(WIP) New modchip coming in

Discussion about Modding the XBOX, including hardware and software hacks.
psyko_chewbacca
Posts: 213
Joined: Fri Oct 05, 2012 5:19 pm
Location: Québec, Canada
Has thanked: 21 times
Been thanked: 125 times

(WIP) New modchip coming in

Post by psyko_chewbacca » Sat Jun 28, 2014 2:11 pm


User avatar
cashonly
Posts: 181
Joined: Thu Jul 19, 2012 5:12 pm
Has thanked: 20 times
Been thanked: 32 times

Re: (WIP) New modchip coming in

Post by cashonly » Mon Jun 30, 2014 6:45 am

nice
These pretzels are making me thirsty!

psyko_chewbacca
Posts: 213
Joined: Fri Oct 05, 2012 5:19 pm
Location: Québec, Canada
Has thanked: 21 times
Been thanked: 125 times

Re: (WIP) New modchip coming in

Post by psyko_chewbacca » Fri Aug 08, 2014 5:26 pm

Hello,

I've set up the code repository over at bitbucket.
https://bitbucket.org/psyko_chewbacca/lpcmod_os

You're welcome to take a look or even ask to contribute.

dan.h
Posts: 133
Joined: Tue May 28, 2013 2:23 pm
Has thanked: 3 times
Been thanked: 10 times

Re: (WIP) New modchip coming in

Post by dan.h » Fri Aug 08, 2014 9:23 pm

Ah are you bennydiamond on assembler? If so amazing work on the modchip, if you sell a batch would love to get hold of one!

User avatar
GhostlyGamer
Posts: 356
Joined: Sun Jul 15, 2012 5:34 am
Location: Seattle USA
Has thanked: 16 times
Been thanked: 15 times

Re: (WIP) New modchip coming in

Post by GhostlyGamer » Sat Aug 09, 2014 9:28 pm

Oh my god. This is amazing. LCD support. Want so hard.
Image
Image

fxmech
Posts: 673
Joined: Wed Aug 01, 2012 9:15 am
Has thanked: 37 times
Been thanked: 46 times

Re: (WIP) New modchip coming in

Post by fxmech » Sun Aug 10, 2014 8:33 am

I'm anxiously awaiting the first run of it, so much so that I may take my first attempt at making my own board. It's very nice to see the developer have such an open approach to his project.

A friend of mine was also thinking we can try using the arduino as the basis of a chip with LCD support. A kind of advanced undertaking for us considering our time constraints as of late, however.

fxmech
Posts: 673
Joined: Wed Aug 01, 2012 9:15 am
Has thanked: 37 times
Been thanked: 46 times

Re: (WIP) New modchip coming in

Post by fxmech » Sun Aug 10, 2014 8:43 am

dan.h wrote:Ah are you bennydiamond on assembler? If so amazing work on the modchip, if you sell a batch would love to get hold of one!
According to psyko's link, it's implied he is the same person. On the page it states to contact psyko here or benny on assembler.

psyko_chewbacca
Posts: 213
Joined: Fri Oct 05, 2012 5:19 pm
Location: Québec, Canada
Has thanked: 21 times
Been thanked: 125 times

Re: (WIP) New modchip coming in

Post by psyko_chewbacca » Mon Aug 11, 2014 8:59 pm

fxmech wrote:A friend of mine was also thinking we can try using the arduino as the basis of a chip with LCD support. A kind of advanced undertaking for us considering our time constraints as of late, however.
If you want to interface the Xbox LPC bus, you will need something way faster than a Arduino. Capturing data going at 33MHz using a processor running at 16MHz (Arduino UNO) with an appreciable success rate is not possible.

Even using a Arduino Due with a Cortex-M3 at 84MHz will be a hard task. It would probably be impossible to do anything else with the Arduino other than interface a LCD screen.

IMO, the best approach is to use High speed programmable logic device to interface the LPC Bus. Small CPLDs cost way less than a fast ARM microcontroller and can work at faster clock frequencies.
fxmech wrote: According to psyko's link, it's implied he is the same person. On the page it states to contact psyko here or benny on assembler.
Yes I am, AssemblerGames' forums do not accept usernames with underscore. No loco chewy will venture onto the InterWebs without its underscore!

SPPV
Posts: 64
Joined: Fri Jul 13, 2012 10:47 pm
Been thanked: 7 times

Re: (WIP) New modchip coming in

Post by SPPV » Mon Aug 11, 2014 10:13 pm

When all is well drop me a line here, my site or Xtras, I will buy 50 units for resale in Canada. I want 1 unit to test and confirm and then its gravy will put em out via my store.

kolosus
Posts: 13
Joined: Sat Oct 19, 2013 1:43 am

Re: (WIP) New modchip coming in

Post by kolosus » Thu Aug 28, 2014 11:02 pm

I'd be interested in a few since I've bad flashed my TSOP on a few xboxes. Looking forward to this.

psyko_chewbacca
Posts: 213
Joined: Fri Oct 05, 2012 5:19 pm
Location: Québec, Canada
Has thanked: 21 times
Been thanked: 125 times

Re: (WIP) New modchip coming in

Post by psyko_chewbacca » Tue Sep 30, 2014 3:40 am

Hello everyone,

I have a little somehting for you.

It's a BETA version of my upcoming OS for my modchip; called XBlast OS.

The cool thing is that I provided the files to flash it to your modchip/TSOP or run it as an Xbox executable(XBE).
Please click here to get it!.

Take the time to read the included readme text file. I'm sure it will answer most of your questions.

User avatar
cashonly
Posts: 181
Joined: Thu Jul 19, 2012 5:12 pm
Has thanked: 20 times
Been thanked: 32 times

Re: (WIP) New modchip coming in

Post by cashonly » Tue Sep 30, 2014 6:33 am

Wow, the functions of the bios sounds good:

"3. Basic operation
XBlast OS has the ability to save settings and eeprom data
on your modchip's flash or onboard TSOP (W49F020 and
SmartXX modchips do not support saving). However, it is
important to know that no actual writing to physical devices
(modchip flash dan/or Xbox eeprom) will be made until a
software request for a reboot or a shutdown is made. This
allows you to freely edit settings and eeprom data and only
commit your changes upon exiting XBlast OS software by
issuing a software reboot/shutdown command or booting
from a flash bank. If you press the power button or unplug
your console, the settings you changed since last boot will
not be saved. Note that any changes made to the HDDs are
commited automatically; same goes for flashing a new Bios.

If no persistent settings are found on your modchip, the front
LED will flash Red/Orange. XBE version will always starts
flashing Red/Orange if no XBlast mod is detected.

LCD Settings menu will only be shown if a supported
modchip is detected. Sorry, no Aladdin XT 4064...

128MB Ram test will not be shown on 1.6/1.6b Xboxes.

Locking/Unlocking HDD option will not be shown if a
connected HDD does not support it.

Saving and restoring Xbox EEProm will not be shown in XBE
version, unless a XBlast Mod is detected.

Flashing a Bios from the HDD lists files from C:\BIOS only.

The Mainmenu screen (called IconMenu) contains several key
elements:
-Type of executable (XBE or ROM)
-Version (currently 0.1 BETA)
-Detected Modchip (currently, XBlast Mod Lite, Xecuter 3(CE) and all SmartXX modchips are supported)
-DEBUG_fHasHardware: modchip ID in hexadecimals
-Debug MB_string: Actual Xbox revision string to identify your motherboard revision.
-Conventionnal Xbox revision number.
-Amount of RAM detected, in Megabytes.
-A set of icons to choose by using left and right arrows on the D-pad and selecting with A.

The settings screen (called TextMenu) works in this manner:
-Navigate entries using Up and Down arrows on the D-Pad
-Enter selection or toggle option by pressing A or Start.
-Go back by pressing B or Back.
-Most settings can be changed using the D-pad's left and right arrows.
-Simple yes/no or True/false options can also be toggled by pressing A
-Settings with numerical values can be changed using the analog triggers(faster)"
These pretzels are making me thirsty!

User avatar
spicemuseum
Posts: 906
Joined: Mon Jul 09, 2012 11:08 pm
Has thanked: 94 times
Been thanked: 75 times

Re: (WIP) New modchip coming in

Post by spicemuseum » Fri Oct 03, 2014 1:57 pm

Chewy, does your new h/w design (and BIOS) support Chameleon/Matrix style flashing of TSOP? If not, could it?

fxmech
Posts: 673
Joined: Wed Aug 01, 2012 9:15 am
Has thanked: 37 times
Been thanked: 46 times

Re: (WIP) New modchip coming in

Post by fxmech » Fri Oct 03, 2014 2:18 pm

spicemuseum wrote:Chewy, does your new h/w design (and BIOS) support Chameleon/Matrix style flashing of TSOP? If not, could it?
Interesting. I have at least one Xbox that would be revived if it can be done... 8-)

xboxprince
Posts: 234
Joined: Mon Feb 17, 2014 1:48 pm
Has thanked: 16 times
Been thanked: 12 times

Re: (WIP) New modchip coming in

Post by xboxprince » Fri Oct 03, 2014 2:45 pm

psyko_chewbacca wrote:
fxmech wrote:A friend of mine was also thinking we can try using the arduino as the basis of a chip with LCD support. A kind of advanced undertaking for us considering our time constraints as of late, however.
If you want to interface the Xbox LPC bus, you will need something way faster than a Arduino. Capturing data going at 33MHz using a processor running at 16MHz (Arduino UNO) with an appreciable success rate is not possible.

Even using a Arduino Due with a Cortex-M3 at 84MHz will be a hard task. It would probably be impossible to do anything else with the Arduino other than interface a LCD screen.

IMO, the best approach is to use High speed programmable logic device to interface the LPC Bus. Small CPLDs cost way less than a fast ARM microcontroller and can work at faster clock frequencies.
fxmech wrote: According to psyko's link, it's implied he is the same person. On the page it states to contact psyko here or benny on assembler.
Yes I am, AssemblerGames' forums do not accept usernames with underscore. No loco chewy will venture onto the InterWebs without its underscore!
Arduino is coming out with a new board that has a 1GHz processor( arduino Tre), would that work?
project media center "part one" (working progress): 6tb trusty 1.4ghz + 128mb v1.4 Xbox water cooled pc case with xenium ice, LCD, HDMI out and optical out.

psyko_chewbacca
Posts: 213
Joined: Fri Oct 05, 2012 5:19 pm
Location: Québec, Canada
Has thanked: 21 times
Been thanked: 125 times

Re: (WIP) New modchip coming in

Post by psyko_chewbacca » Fri Oct 03, 2014 5:26 pm

spicemuseum wrote:Chewy, does your new h/w design (and BIOS) support Chameleon/Matrix style flashing of TSOP? If not, could it?
I would like to but I don't know the technicality of the process involved using Matrix chip. I know it requires a special Evox BIOS and that you wire the modchip to TSOP's A15 point. Other than that, what's being done to this pin? Why the A15 pin and no other(and why not A19 or A18...)? Do the special Evox BIOS sends special data to the modchip through the LPC port while doing its magic?
Unless someone points me to the proper documentation I can't do much about it sorry. If anyone can transfer all the knowledge to me, I would integrate it to the modchip of course.
xboxprince wrote:
psyko_chewbacca wrote:
fxmech wrote:A friend of mine was also thinking we can try using the arduino as the basis of a chip with LCD support. A kind of advanced undertaking for us considering our time constraints as of late, however.
If you want to interface the Xbox LPC bus, you will need something way faster than a Arduino. Capturing data going at 33MHz using a processor running at 16MHz (Arduino UNO) with an appreciable success rate is not possible.

Even using a Arduino Due with a Cortex-M3 at 84MHz will be a hard task. It would probably be impossible to do anything else with the Arduino other than interface a LCD screen.

IMO, the best approach is to use High speed programmable logic device to interface the LPC Bus. Small CPLDs cost way less than a fast ARM microcontroller and can work at faster clock frequencies.
fxmech wrote: According to psyko's link, it's implied he is the same person. On the page it states to contact psyko here or benny on assembler.
Yes I am, AssemblerGames' forums do not accept usernames with underscore. No loco chewy will venture onto the InterWebs without its underscore!
Arduino is coming out with a new board that has a 1GHz processor( arduino Tre), would that work?
I guess it would work but it would probably be the only thing it can do: decode LPC data and drive the LCD. You can only have a single threaded application that runs without any OS. Context switching is impossible without loosing data.
Here's why. Looking at the benchmarks made by TI for a Cortex-A8 (http://downloads.ti.com/dsps/dsps_publi ... times.html) interrupt latency is way to long to switch context when a hardware interrupt signal is received (start of a LPC transaction). I takes in average 966ns for your CPU to drop what he was doing and assert that there was an interrupt(you're not even starting to execute code in your interrupt handling routine). A whole 17 cycles LPC transaction takes 515 ns so you've missed it by much. Software-only interrupt(like changing task) takes less time but you'd still have to code small and low-priority tasks that executes for only a handful of cycles each time; not much to be done there.

Cortex-A8 is not the kind of architecture well suited for that type of work. It's designed as a low-power multi-purpose powerhouse. It's an other kind of embedded system when looking at smaller microcontrollers like PIC and AVR(or even Cortex-M serie).

So to sum up, I think it would be possible to do it with a 1GHz A8, just not very effective.

User avatar
professor_jonny
Posts: 1233
Joined: Thu Jul 05, 2012 5:41 am
Location: New Zealand
Has thanked: 56 times
Been thanked: 174 times

Re: (WIP) New modchip coming in

Post by professor_jonny » Fri Oct 03, 2014 9:47 pm

There is magic in the matrix bios that makes it possible to flash the tsop there is also the d6 tsop bios that works the same as the matrix bios with some exceptions.

with the d6 bios you boot holding D0 and A15 low then let d0 go when the eject light flashes green.
then at the splash screen with the yellow/blue image let go of A15 them and flash the bios you want with any flash program.
NOTE: this will only work if the bios contains a valid bootable image retail or other, no idea why tho.

I don't know what makes the matrix bios /chip combo different but it will flash a dead bios I believe.

but the 3 wire trick works some times on a big tsop as a work around.

This may be handy:

http://forums.xbox-scene.com/index.php? ... c508-code/

A15 is one of the address lines on the tsop and it makes the xbox skip reading specific data for what reason I don't know, it may make it stop reading the end of the bios who knows.

fxmech
Posts: 673
Joined: Wed Aug 01, 2012 9:15 am
Has thanked: 37 times
Been thanked: 46 times

Re: (WIP) New modchip coming in

Post by fxmech » Sat Oct 04, 2014 4:01 pm

I would be happy about the modchip booting a bad TSOP board and reflashing it.

Matrix does that. I recall reports of the cheapmod/26 wire being able to do this as well. I am unsure how the processes or logic work.

It's a pretty cool feature, but a modchip would inherently fix a bad TSOP in the first place. Likely won't be used by too many people...

User avatar
Xphazer
Posts: 522
Joined: Wed Jul 04, 2012 4:39 am
Location: Montréal
Has thanked: 233 times
Been thanked: 99 times

Re: (WIP) New modchip coming in

Post by Xphazer » Sat Oct 04, 2014 9:51 pm

Amazing work psyko_chewbacca! I can't wait to have one!
Personally what interests me the most is Boblight support. :wink:

If you don't have any Matrix or Chameleon I have a Chameleon and would be more than happy to lend it to you if you're interested in reverse engineering what the A15 pad does on the modchip.
I also have couple other modchip I can lend you for testing if you're interested (X3CE, Xenium, Xenium(clone), X-Bit, X-Chip and a couple of 3rd gen).

Also, slightly off-topic but seeing that you have spoken about TSOP splitting in your ASSEMbler forum thread, what do you think of this? http://www.eurasia.nu/modules.php?name= ... le&sid=183

psyko_chewbacca
Posts: 213
Joined: Fri Oct 05, 2012 5:19 pm
Location: Québec, Canada
Has thanked: 21 times
Been thanked: 125 times

Re: (WIP) New modchip coming in

Post by psyko_chewbacca » Mon Oct 06, 2014 3:39 am

professor_jonny wrote:There is magic in the matrix bios that makes it possible to flash the tsop there is also the d6 tsop bios that works the same as the matrix bios with some exceptions.

with the d6 bios you boot holding D0 and A15 low then let d0 go when the eject light flashes green.
then at the splash screen with the yellow/blue image let go of A15 them and flash the bios you want with any flash program.
NOTE: this will only work if the bios contains a valid bootable image retail or other, no idea why tho.

I don't know what makes the matrix bios /chip combo different but it will flash a dead bios I believe.

but the 3 wire trick works some times on a big tsop as a work around.

This may be handy:

http://forums.xbox-scene.com/index.php? ... c508-code/

A15 is one of the address lines on the tsop and it makes the xbox skip reading specific data for what reason I don't know, it may make it stop reading the end of the bios who knows.
Well that gives some explanation on how this is done but I still cannot understand the logic behind it. Was the EvoX D6 BIOS created by the same guys who made the Matrix and Chameleon chips? If so, that could at least understand how A15 and D0 lines on the chip could be controlled and toggled by the modchip itself as the D6 BIOS boots.

From what I read, Chameleon requires you to solder both D0 and A15 and will simply do all the work as opposed to the matrix chip where you need to physically play with the wires. How can the chameleon chip knows when to toggle a signal line and make it work most of time without some signal triggering it at a precise time? My guess is that the D6 BIOS send commands to the modchip via the LPC bus to toggle

I'll have to check that D6 BIOS if it sends any data on the LPC bus. Currently, I cannot think of any way to make this work without manually handling wires.
Xphazer wrote:Amazing work psyko_chewbacca! I can't wait to have one!
Personally what interests me the most is Boblight support. :wink:

If you don't have any Matrix or Chameleon I have a Chameleon and would be more than happy to lend it to you if you're interested in reverse engineering what the A15 pad does on the modchip.
I also have couple other modchip I can lend you for testing if you're interested (X3CE, Xenium, Xenium(clone), X-Bit, X-Chip and a couple of 3rd gen).

Also, slightly off-topic but seeing that you have spoken about TSOP splitting in your ASSEMbler forum thread, what do you think of this? http://www.eurasia.nu/modules.php?name= ... le&sid=183
Thanks for the support! I too would like to see some boblight action but I think I saw too big when I came up with my first design. I backtracked a little and starting with a simpler device for now. Once the XBlast Lite is stable and mature, then I'll start on making my original design work. The biggest issue is making the CPLD hold up the LPC bus while the ARM microcontroller makes the transaction. I have something good going on but it's not quite there yet. I might have to get a beefier CPLD for this. First, let me get that XBlast Lite out in the wild. Then we'll talk about that boblight support!

I don't have either a Matrix or a Chameleon. I don't want to borrow them(for now at least) but if you could take your chameleon and trace the signal from the A15 pad up to the CPLD pin. Does it go through a transistor or logic IC? Are there any resistors (pull-up/down), capacitors on the same signal trace? Any info on how the the A15 signal line is designed could give be some clue on how to make this work.

About the TSOP slitting, I did read the article you linked some time ago. The guy does make a valid point about the fact that only grounding the A19,A18 lines are a bad thing when you leave them floating to boot different banks. You need to force those lines either way, ground or VCC but not leave them floating when you have different BIOS in the same flash chip.

Technically, he's right when he says that the Xcodes read by the MCPX should come from the same binary image as the X86 code executed but the reality is that once the Xcodes are executed, your Xbox will either start executing x86 code or not. There's a great deal of chance that the Xcodes contained in a hacked MS BIOS be identical to the ones contained in another hacked MS BIOS. Even if they're read from a cromwell BIOS, I don't think it would really matter; your Xbox does boot with Cromwell no? I guess this issue was raised during the development of the Cromwell BIOS but is now a little over cautious since all Xboxes with flashable TSOP boot from the same set of Xcodes and they seem to produce a stable result!

But anyway, this whole Xcode issue doesn't matter if you simply use double throw switch that will force the address line to either GND or VCC and never leave them floating. That's what the XBlast modchip does (theorically, did not test that feature yet). If you select to control the TSOP using the XBlast, the modchip will either apply VCC or GND depending on which bank you want to boot. If you don't want to split the TSOP, modchip will simply stay in high impedance and leave the control of the address line to the MCPX chip.

And the final message in the article linked that states your MCPX will explode if you apply VCC to A18 and/or A19... Come on... You've all been constantly forcing D0 to ground to boot from modchip and I never heard any story about an exploding flash chip... Why would it be so different for the MCPX. Even worst, people ground LFRAME signal on 1.6(b) that comes from the MCPX and that signal does source a lot of current. Even after so many years, I'm fairly confident that most 1.6 Xbox can still boot from onboard BIOS if you remove the modchip!

Anyway, the worst that can happen in my opinion is that you'll burn A18 and A19 line drivers in MCPX chip by forcing them. That won't happen for a very long time since these are only driven when talking on the ISA bus(only device on the ISA bus is the TSOP flash chip). The rest of the time, these pins are tri-stated. Even so, why would it matter if those 2 lines are burned inside the MCPX chip, you control them manually to boot different banks now. Unless you go back to a single 1MB bank, you won't ever need the MCPX to control these anymore.

Post Reply