I'm quite unsure exactly how to do this, i've soldered in my ROM and performed a standard 2->1 chip conversion. I have the /OE pin removed and wired to PIN 4 of the MAD-R, and as the original ROMs are 1Mb i have connected A20 to PIN 13 of the MAD-R. As i'm using the 24Mbit text width hacked version of SoM in this cart i have wired A21 to SNES #46 on the card edge connector. I am however still getting a black screen and the game won't boot. Anyone have any advice?
(http://i.imgur.com/TFRQ0eb.jpg)
(http://i.imgur.com/oRbFquN.jpg)
Did you fill the end of the flash chip with all FFs? A 24Mbit file means there would be 8Mbit free in a 32Mbit flash chip.
Also, try rewiring just flash adapter pin 2 to MAD pin 13 and adapter pin 33 to MAD pin 4.
Why didn't you solder in a DIP socket? If you need to flash that chip again you're gonna have a bad time.
I want this to fit in the cart (so no socket), and i'm pretty adept at surface mount repair/chip removal.
The program I used to flash the chip automatically fills out the chip with FFs. I don't see how wiring Pin 2 (A21) to MAD pin 13 would work. MAD pin 13 is routed to SNES #45 (A20) as is expected when a dual 1MB rom setup is used, so this is where I connected my adapter Pin 1 (A20).
Note: in a 2x 2MB rom setup MAD pin 13 is wired to A21 when swaping for a flash ROM. Examples of this include Mother 2, Final Fantasy VI and any other dual 2MB chip boards.
If the board wasn't originally meant to support 32Mbit on the P0 socket, there may not be an easy way of altering it to now. Try wiring Pin 2 to MAD 13 and see if that helps.
Method 1: Adapter Pin 2 (A21) -> MAD 13 does not work (here i have tried Pin 1 both loose and tied to GND)
Method 2: Adapter Pin 1 (A20) -> MAD 13 w/ Adapter Pin 2 -> MAD 12 (routed to SNES #47) does not work
Method 3: Adapter Pin 1 (A20) -> MAD 13 w/ Adapter Pin 2 -> GND does not boot (dark static fuzz)
Method 4: Adapter Pin 1 (A20) -> MAD 13 w/ Adapter Pin 2 -> SNES #46 does not boot (dark static fuzz)
Method 3 SHOULD allow the game to boot (tested in emulator w/ top 8 mbit removed to simulate A21->GND) but freeze on the first cutscene, however i just get very faint static fuzz on the screen.
Method 4 SHOULD work completely by my logic.
I understand that the cart was not originally configured for 32Mbit games, but wiring A20,A21,Vcc and /OE appropriately SHOULD allow the cart to work.
Critical connections on board (checked with multimeter):
- SNES #45 (A20) -> MAD-R pin 13
- SNES #46 (A21) -> NOTHING (Unrouted, attached by solder spot)
I think that maybe the MAD-R chip is the problem, and is not triggering /OE on pin 4 as would be expected from a MAD-1
I tried this very same thing with a single 32 pin ROM board, I couldn't figure out how to get a larger ROM to work.
I was able to do this with Final Fantasy IV (which has a 74LS139 decoder) but that is a 12Mbit translated ROM and doesn't use the A21 address line. I need to get my oscilloscope out and check the /OE line with the unit powered on, I still think that maybe the MAD-R pin 4 isn't triggering /OE correctly
Post Merge: December 12, 2015, 01:33:26 am
nesdev forums thread (http://forums.nesdev.com/viewtopic.php?p=102869#p102869)
Turns out i'm right on this, MAD-R pin 4 is RESET and NOT /OE
Have you tried simply removing the P1 mask rom? I always do and haven't had any issues with my repros.
Edit: I also run a wire from the adapter /OE (pin removed) to MAD pin 4.
Yes, removing the 2nd MaskROM should make it work on 2XXM boards since it tries to read the 2nd MaskROM and then crashes. You will also need to rewire the MAD-R accordingly to fit a 1XXM board.
Quote from: muckyfingers on December 12, 2015, 04:16:30 am
Have you tried simply removing the P1 mask rom? I always do and haven't had any issues with my repros.
Edit: I also run a wire from the adapter /OE (pin removed) to MAD pin 4.
As previously discussed MAD-R pin 4 IS NOT /OE.
From experience with all my other repros I shouldn't need to remove P1. I will try using an XNOR gate to toggle /OE.
XNOR LOGIC
/OE_1 /OE_2 /OE Output
0 0 1 ; both chips enabled, null case from MAD-1/MAD-R
0 1 0 ; chip 1 enabled, output true
1 0 0 ; chip 2 enabled, output true
1 1 1 ; both chips disabled (/RAM_OE or /RESET) , output false
to completely prevent output conflicts I may lift MAD-R pin 1 and 16 and wire Chip 2 /OE to Vcc
Post Merge: December 12, 2015, 07:14:54 pm
Tried using a HD14077B XNOR,
-Lifted PIN 1 and 16 from MAD-R
-Bridged P1 /OE to Vcc to disable the chip
-wired PIN 1 and 16 from MAD-R to 1A and 2A (pins 1 and 2) on the XNOR
-wired adapter /OE to Q
A (pin 3) on the XNOR
-wired Vcc and GND from P1 adapter to XNOR
Result: Black screen w/o the random fuzz i got before
I had MAD-R before and it worked fine. I have no clue why you overcomplicate these things.
This is what I would do:
Remove P1 and P2 completely, no matter if you only use P1.
Install TSOP Adapter + Flash into P1 with all pins connected.
Does it work? Stop here.
If it doesn't I'd try this:
Check connections to A20,A21 if they're present or grounded.
If so, re-route these connections.
Check where /OE is going. Is it going to MAD-R Pin 16 for P1?
I would still try to do the following if all of the above does not work:
Desolder MAD-R completely and connect it like this:
__ __
NC |01\/16| NC
RAM /CE |02 15| SNES #40
NC |03 14| A19
ROM /OE |04 13| A20
+5V |05 12| A21
+5V |06 11| SNES #49
+5V |07 10| GND
GND |08 09| SNES #26 and RAM #26 (only in 6264)
------
If all else fails check your soldering from the EEPROM.
Tried this out, same result with a black screen. I checked the ROM when I flashed it yesterday by reading back the data and loading the .bin into zSNES (worked perfectly but gave bad checksum as the original file was 24mbit and my programmer filled the remaining 8mbit with FF). As a last resort I can desolder the TSOP chip and make sure that it hasn't lost data in all this tweaking.
As you can see in the attached photos there are no solder bridges on the TSOP chip, and I have fluxed/reflowed these pins a couple of times now to ensure that any possible solder bridges are removed.
(http://i.imgur.com/5owYaoJ.jpg)
(http://i.imgur.com/8x2HNu3.jpg)
I have verified that all Jumper points on the adapter board are sufficiently bridged using my multimeter. I have also attached a photo below of the soldering between the cart and the adapter board, note /OE pin is not present.
(http://i.imgur.com/D2LS1FD.jpg)
I have removed the MAD-R, lifted pins 1,5,7,10,12,14,15 and soldered it back in.
MAD-R Pin connections
PIN 1: N/C lifted
PIN 2: RAM /OE (already present connection on board)
PIN 3: N/C soldered in (not connected to anything on the board)
PIN 4: ROM /OE (adapter board) soldered in
PIN 5: Vcc lifted (connected to PIN 6)
PIN 6: Vcc (already connected on board)
PIN 7: Vcc lifted (connected to PIN 6)
PIN 8: GND (already connected on board)
PIN 9: SNES #26/RAM #26 (already connected on board w/ 6265 RAM)
PIN 10: GND lifted (wired to PIN 8 )
PIN 11: SNES #49 (already connected on board)
PIN 12: A21 lifted (connected to adapter pin 2 which is connected to SNES #46)
PIN 13: A20 connected through board to SNES #45 and wired to A20 on adapter
PIN 14: A19 lifted (connected to A19 on adapter board)
PIN 15: SNES #40 lifted (connected to A15 on adapter which is connected to SNES #40 through the board)
PIN 16: N/C lifted
I still have a feeling that the thread above was right and that MAD-R pin 4 is RESET as verified by my oscilloscope earlier. (observed a single +5V pulse on system startup)
Looks like the bottom wasn't soldered correctly on the photos. Looks like there's holes but that could be wrong.
Either way, I have no clue otherwise. It works fine for me with any game yet, no matter if MAD1,2,R.
Probably just the lighting, perfect little cones of solder around all the pins. I've reflowed every single one to make sure none have a cold solder joint. Will desolder the EEPROM and check to make sure the file hasn't somehow been corrupted. Thanks again for your help man.
Tried reflashing and verifying the ROM. It's perfectly fine, and the .bin I ripped from it loads fine in ZSNES.
I switched over to a new adapter board and rewired the MAD-R in the same configuration as the MAD-1 in Chrono Trigger. I now get either scrolling fuzz or I get flashes of garbage on screen if I'm lucky.
The .sfc file i'm using (Secret of Mana w/ enhanced text) is 24Mbit and Hi-ROM so it makes sense to use the same configuration as Chrono Trigger, but i'm still seeing issues with the cart. I've ruled out all connection issues with my multimeter, carefully inspected the flash chip and all the solder joints to the adapter board, and have used every logical MAD configuration with the original MAD-R chip to no avail.
I'm sick of the MAD-R shit, so does anyone have a spare MAD-1 lying around I could buy?
I was right, the MAD-R was the problem. Got a MAD-1 from HVC-MAN today (Thanks again dude!) and soldered it in. I wired it exactly the same as the MAD-1 in Japanese Chrono Trigger and it worked perfectly!
So the game works but the RAM does not appear to be saving. When I save the game the save vanishes and the files appear empty. This is WITHOUT turning off the console or resetting, just going back to a save point and saving again. I have tried the pinout suggested by xIceMan
Quote from: xIceMan on December 13, 2015, 01:35:58 am
I had MAD-R before and it worked fine. I have no clue why you overcomplicate these things.
__ __
NC |01\/16| NC
RAM /CE |02 15| SNES #40
NC |03 14| A19
ROM /OE |04 13| A20
+5V |05 12| A21
+5V |06 11| SNES #49
+5V |07 10| GND
GND |08 09| SNES #26 and RAM #26 (only in 6264)
------
except with PIN 10 to +5V as the ROM is HiROM and not LoROM.
Again, the game booted and played fine but would not save. Any ideas?