|
|
ForumsSega Master System / Mark III / Game GearSG-1000 / SC-3000 / SF-7000 / OMV |
Home - Forums - Games - Scans - Maps - Cheats - Credits Music - Videos - Development - Hacks - Translations - Homebrew |
View topic - SMS FCR's
 |
| Author | Message |
|---|---|
|
Angus
|
 SMS FCR's
 Posted: Sun May 21, 2000 12:18 am
|
|
I read the SMSARCH document today and it was very useful because i want to build a SMS ROM 'undumper' using a 512kB EEPROM, as a college project. Reading the document, I found out that i must build a circuit to simulate the FCR's, so i have a couple of doubts....
first it says: "According to the previous section, only the first 1K of ROM page 0 is available on power up. The reason is because the values in the FCR's are not known. However, it appears that cartridges make the provision that Frame 0 includes page 0, and Frame 1 includes page 1 on start-up. This greatly simplifies accessing the first 32 kilobytes; no writes to the FCR's are necessary. Therefore, in practice, the first 2 pages are available immediately after power-up." Does it mean that when the system powers up, FCR0 must be set to $00 and FC1 must be set to $01 in order to my fake cartridge work? Next, it says: "The SMS uses three 16-kilobyte frames, allowing a total of 48 kilobytes of ROM to be "visible" at a time." and then.... "Accessing the remaining pages requires using the FCR's. The general paging scheme mentioned in section 2.1 indicates that any page can be mapped into any frame. It appears, however, that depending on the size of the cartridge ROM, only certain frames can be used. Cartridge Type Frame used for Paging (all pages except 0 and 1) One-Mega (128 kilobyte) Frame 2 Two-Mega (256 kilobyte) Frame 1 Four-Mega (512 kilobyte) Frame 1 Then, all of the remaining pages in a "One-Mega" cartridge should be mapped into Frame 2 using FCR 0xFFFF. Likewise, for two and four mega cartridges, all remaining pages are mapped into Frame 1 using FCR 0xFFFE." Since only frame 1 or frame 2 is used, do I really need 3 FCR's? And do I really need a FCR for frame 0 since it seems to be used only during power up? How many registers do i really need to build my fake cartridge? i hope anyone can help... |
|
|
Posted: Mon May 22, 2000 4:06 pm
|
Hello. I designed and built the SMSARCH and wrote the SMSARCH document. Hopefully I can help you.
That would be my recommendation. I noticed that I could reliably read page 0 and page 1 of the cartridge ROM right after power-on without having to write to FCR0 or FCR1. It is likely that some game software assumes that these registers are initialized to 0 and 1 respectively, so your fake cartridge should probably ensure these registers are initialized to 0 and 1 also.
To make your cartridge as general as possible, I would recommend having 3 FCRs. However, it appears that commercial software does not use FCR0, so you could possibly leave it out, and just include FCR1 and FCR2. If you leave out FCR0, make sure that ROM page 0 is always mapped to addresses 0x0000 through 0x3FFF. Eric Quinn |
|
  
|
|
|
Angus
|
Posted: Mon May 22, 2000 4:40 pm
|
|
Eric, first of all, thanks for your help, and congratulations for your document. It is really very good and was very useful for me. Did you ever finished it? If so, please mail me a copy:
joaolac@zaz.com.br Well, i still have some questions for you... Since FCR0 is not used, and FCR1 and FCR2 are not used together by any cartridge, I am thinking about using only one general FCR, initialized with $00, with its outputs always enabled and that could always be written when WR=1. Do you think that would work? and when you say: "If you leave out FCR0, make sure that ROM page 0 is always mapped to addresses 0x0000 through 0x3FFF." I don't get it... if i use only one register, initialized with page 00, don't you think that would work? isn't frame 0 used only when the game starts? or is it used again even after frame 2 or frame 1 is used? i thinking about sending a schematic of my project to you... maybe you could tell me if there is anything that could not work... |
|
|
Posted: Mon May 22, 2000 7:53 pm
|
Out of experience I can say 95% of the cartridge don't even use the RESET features so it's more unlikely that default pages could be set for sure. Beside I never saw a SMS/GG game not setting the first and second pages at the beginning, but they could exists if the cartridge use the RESET signal to map itself this way on power. |
|
|
|
|
|
Posted: Mon May 22, 2000 9:08 pm
|
I do not think this will work. Each of the three page frames is allowed to hold a unique page of ROM. Some software may count on different pages being in frame 1 and frame 2, even though only one of the frames is used for page swapping. If you only use one FCR, there's no good way to determine whether it should be addressed using 0xFFFE (for frame 1) or 0xFFFF (for frame 2) since different cartridges use one or the other of these addresses. Furthermore, you can't simply map the FCR to both addresses since frame 1 and frame 2 can contain different ROM pages. I think you really need to FCRs, one for frame1 and one for frame 2.
Frame 0 contains ROM page 0 which is where the INTR and NMI interrupt handlers are located. These handlers are used at least 50 times a second. You will always need to make sure that page 0 is mapped into frame 0. If you don't use a FCR, you will need to hard-wire those addresses to map to page 0 in ROM.
That would be O.K. My e-mail address is: e-quinn@elnet.com (or just click on my name above this message.) Eric Quinn |
|
|
|
|
|
Angus
|
so i don't need to initialize FCR1?
Posted: Tue May 23, 2000 1:43 am
|
|
I'm not sure if i understood your reply because my english sucks =)
since i want to build a project to use only with the most popular titles, i don't think my project have to include those 5% that use those RESET features you mentioned. so do you believe i really don't need to initialize FCR1 with $01? am i right? i wouldn't need to initialize FCR0 anyway cause 74XXX registers are initiated with $00... but i wouldn't want to build a circuit to inialize FCR1 with $01... =) i do hope that's not necessary!
|
|
|
Angus
|
i'll use 3 FCR's, end of question =)
Posted: Tue May 23, 2000 1:44 am
|
|
well, i have decided to build it with 3 FCR's to make sure it will work. as soon as i have scanned the schematics i'll send them to you for your "approval"... =)
eric, do you use icq? mine is 1389871 and my e-mail is joaolac@zaz.com.br |
|
|
Re: i'll use 3 FCR's, end of question =)
Posted: Tue May 23, 2000 4:30 am
|
I think using 3 FCR's is the best idea, even if it is more difficult to do. Your fake cartridge will then have the best chance of working with all commercial game software.
I currently do not use ICQ. I'd prefer just to use e-mail. My e-mail address is e-quinn@elnet.com. You can e-mail the schematics to that address. Good luck. Eric Quinn |
|
|
|
|
|
Angus
|
ok, i'll be sending tomorrow
Posted: Tue May 23, 2000 4:52 am
|
|
ok, eric... i'll probably be sending the schematics tomorrow.
Even tough i WILL use 3 FCR's (now i'm pretty convinced about that!), based on what you said i am wondering if i need to connect inputs D7-D0 and IE (input enable) to FCR0, since it will always be set to page 00... but even if i don't connect those inputs, i still need a tri-state 8-bit latch, so i'll use a register anyway... but since ttl registers are always set to 00 on startup, i don't think FCR0 would need any data inputs other than the OE (output enable). do you agree? |
|
|
Re: so i don't need to initialize FCR1?
Posted: Tue May 23, 2000 9:14 am
|
It would be nonsense anyway. Each cartridge can technically include any custom features, you can't have one cartridge doing everything. |
|
|
|
|
|
Posted: Tue May 23, 2000 9:15 am
|
I'll be happy if you could send your schematics/infos on the forum. I am interested so many other people should be. |
|
|
|
|
|
Angus
|
i'm so stupid....
Posted: Tue May 23, 2000 11:36 am
|
|
Damn, i'm stupid! Eric, if you really think FCR0 is always set to page 0, then there's no need for a 5-bit latch for frame 0! I'll just use a tri-state buffer, with OE0'=(A14'.A15')' on it's output enable, 0 on its data input, and A18-A14 wired to it's tristate output. that would save me one or two IC's. are you ok with that? =)
|
|
|
Angus
|
yes or no?
Posted: Tue May 23, 2000 3:34 pm
|
|
so you mean i really DON'T need to initialize FCR 1, right?
|
|
|
Eric
|
Re: i'm so stupid....
Posted: Tue May 23, 2000 3:45 pm
|
Yes, that seems fine. Two FCR's (frame 1 and frame 2) should work with nearly all commercial software. (As we discussed before, 1 FCR is not enough.) Eric Quinn |
|
|
Angus
|
zoop, do you agree with eric?
Posted: Tue May 23, 2000 3:48 pm
|
|
i am looking over over at the schematics and i really wish i could use only one register for both frame 1 and frame 2. it would save a lot of IC's and a lot of connections...
eric, why would a software use a frame without maping a page for it? i am sorry, but i still don't get it.... =( zoop, do you agree with eric? do you believe i really need different registers for each frame, even though they are never used by the same cartridge? eric, a few more for you: do you think i must include the rom/ram register (FFFC), even if will not use my project with any game with back up battery ram? i didn't understand how the master system use this register.... is it located at the cartridge? |
|
|
Re: yes or no?
Posted: Tue May 23, 2000 3:56 pm
|
I feel you dont NEED, but anyway it is always better to do so. Will not hurt. |
|
|
|
|
|
Re: zoop, do you agree with eric?
Posted: Tue May 23, 2000 3:58 pm
|
I don't remember being asked this question actually, but OF COURSE! you need them both. And I don't see why they couldnt be used by the same cartridge.
You can forget about it in the first time I think. Maybe in the later stage of your project when normal games will work you could expand your cartridge to support battery backed ram. |
|
|
|
|
|
Angus
|
didn't eric say so?
Posted: Tue May 23, 2000 4:09 pm
|
well if i did understand it, Eric does say that on SMSARCH Document. he says 2 and 4 mega cartridge don't use frame 2... I could be wrong though... |
|
|
Re: didn't eric say so?
Posted: Tue May 23, 2000 4:42 pm
|
O.k., here's the situation, as I understand it: I built a cartridge reader. All it needed to do was read every byte from a cartridge. The person who helped me informed me that 2 and 4 Mega carts didn't implement the FCR for frame 2. The only frame that could have pages mapped in and out was frame 1, so only FCR 1 existed. This does not mean that frames 0 and 2 didn't have memory. As mentioned before frame 0 always mapped ROM page 0. Frame 2 could have any ROM page (most likely page 2 or 0). For the sake of a cartridge reader, though, it didn't matter, I could read any ROM page through frame 1 by writing to FCR 1. Similarly for 1 Mega carts, FCR 1 wasn't implemented, so I wrote to FCR 2 and read every ROM page (except page 0) from frame 2. Angus, your project is different, you are trying to build a fake cartridge. Consequently, you need to include both FCR 1 and FCR 2. Any software from a 1 Mega cart will try to write to FCR 2 to map ROM pages, and any software from 2 and 4 Mega carts will try to write to FCR 1 to map ROM pages. If you want your cartridge to work with both, you must include both FCRs. As far as intializing them, I believe the following is the best approach: Frame 0 should always have ROM page 0, Frame 1 should be initialized to have ROM page 1 (FCR 1 = 1). Frame 2 should be intialized to have ROM page 0 (FCR 2=0). I believe this is an approach used by many emulators, and it seems to work. Remember, though, just because an FCR doesn't exist, that does not mean that frame is not being used, it just means that whatever page is in that frame at start-up is never swapped out. I wouldn't assume that the memory page is not used. Good luck. Eric Quinn |
|
|
|
|
|
Angus
|
ok, i understand, but.... common adressing could be done
Posted: Tue May 23, 2000 9:49 pm
|
|
"Remember, though, just because an FCR doesn't exist, that does not mean that frame is not being used, it just means that whatever page is in that frame at start-up is never swapped out. I wouldn't assume that the memory page is not used."
This sentence is the only reason i see for not using only one common register for frame 1 and 2, so i'll probably build my circuit with both FCR1 and FCR2. But, if only one of the 2 frames were used for sure, i really believe i could implement only one register. There would be no addressing problem, i'll show you how: i'll refer to the intended common register for frame 1 and 2 as FCRX or FX, and to frame 0 as FCR0 or F0. so, the logical functions for enabling FCR writing and their outputs are: FXIE = A1.W (this enables writing for both FCR1 and FCR2) F1IE = A0.W (writing enable for FCR0, won't implement) FXOE = A15exorA14 (this enables outputs for both FCR1 & FCR2) F1OE = A15'.A14' (this enables outputs for FCR0) Am I totally wrong or could this really be implemented IF only frame 1 OR frame 2 were used by each cartridge? i agree that one frame could be used if it can't be swapped. probably page 1 is never mapped to frame 2, even on 1 mega cartridges, right? if that's so, then i won't argue anymore... or will find an alternate solution. |
|
|
Re: ok, i understand, but.... common adressing could be done
Posted: Tue May 23, 2000 10:55 pm
|
Yes, I believe you're right. This seems like it would work. However, I don't know if the WRITE signal is only asserted when writing to FCRs. The WRITE signal is probably a generic bus signal used for writing to RAM also. Simply testing A0, A1, A14, A15 and W might not be sufficient to determine when FCR's are being written. For example if software writes to RAM address 0xC002, the signal FXIE will be asserted enabling writing to the FCR, which you definitely don't want to do in this case. You may need to do improve the address decode, like this: FXIE = A15.A14.A13.A12.A11.A10.A9.A8.A7.A6.A5.A4.A3.A2.A1.W
That seems likely, but I have no evidence to support that claim. Eric Quinn |
|
|
|
|
|
Angus
|
schematics
Posted: Wed May 24, 2000 2:55 am
|
I do hope WRITE is not used for writing to RAM, or the circuit board would have to be a whole lot bigger!!! Gotta check that out.
it would be good if someone could check that out.... as you can see on my schematic, the circuit would be a lot smaller if I could use one FCR less. you guys can check my schematic at: http://www.spaceports.com/~vinci/smscart-rw1.gif tell me what you think about it. |
|
|
Re: didn't eric say so?
Posted: Wed May 24, 2000 9:05 am
|
That is cartridge hardware dependant. All games I disassembled used FCR1/FCR2 as they wanted.
Frame 0 can be mapped too, with the first kilobyte staying at 0. |
|
|
|
|
|
Angus
|
Re: schematics
Posted: Wed May 24, 2000 12:29 pm
|
|
i've found a wiring mistake on the schematics i sent you, so deleted it from the server. I'll fix it to today and send the corrected version, as well as posting it pack to that url.
|
|
|
Adam Klotblixt
|
Posted: Wed May 24, 2000 1:18 pm
|
|
I had plans on building a similar device, but scrapped the idea because of the complexity of the FCR's...
But yesterday I opened up my friends Phantasy Star and found 3 main chips: ROM, RAM and what I presume is an FCR memory-mapper!!! It would be a fairly quick hack to remove the ROM and replace it with a ZIF-socket and a 512K FLASH-EPROM or EEPROM. I'm going to the local flea-market on Saturday to pick up a 55xx, 75xx or 95xx cartridge and see if my suspicions are correct. Playing old games on an emulator is fun, but playing them on the real thing is MUCH better! /Adam |
|
|
Angus
|
schematics 2.0
Posted: Wed May 24, 2000 2:29 pm
|
|
i fixed the wiring errors and rearrenged the ic's, and the schematics are back online... let me know if you think something is still wrong or could not work...
http://www.spaceports.com/~vinci/smscart-rw1.gif |
|
|
Angus
|
how will you write the EEPROM
Posted: Wed May 24, 2000 2:39 pm
|
|
Have you already developed anything to write the binary SMS file to the EEPROM? That's a whole other project!
|
|
|
Posted: Wed May 24, 2000 3:39 pm
|
Be careful here. I do not think that third chip is a FCR memory-mapper. I suspect it only has something to do with the batter-backup RAM. Most cartridges (without battery RAM) only have a single ROM chip, implying the FCR's are embedded in the ROM. Phantasy Star is not the game you want to be experimenting with. May I suggest you try your experiment with another battery-backup RAM game, like Penguin Land or Miracle Warriors? Eric Quinn |
|
|
|
|
|
Adam Klotblixt
|
Posted: Thu May 25, 2000 6:01 am
|
Well, I do believe it is a memory mapper chip, it has too many legs to be a simple battery-monitor. There's also a load of analog parts (including a few transistors) that I suspect is doing the battery-regulation. I know that the RAM-less carts only have one chip (really pissed me off when I opened up my carts), so I'll just make sure i buy a RAM-cart (hence 55xx, 75xx or 95xx cart numbres). I already have an EPROM-programmer that does EEPROM and FLASH as well, so this is probably the fastest way for me to play real games on real hardware. It's the programmer found on http://www.willem.org/nprome.htm I recommend it highly since it's cheap and very versatile. /Adam |
|
|
Posted: Thu May 25, 2000 3:52 pm
|
Yes, it may be a memory mapper chip, but probably only for the FCR at 0xFFFC (the battery-backup RAM control register). Since all other cartridges have their FCR's embedded in the single cartridge ROM chip, I don't see any reason why Sega would create special ROM chips, just for battery-backup RAM games, with all the FCRs external. There's a good chance that the FCRs at 0xFFFD, 0xFFFE, and 0xFFFF are still embedded in the main ROM chip. In any case, good luck with your project. Keep us informed as to how it's going. Eric Quinn |
|
|
|
|
|
Not all games have the FCRs embedded in the ROM!
Posted: Sat May 27, 2000 6:20 pm
|
|
Not all games have the FCRs embedded in the ROM.
I have a "The Ninja" cartridge which contains a standard ROM plus a bankswitching device... there may be other games with this arrangement also. I've successfully modified the cartridge to take a standard 128k EPROM - see my recent post entitled "FCRs, Bankswitching and The Ninja". Mike |
|
|
|
|
