Resize 1024kb bios to 256kb or 512kb to flash bios split...

[verawarez]

Hi guys now im recently splitting 1.0 bios to 4 banks, i backup bios with evox but gives me 1024 size backup

I know that with raincoat it doesnt autosize to 1024kb but the problem i dont know how to use it

Many questions...

1 could i split the 1024 bios to 4x 256kb bios? or 2x 512kb ones?

2 can i backup 256kb or 512kb bios size (tsop splitted one)

3 is there any option to have a 256kb backup of 1.0 1.1 bios xbox?

last, is there any option or tutorial to get raincoat working? or to make backup with raincoat?

[Marty]

You don't split 1mb BIOS files. You split the TSOP chip with a BIOS flasher.

If you want a 4-way BIOS option, you'll need to wire your XBOX up for a 4-way split and flash 4 individual 256kb BIOS files.

Not sure why you'd want 4 different BIOS options though....

[soilengreen]

[Bowlsnapper]

10 hours ago, verawarez said:

Hi guys now im recently splitting 1.0 bios to 4 banks, i backup bios with evox but gives me 1024 size backup

I know that with raincoat it doesnt autosize to 1024kb but the problem i dont know how to use it

Many questions...

1 could i split the 1024 bios to 4x 256kb bios? or 2x 512kb ones?

2 can i backup 256kb or 512kb bios size (tsop splitted one)

3 is there any option to have a 256kb backup of 1.0 1.1 bios xbox?

last, is there any option or tutorial to get raincoat working? or to make backup with raincoat?

If you're talking about 4 DIFFERENT bioses, that sounds like a lot of work to avoid buying a chip. lol.

[verawarez]

I'm doing some testing, if a bank is not flashed properly i could restore just changing to other bank

I have some modchips too anyway, thankx

[Bowlsnapper]

17 hours ago, verawarez said:

I'm doing some testing, if a bank is not flashed properly i could restore just changing to other bank

I have some modchips too anyway, thankx

Was just makin a joke, my apologies.  Well if it's just for fun and experimental, then lets see if we can do it! Didn't mean to make it sounds like a dumb idea. If you can pull it off, you would probably be the first! So you're saying theoretically, you could boot XBlast from one bank and try flashing again if something goes wrong. But if XBlast detects a certain chip type, it will probably try to flash a full 1MB again not just 256, so you would have to keep that in mind. But I'm sure you know that.

[WildMage]

So, basically, when a part of a TSOP bank is grounded, the Xbox can't see that part of the chip anymore.  By grounding specific parts of the chip away like that, the Xbox will skip over those parts and boot the first bios file it sees.  So to accomplish what your shooting for, you want to build your custom 1024k bios file that contains all four banks, by just simply combining all the 256k files together into a single large file.  This is easy to do from a DOS or Linux command prompt.

 

In DOS you use something like 'copy biosA.bin + biosB.bin + biosC.bin + biosD.bin bigbios.bin'. Then you take the 1024k bios file and flash it to your TSOP with no switches active, so the chip is ungrounded and you have access to the entire 1024k.  Then you would use your switches to access the specific bank you want to load from.

 

That's how I've always done it at least.

[Bowlsnapper]

4 hours ago, WildMage said:

So, basically, when a part of a TSOP bank is grounded, the Xbox can't see that part of the chip anymore.  By grounding specific parts of the chip away like that, the Xbox will skip over those parts and boot the first bios file it sees.  So to accomplish what your shooting for, you want to build your custom 1024k bios file that contains all four banks, by just simply combining all the 256k files together into a single large file.  This is easy to do from a DOS or Linux command prompt.

 

In DOS you use something like 'copy biosA.bin + biosB.bin + biosC.bin + biosD.bin bigbios.bin'. Then you take the 1024k bios file and flash it to your TSOP with no switches active, so the chip is ungrounded and you have access to the entire 1024k.  Then you would use your switches to access the specific bank you want to load from.

 

That's how I've always done it at least.

However, XBlast will have a hard time being used because it has to boot with the intended flash area fully exposed. This would have to be a one time TSOP flash. There will only ever be one chance to do it.... Unless a programmer is involved. Or, perhaps XBLAST will see the TSOP and detect it properly regardless of how much is exposed and the proper switch configuration can be engaged after XBlast has booted from its 256K bank.

[KaosEngineer]

On 3/22/2024 at 5:43 AM, verawarez said:

Hi guys now im recently splitting 1.0 bios to 4 banks, i backup bios with evox but gives me 1024 size backup

I know that with raincoat it doesnt autosize to 1024kb but the problem i dont know how to use it

Many questions...

1 could i split the 1024 bios to 4x 256kb bios? or 2x 512kb ones?

2 can i backup 256kb or 512kb bios size (tsop splitted one)

3 is there any option to have a 256kb backup of 1.0 1.1 bios xbox?

last, is there any option or tutorial to get raincoat working? or to make backup with raincoat?

All stock Xbox BIOSes and most modified Xbox BIOSes are only 256 KBs in size. 

Microsoft (MS) installed a 1MB flash chip on v1.0 and v1.1 motherboards and Evox dumps the entire flash memory chip - 4 copies of the same BIOS.  And, with v1.2 and later motherboards MS changed the flash memory chip to 256KBs; however, the replacement dashboard's backup routine did not change and the same chip is accessed 4 times to dump a 1MB file too.  This occurs because the memory address decoding is incomplete throughout the entire 16MBs of memory space MS allocated for the TSOP.

For a v1.0 or v1.1s 1MB flash, it can be accessed at 16 different evenly divisible 1MB addresses throughout this space to dump the 1MB flash memory chip.

And for v1.2-1.4s, the 256KB flash can be accessed at 64 different 256KB starting addresses to dump the same data.

 

If the BIOS's base size is 256KBs, you can split the file into 4 256KB sections one after the other to obtain 4 copies of the same BIOS.

 

 

 

[WildMage]

On 3/30/2024 at 11:07 AM, Bowlsnapper said:

However, XBlast will have a hard time being used because it has to boot with the intended flash area fully exposed. This would have to be a one time TSOP flash. There will only ever be one chance to do it.... Unless a programmer is involved. Or, perhaps XBLAST will see the TSOP and detect it properly regardless of how much is exposed and the proper switch configuration can be engaged after XBlast has booted from its 256K bank.

Yeah, I didn't think about this, but an XBlastOS bios would be pretty pointless for a split TSOP, unless it was on the first and therefore ungrounded bank, wouldn't it? I've only ever had luck flashing the entire ungrounded TSOP at once. If any switch is hit, the flash program just couldn't even see the TSOP, although I admit I've never messed with anything but the Xblast xbe, and standard gaming bios files.

 

I used my switched 1.0 bios to have one VGA bios, and an almost identical bios that was not VGA, and maybe another bank to switch to a different OS, like one that boots into Avalaunch instead of XBMC.  It was all done pretty much because I could though, and it never saw much real world use.

[Bowlsnapper]

1 hour ago, WildMage said:

Yeah, I didn't think about this, but an XBlastOS bios would be pretty pointless for a split TSOP, unless it was on the first and therefore ungrounded bank, wouldn't it? I've only ever had luck flashing the entire ungrounded TSOP at once. If any switch is hit, the flash program just couldn't even see the TSOP, although I admit I've never messed with anything but the Xblast xbe, and standard gaming bios files.

 

I used my switched 1.0 bios to have one VGA bios, and an almost identical bios that was not VGA, and maybe another bank to switch to a different OS, like one that boots into Avalaunch instead of XBMC.  It was all done pretty much because I could though, and it never saw much real world use.

Yeah, the XBE, perhaps, but you  would have to build the bios on the PC every time, as long as you were okay with that.

[WildMage]

So, I did some testing with my 1.0 with a single TSOP split switch (512k/512k), with Xblast.xbe.  Xblast correctly identifies the TSOP chip, regardless of the switch is in the A position or the B position.  However, when flashing, it will crash and freeze halfway through flashing the A position, or it will complete the flashing and reboot on the B position.  Either way, it will flash the bios (mostly) correctly even if it freezes.

I say mostly, because I then ran the latest BiosChecker XBE with CRC codes added for the various version Cerbios files I was testing, and it resulted in very strange results, some that showed obvious corruption.  One of the switch positions showed 256k slot A as what was really in slot A, then position 2 as what was in slot B, then position 3 as what was in slot A, and position 4 as what was in slot B.  The other switch position showed something like a correct slot 1, and correct slot 3 for what was flashed, but a CRC that didn't match anything flashed to it in slot 2 or slot 4.

I was able to recover from a bad flash, by completely booting the system, and then completely loading Xblast, and THEN flipping the switch from the bank that worked, to the bank that wouldn't boot, and then flashing in Xblast.  I had to do this, because some test BIOS'es I flashed just failed to boot at all.

Due to this testing, I would go with what was pointed out by @KaosEngineer in the 2017 'TSOP splitting' thread, that explained "Why you shouldn't do TSOP Splitting": https://web.archive.org/web/20070205091445/http://www.xbox-linux.org/wiki/Why_you_shouldn%27t_do_TSOP_Splitting 

After running on the split bios on an old, fairly beat up system, I attempted to use XCAT 1.0 to archive what was on a few different pulled HDD's, and after running for sometimes hours, it would eventually just crash and completely shut down long before the archiving was complete.

The crashing might be unrelated, due to the console being beat as hell, but the results of the BiosChecker were very odd and disconcerting either way, and the article linked above does kind of explain the dangers of attempting to split a bios, and how it's honestly not really worth it, if stability is important to you.