14 replies to this topic

[Xenepp]

I have an Iron Soldier cart here which has never worked for me. I know the Jag is ok since Cybermorph and Doom work fine. This cart has some corrosion around the mask rom pins and continuity to the cart edge seems good and I have gotten the edge gleaming so I suspect one of the mask roms has bitten the dust.

So I have a programmer and some large eproms, but I've been trying to figure out how to split them for burning (and likewise dumping for testing in an emulator).

I saw a Brett Hull proto cart and it had the sum16s right on the labels so I figured that would make life much easier, I could split the rom for that game in various ways until I got the right checksum but after splitting in half, by byte and by word, both little and big endian, I still can't get a proper match. I came close by splitting by word but close isn't correct so I assume there's something I'm missing (unless the publicly available dumps are bad or a different revision?).

I realise I can just buy a working cart as it's not a rare game but since I enjoy repairing games I'd much rather try that as I have all I need to do it right here.

Edited by Xenepp, Mon Apr 25, 2011 3:12 PM.

[GroovyBee]

You've sorta answered your own question. Its a Brett Hull proto cart, so you'd need the exact same ROM image taken from the cart (in the image in question) for the checksums to match. I suspect that the final game's ROM image is going to have different checksums.

[Xenepp]

GroovyBee, on Mon Apr 25, 2011 3:23 PM, said:

You've sorta answered your own question.

Not really, I'm asking how to split a game into two roms for burning to eproms as Iron Soldier is a two ROM cart.

Also I'm using the proto rom of Brett Hull to test with so it should match since as far as I can tell, the picture I found was the same cart that is available in the NonGoodJag set which is explicitly marked as a prototype.

Hi ROM = 8044
Lo ROM = AD3F

There are also two Brett Hull dumps in the nongoodjag set and neither match no matter what I do. They both have to be protos too since it was never released.

Quote

Its a Brett Hull proto cart, so you'd need the exact same ROM image taken from the cart (in the image in question) for the checksums to match.

I assume you mean the same revision, not the same cart since the checksum should only change when the original data does, not the physical roms.

Edited by Xenepp, Mon Apr 25, 2011 3:47 PM.

[Chuplayer]

Maybe this would help? http://www.atariage....to-real-eproms/

If you get it working, please let me know how you did it. I'd really love to know. I'm thinking about getting an EPROM programmer myself, but I don't know a thing about splitting the files.

Edited by Chuplayer, Mon Apr 25, 2011 5:29 PM.

[Tyrant]

You'll want to use an Atari tool by the name of ROMSPLIT, but I'm afraid I don't know where you can find it. It's not something I've had need of.

[CyranoJ]

We might be able to implement that in JiFFI

I will suggest it to GGN.

[Chuplayer]

Tyrant, on Mon Apr 25, 2011 5:50 PM, said:

You'll want to use an Atari tool by the name of ROMSPLIT, but I'm afraid I don't know where you can find it. It's not something I've had need of.

Is it a command line DOS thing? Because I believe I found an ST-based ROMSPLIT. I guess you could use that in an emulator. I don't see anything in the readme talking about the Jaguar, though.

Attached Files

•  ROMSPLIT.ZIP   11.52K   20 downloads

[ggn]

Frankly I'm a bit puzzled here. Every eprom/eeprom/etc programmer software worth its salt usually has ROM file splitting built it.

Did you build the programmer yourself?



Also, since the program above is written by Mr Crashware himself, I recommend against using it


[EDIT]D'oh! I forgot a sentence there

Edited by ggn, Tue Apr 26, 2011 2:02 PM.

[Xenepp]

CyranoJ, on Mon Apr 25, 2011 6:02 PM, said:

We might be able to implement that in JiFFI

I will suggest it to GGN.

That would be useful. I discovered by taking apart and dumping another cart with two roms that they are 27C160 equivilents and need to be byteswapped and be split word-wise (no idea what you'd need to do for four rom carts).

My only problem now is I think it either needs the header stripping (JiFFI doesn't see one) or needs to start at a different address thanks to the encryption since when I burned new roms it red screens on the Jag. It waits a moment before red screening though so that's more than it did before.

It'd be good to see this run after having this sat about for over a year. I'm open to suggestions!

[Chuplayer]

Xenepp, on Thu Apr 28, 2011 4:57 PM, said:

CyranoJ, on Mon Apr 25, 2011 6:02 PM, said:

We might be able to implement that in JiFFI

I will suggest it to GGN.

That would be useful. I discovered by taking apart and dumping another cart with two roms that they are 27C160 equivilents and need to be byteswapped and be split word-wise (no idea what you'd need to do for four rom carts).

My only problem now is I think it either needs the header stripping (JiFFI doesn't see one) or needs to start at a different address thanks to the encryption since when I burned new roms it red screens on the Jag. It waits a moment before red screening though so that's more than it did before.

It'd be good to see this run after having this sat about for over a year. I'm open to suggestions!

Somebody in one of the links I posted said to strip the first 8kB off of the ROM or pad it with 8kB of zeroes. Jiffi has a headerless ROM option, and it produces a .rom that is 8kB shy of 4MB.

So, I don't know. Hex editor? Overwrite to 1FFF with zeroes? I don't really understand what headers have to do with physical Jaguar carts.

[Tyrant]

The header is there as a security measure to ensure Atari got paid royalties for each title released on the system. It consists of a cryptographic hash of the cartridge contents. A retail Jaguar won't boot without the header being in place and valid. There is a universal header you can use to replace it, if, for some reason, the rom image you're trying to burn doesn't have it's original header.

[LinkoVitch]

It would require you replace both ROMs in that case, I may have miss-read but are you not after replacing just one?

If you have the ability to dump carts, perhaps it may be worth dumping your faulty one, then compare with hex it vs a working dump of the cart. You should be able to determine from this which bit(s) are not correct and therefore identify which tracks/pins are not happy possibly for closer inspection/resoldering.. may just be a dry/cracked joint.

If they are 16bit chips I would imagine each chip gets one 16 bit word correct endianness for a 68K (I can never remember which way around it is )

I have attached an EXE version of ROMSPLIT I found in my archives of random, in-case that is of any help.

HTH

Attached Files

•  romsplit.zip   17.5K   20 downloads

[Xenepp]

Well it turns out I was being a bit of a twit, I accidently saved the original file as half it's size after doing the byteswap so when I burned the EPROMs they were only half full.

I now have a working Iron Solder cart. Thanks for your help folks.

So to recap, all I did was byteswap the rom file, then dissect it by word using winhex and then write those to two 27c800 eproms. The first file spit out from winhex being the high words, the second being the low (top rom, bottom rom on the cart).

For those interested, both mask roms were faulty. Several bits were incorrect on both when comparing them to the roms I found online and ultimately used.

[Chuplayer]

Thanks for the info, Xenepp. One quick question, what speed EPROMs did you use? Does it matter for the Jaguar? I've seen 100ns, 90ns, 50ns, and I can't find out if it matters for the Jaguar.

[Zerosquare]

It doesn't. ROMs on the Jaguar are allowed a 10 cycle latency by default ; since the clock is about 26.6 MHz, that gives a maximum access time of about 376 ns, which even the slowest EPROM can handle without problems.