Super Mario RPG on a PAL Snes

So I've tried to follow the tutorial from http://www.mmmonkey.co.uk to allow my NTSC copy of Mario RPG to work on PAL Snes. Lifted the pin on the SA-1 chip and the game boots but after the Square logo shows background (clouds) and the music playing, no sprites. What was wierd every time I pressed L+R and start and select the game would reset or go back to the Nintendo and Square logo.
Read some posts on nesdev.com forum and someone has had similar issues that he fixed, in his case pin 126 needed a bit of solder or something. So I fiddled with mine until I broke the pin on the SA-1... Frustration has set in but I wasn't ready to give up just yet. After a smoke I decided to remove the NTSC mask rom and the SRAM and solder it into PGA European Tour PAL since it's only got 64K and Mario RPG requires 256K.

Did all that, started the game but graphical corruption were present, although this time the game didn't freeze, Peach and the butterflies were there... Rechecked the soldered points by looking at them only and noticed one of the pins on the SRAM didn't have enough solder so I put a bit more and it was much better but the corruption was still present in some scenes..

Ended up desoldering the mask rom because to me it looked like it was on a bit of an angle, soldered it back, same thing.

I believe the SRAM seems to be the problem, might need to check the connections there. People that have made their own repros said something about clearing the SRAM, but I take it that's because they used an SRAM from another game not Mario RPG?

I'm ready to give up on this after hours of working on it and ruining the cart :/

Any suggestions, thanks!

 

Bin it and buy a new one. Then mod your SNES with a supercic.

 

I'm trying to get this working. So any suggestions like get an NTSC Snes, mod it with a superCIC and 50/60hz switch are not really helpful in this case

 

Sir (fellow Aussie!) You can supercic your snes. The only downer is you will have to modify the case to fit us carts (if you wish). Probably sound like a jack-ass stating the obvious.

It would make your life easier for sure as it ensures compatibility for all.

I have recently done a 1 chip us console for myself and have 3 pal super-cic'ed consoles i am selling on ebay slowly.

If you are stuffing around in a cart, you are more than capable to do the supercic matey.

But i understand wanting to get something to work come hell or high water.

 

This. I know it's not the answer you wanted, it's by far the best option.

 

I will investigate the SuperCIC but for that I need another Snes, thanks for the recommendation mateys. Anyway I have Action Replay MK3 to play NTSC games, except for Mario RPG obviously.

ere is a photo of the two boards, Right original NTSC, left is PAL PGA European Tour with the chips soldered. I only noticed the differences last night, the extra FBX on PGA next to the R5 and R6. Forgive me but I do not know what they are.
I was beginning to think that it might be a problem but I have saved a video of the game playing on the original NTSC board with just SA-1 pin 127 lifted and the same graphical glitches were present just before it would freeze.

The plan is to test continuity from the SRAM to the SA-1, as I managed to decrease the graphic corruptions by adding more solder on one pin. I have also removed the SRAM battery over night to clear the SRAM? and tried it without a battery, made no difference

Links to what it looks like on TV:
http://img35.imageshack.us/img35/4731/2v6l.png Looks normal
http://img59.imageshack.us/img59/9436/nz42.png The problem starts here in the intro, before that is perfect, so is the start menu
http://imageshack.us/a/img23/1989/9xtp.png Gameplay :/

 

Looks like a RAM addressing problem, all right.

This is probably caused by the upper address lines not being connected to the SRAM chip. If you look at the type numbers on the board, you will notice they are different - the original Mario RPG board is a SHVC-1L5B and the new board is a SHVC-1L3B - so the boards are actually designed to take different SRAM sizes. The only pins that differ between the 64 and 256kbit SRAMS are pins 1 and 26, so start out by finding out where they go on the 1L5B board (presumably the SA-1) and then check them on the 1L3B board (the corresponding pins will likely either be connected to VCC or just floating).

If they are tied to a rail, cut the corresponding trace and then wire the pin on the SRAM to the corresponding place on the SA-1.

And don't worry about the missing ferrite beads - they are interference suppression components that were likely included on the original design just in case they were needed and then removed in a later PCB revision when it became clear they weren't.

 

Thank you so much for your reply. I am at work but couldn't wait to get started on this, all I had time for was to find where the two pins go on the 1L5B

PIN 1 -> SA-1 87
PIN 26 -> SA-1 103

The rest will have to wait

Also thank you for explaining what those FB's stand for as well!

 

I have found that on the PGA European Tour cart 1L3B
SRAM PIN 1 -> SA-1 87 (identical)
SRAM PIN 26 -> U4 PIN3

TriMesh if you would be so kind to look at the attached image and confirm where I need to cut the trace? I placed a yellow X and solder 26 to U4 Pin3?

I really do appreciate your help

1L3B:
http://img31.imageshack.us/img31/879/w8ee.jpg

 

OK, looking at the PCB, I have a strong feeling that those pads next to the RAM supervisor chip are what you need to play with.

http://bayimg.com/LaclnAAFj

If you cut the trace between the two pads I've marked as "Link #1" in that photo, then you should find that the vias with the arrows pointing to them are connected to the specified places. In that case, just link the two pads marked as "link #2", and the RAM addressing should be correct.

Edit: In fact, looking at the two photos you made in the earlier post, I can see that "Link #1" is cut on the original PCB - I can't actually verify that "Link #2" is made, but it looks like it might be...

 

I have checked and you are completely right. Link#1 is cut and Link#2 is definitely connected. I will try this when I get home, very excited!!
You are amazing

 

I tried to find my copy of Super Mario RPG, but it seems to be in storage somewhere and I'm not sure where.

But I did find these photos on SNES central - and they seem consistent with this theory:

http://www.snescentral.com/pcbboards.php?chip=SHVC-1L3B-11
http://www.snescentral.com/pcbboards.php?chip=SHVC-1L5B-11

Note that both of these boards are NTSC (hence the SHVC prefix) and the same (-11) revision. They also both appear to have been made in June, 1996, so any inconsistencies should be down to intentional changes.

As far as I can see, the actual PCBs are identical except for the number on the silkscreen and those two jumpers - it's possible there are other changes on the back, but since you have already verified that pin 1 on the SRAM is wired to the SA-1 even on the board with the 64kbit SRAM, it's hard to think of anything.

This seems to have been an intentional change, too - there is also a photo of an older (-02) revision of the 64bit board on SNES central:

http://www.snescentral.com/pcbboards.php?chip=SHVC-1L3B-02

And this clearly has pin 1 of the RAM wired to VCC.

 

I have managed to cut the trace and join the two right ones together with a wire, it's working yaaaay. Been playing it for 2 hours so far, no graphical or sound issues.

One thing I have noticed, if I don't push the start button and let the introduction play, the music finishes about 10-15 seconds sooner (haven't really timed it). So towards the end of the intro when Princess Toadstool appears there's no music playing only silence. Not sure what the deal is there but to be honest if that's all I can live with it. Hopefully I'll finish the whole game just to make sure there are no other problems.

Thank you so much for all your help TriMesh, you are truly a gentleman and have made my day with this

 

My guess is that the music is still playing at the same speed it always did (since the SPC-700 has the same clock frequency on both NTSC and PAL systems) - but that the graphics in the intro are running 16.6% slower because it was originally timed on a NTSC system with a 60Hz frame rate and you are now running it on a PAL system that uses 50Hz.

I'm glad you got it working - and we also added a little bit more knowledge about the difference between those 2 PCBs

 

That does make sense and is not really an issue. For some reason I always thought there had to be a hacked PAL version of this game that ran at 50hz which people used to make repros run on PAL consoles
Thank you

 

Good luck in the Yoshi race, that'll be a doozy with off sync music.

 

There are apparently a couple of versions of this cart, one works, one not. I did the mod some time ago on my NTSC version of this game, and it works great on an unmodded PAL console now. At some point I thought of rehousing it into a PAL game cart and selling it as what it is; a modded NTSC version of this game working on a PAL console would get a good price after all even as listed as not original housing.

 

I'm up to my fifth star, no problems so far. I couldn't have done this without TriMesh

 

Hi Pikmin,
When you lifted the pin off the SA-1 chip did you connect it to 5v? It has to be connected to 5v to set the SA-1 chip into PAL mode.
I would be checking that on your US cart and see how you go.

UPDATE:
Just read the rest of the thread and you sorted it all out.