTag:qwertymodo
Steam ID:qwertymodo

Re: SNES Nonvolatile SRAM board

Post by qwertymodo » Fri Jun 29, 2012 10:17 am

Can anybody confirm this schematic? The part marked FM18L08 is actually going to be an FM28V020, but the footprint and pinout are the same, so for the sake of the schematic, it's fine (I'm just too lazy to add the new part). The _A/A_ and _D/D_ signals indicate the two sides of the level shifter. _A/_D is the "A" side of the shifter, which is the +5V rail (VCC), and A_/D_ are the "B" side, which is 3.3V (VDD). The connections are rearranged to match up better to the FeRAM pinout (so there are fewer trace crossings), but that shouldn't matter so long as the A and B sides match up. The part numbers don't show up right for the shifters, but the 16-bit shifter (ABUS) is a TI SN74LVCH16T245, and the 8-bit shifter (DBUS) is a TI SN74LVC8T245, and the regulator is a National Semiconductor LM3940.
Spoiler:
Image

Post Reply