Last july on TI-Planet, we supposed that the next Nspire OS was going to include support for a 16th language, arabic.

This evening, we've finally found some evidence of it:

So let's switch the language to arabic, and discover the first captures of our next Nspire OS:


Wow!
It's not just a translation like when two chinese languages were added in OS 3.0... The home screen has completely been inverted horizontally!

Well, not just the home screen... All the interface items are switched from right to left too!

The input has been adapted too: in arabic mode, when you enter some text it is inserted from the right of the screen, and by default right-aligned.

Note the "AR" mode in the top left-hand corner.
Enabled and disabled by the key, it lets you directly input arabic characters using the alphabetical keys.



Such a major modification of the interface in a few months seems quite impressive. Was it planned from the start with some kind of a stylesheet?...





Source:
http://tiplanet.org/forum/viewtopic.php?f=43&t=10818

Many TI-Nspire ClickPad and TI-Nspire+/CAS+ prototypes have been freely given by TI to students, teachers or engineers for beta-testing for years.

This became a problem when some testers tried to resale their prototypes on online shops, as they could easily be mistaken for production models, although they had early versions of the OS and couldn't be easily upgraded.


It seems the beta-testing of TI-Nspire TouchPad or TI-Nspire CX went different, as the same problem didn't happen.


But anyway, here is it finally - discover our first TI-Nsprie color prototype, the TI-Nspire Color, an EVT (Early Validation Tests) prototype of the non-CAS TI-Nspire CX:




Expect great discoveries soon, and why not some hints on the TI-84 Plus C, the future color version of the TI-84 Plus, as something happened at that time: non-CAS TI-Nspire CX dropped the TI-84 Plus emulator support, although the TI-84 Plus ROM remained available in the 3.0.1 non-CAS CX OS.



Source:
http://tiplanet.org/forum/viewtopic.php?t=10723

In a previous news, we discovered that Boot1/Boot2 version numbers couldn't be trusted to determine if a TI-Nspire handheld was a prototype or not.

Yes, we were told about a TI-Nspire prototype running both Boot1 1.1.8916 and Boot2 1.1.8981.



Today, we've finally got such a prototype between our hands.

And we are beginning to understand - let's check the OS:


1.4.8092... This prototype OS is more recent than everything we've found up to now.

It seems that the goal of the prototype versions of Boot1 1.1.8916 and Boot2 1.1.8981 was enabling TI engineers to go on testing trustfully development OSes on their prototypes.



But there is more - just have a closer look at the OS version string:
Full version string is "1.4.8092 debug" - this time it's not a simple development OS sent for testing to teachers/students...
It's a debug development OS coming from the TI development team!

A very nice find - as usually, prototypes used internally for development at TI have to be destroyed once they're no longer needed.



So what's special about this OS?
It creates a debug file "/documents/navnet.log.csv" where is stores lots of informations about the NavNet link engine.
Logged lines are mirrored to the RS232 console and all start with [64].

Just check the OS bootlog:
Beginning system initialization.
Preparing file system. This takes a while ...
   POSIX layer initialized.
   POSIX initialized devices.
Datalight Reliance v2.10.1150
Copyright (c) 2003-2006 Datalight, Inc..
Datalight FlashFX Pro v3.00 Build 1358
Nucleus Edition for ARM9
Copyright (c) 1993-2006 Datalight, Inc..
Patents: U.S. # 5860082, # 6260156 U.S..

   POSIX file system initialized.
File system ready.
* P1R2 battery door detection method
System build date: May 13, 2008, 11:30:55 p.m.
Available memory: 18624748 bytes
Purging temporary files ...
Launching system ...
Setting log file to / documents / navnet.log.csv with a buffer of 4096 bytes
Setting Console Log Level = 3
RET - UDP Echo: TI_Echo_UDP_Init Called + + +
RET - ECHO UDP Socket ID We created Successful. Socketid = 1
RET - UDP ECHO: TI_UDP_Echo_BindIPAddr + + +
RET - UDP ECHO: Successful NU_Bind
RET - UDP ECHO: --- TI_UDP_Echo_BindIPAddr
RET - UDP ECHO: EchoUDPListenerThread. Inside Thread + + +
RET - UDP ECHO: Init Called ---
[64] Register UDP stream connector, result = 0
[64] Called TI_NN_Init. LowAPI LowAPI LowAPI 02-06-2008
[64] CM: init_conn_mgr: Entered (NavStack: connection manager: 1/31/2008).
[59] CM: processLocalNodeRecvDeferredQueue: Entered (NavStack: stream manager:
/ 22/2008).

[64] Entered TI_NN_StartService. Service Id: 5000 + + +
[64] Start service: port 5000 '. NS_SVC.5000' service.
[64] End of TI_NN_StartService. 5000 Service started. ---
[64] Start sync services
[64] Entered TI_NN_StartService. Service Id: 4060 + + +
[64] Start service: port 4060 '. NS_SVC.4060' service.
[64] End of TI_NN_StartService. 4060 Service started. ---
[64] MS: MessageServiceRegister called.
[64] Entered TI_NN_StartService. Service Id: 4051 + + +
[64] Start service: port 4051 '. NS_SVC.4051' service.
[64] End of TI_NN_StartService. 4051 Service started. ---
[64] MS: MessageServiceRegister called.
[64] MS: MessageServiceRegister called.
[64] MS: MessageServiceRegister called.
[64] MS: MessageServiceRegister called.
[64] MS: MessageServiceRegister called.
[64] MS: MessageServiceRegister called.
[64] MS: MessageServiceRegister called.
[64] MS: MessageServiceRegister called.
[64] OS install Start Service
[64] Entered TI_NN_StartService. Service Id: 4080 + + +
[64] Start service: port 4080 '. NS_SVC.4080' service.
[64] End of TI_NN_StartService. 4080 Service started. ---
[64] Start node info service.
[64] Entered TI_NN_StartService. Service Id: 4020 + + +
[64] Start service: port 4020 '. NS_SVC.4020' service.
[64] End of TI_NN_StartService. 4020 Service started. ---
[64] Entered TI_NN_StartService. Service Id: 4042 + + +
[64] Start service: port 4042 '. NS_SVC.4042' service.
[64] End of TI_NN_StartService. 4042 Service started. ---
[64] Start screen Service
[64] Entered TI_NN_StartService. Service Id: 4024 + + +
[64] Start service: port 4024 '. NS_SVC.4024' service.
[64] End of TI_NN_StartService. 4024 Service started. ---
[64] Entered TI_NN_StartService. Service Id: 4052 + + +
[64] Start service: port 4052 '. NS_SVC.4052' service.
[64] End of TI_NN_StartService. 4052 Service started. ---
[64] Called TI_NN_RegNotifyCB. + + +
[64] CM: RegisterNotifyCallback called.
[64] End of TI_NN_RegNotifyCB. Notify Callback registered. ---
[64] New API: Load Effective connector all services started.
[64] Initializing 2 Connector (s).
NavNet Ready.

Something which could have been very useful if we got it sooner, when adding TI-Nspire support in TiLP for example.



In fact, this is not our first TI-Nspire debug OS.

Summer 2011 at TI-Dallas, Adriweb found a TI-Nspire running a debug version of the 1.8 Nspire OS, which was never released as we went straight from 1.7 to 2.0:



Sources:
http://tiplanet.org/forum/viewtopic.php?t=10671
http://tiplanet.org/forum/viewtopic.php?t=10669

In a previous news, DJ_O told you Extended was working on USB support in Ndless.




Since, Tangrs has started a Linux port for the TI-Nspire CX.

He's now started working on USB support and has managed to use an USB keyboard and to play videos on an external USB storage device!




Development topic:
http://ourl.ca/17131/321577;topicseen#new

Tonight, let's finally discover the long awaited images of the inside of a TI-Nspire CM-C, the one Adriweb went searching in Asia before giving lending it to me after coming back.

As already predicted in our previous news, you will discover a motherboard organized very differently from the TI-Nspire CX one:

Unlike the TI-Nspire CX, there are two boards on the TI-Nspire CM instead of three on the TI-Nspire CX:

• the screen borad (motherboard) is a RABBIT_MB_DVT_6420 on the TI-Nspire CM instead of a FireBird_Color_MB_6422 on the TI-Nspire CX
  
• the keypad board is a RABBIT_KB_NONCAS_DVT_4420 on the TI-Nspire CM instead of a FireBird_Noncas_Color_KB_EVT_4421 on the TI-Nspire CX

It seems that we were right once again, when announcing in a previous news that the TI-Nspire CM code name was Rabbit

As in another previous news, we notice the keypad board is once again specific to the non-CAS model. The CAS keypad board would probably be a RABBIT_KB_DVT_4420.



The third card on the TI-Nspire CX which is missing here on the TI-Nspire CM is the dock connector.

However, the J4 test / debug connector is there in the bottom right-hand corner of the screen board and should be usable in place of the dock connector to get the bootlog, like we did on the TI-Nspire Lab Station Cradle.



The TI-Nspire CM is showing us two main chips on the screen board instead of three on the TI-Nspire CX:

The visible chips are:

• the ASIC
• a TI 6C053A1 present on all Nspire models since the TI-Nspire TouchPad whose purpose is unknown

We allready said that the ASIC was including the processor but also the Flash NOR ROM chip (boot1) and the RAM.

The first question which comes to mind as the TI-Nspire CX have 64MB of RAM and the TI-Nspire CM 32MB is whether ASIC is the same or not ... If it is the same, it means that there are 32MB of RAM which have been disabled some way...



The chip which is lacking is the 128MB Flash NAND ROM. Was it moved to the ASIC too? Or was it just moved on behind the board? Let's see...
Yes, the missing 128MB NAND chip is on the back:



Source:
http://tiplanet.org/forum/viewtopic.php?t=10394&lang=en

We've been working a lot on TI-Nspire ClickPad prototypes, which revealed to be very interesting models.



On the one hand, a very small hardware difference made their NOR Flash chip writeable:

Thanks to this, we were able to reprogram the boot1 contained in this chip, turning prototypes into production models accepting the OS published by TI.



On the other hand, we noticed that the prototypes ASIC CAS/non-CAS flag wasn't set.

Thanks to this, we cound turn non-CAS TI-Nspire prototypes into TI-Nspire CAS.



It seemed obvious to us, that on production models, there was probably no way to reprogram the ASIC.



But, somebody (who apparently wants to remain anonymous) followed our work very seriously and decided to do the impossible: make NOR chips rewriteable on production TI-Nspire ClickPad!

Even if no flashing has been attempted yet to our knowledge, the NOR chip from this modified production TI-Nspire ClickPad now reports itself as writeable when asked for its properties through an Ndless program.



The Boot1 is the central piece in TI-Nspire models security.
It's the one to check Boot2 and diags checksums and RSA signatures before launching them or not.

Being able to reprogram the Boot1 means being able to run any original, modified or 3rd party diagnostic or boot2 image!

Those who think in the short term certainly have in mind installing a CAS OS on a non-CAS TI-Nspire, thanks to a modified or 3rd-party Boot2.

The others are probably allready thinking to installing Linux, thanks to the U-Boot launcher we dumped on some old TI-Nspire prototypes.



However, such mod if a comprehensive documentation is ever released for it, will only deal with TI-Nspire ClickPad models. TI-Nspire TouchPad have their NOR chip included in the ASIC: there is probably no way to open that chip without breaking everything.





Source:
http://tiplanet.org/forum/viewtopic.php?t=10389&lang=en

More information: (fragmented and incomplete so far)
http://tiplanet.org/forum/archives_voir.php?id=6835
http://www.cemetech.net/scripts/countdown.php?/text/production_clickpad_information.tar.gz&path=archives
http://ourl.ca/16358/305205 (point 3)
https://groups.google.com/forum/?fromgroups=#!topic/tinspire/GQJO45-kXyk (point 3)

Basic TI-Nspire and TI-Nspire CAS share the same hardware. Yet, installing a CAS OS on a basic TI-Nspire is not possible, nor is possible installing a basic OS on a TI-Nspire CAS. Thus, this impossibility is not due to the hardware but to many software protections without which TI wouldn't be serious when selling both models at different prices.
The TI community always wanted to use TI calculators at their best and long looked for using a CAS OS on a basic TI-Nspire.

The first solution was found using Ndless 2.0, a software - whose creator wishes to stay anonymous - decrypting the OS on computer and the little calculator software OSLauncher, developed by Lionel Debroux. Starting from a basic TI-Nspire with OS 1.7, 2.0.1 or 2.1.0 and running Ndless 2.0, it was possible to launch the CAS version of the OS sharing the same number and vice versa.
Note that the OS was only launched - not installed. The mod was temporary, cancelled by any reboot of the calculator.

Since OS 3.0.1, installing an OS 2.1 or older isn't possible anymore - officially. This is due to a minimum installable version number written in the NAND chip in a zone out of the file system. Even though there is an unofficial method to get rid of this problem, this method cannot be used on TI-Nspire CX, as the first OS supported by their new hardware is 3.0.1.

From now on, this is not Lionel but Compu who updates OSlauncher for using it with Ndless 3.1.
We did not write much about it because of many inconveniences of this new version, making it almost impossible to use for most users:

• OS launching fails most of the time (95%)
• you have to put the launching program in the Ndless startup folder to increase the success rate
• OSes 3.1 and 3.2 just freeze when hot-launched - you can only launch OSes 3.0.1 and 3.0.2, which are different from OS 3.1 and mess the ressources id:
  
  
• like the above version, the mod is temporary and cancelled by any reboot
• you have to reinstall Ndless on the hotlaunched OS and cannot do it from the calculator
• and finally, it doesn't work on TI-Nspire CX at all

In a previous news, I revealed that my basic TI-Nspire prototypes were detected as TI-Nspire CAS when using OS 3.1 or later.
Such an information must have gone unnoticed this summer, yet everything finally came clear in my head. A routine detecting the type of hardware has been changed since OS 3.1, making OSLauncher not working anymore. We may suppose that OS 3.1 was modified for this purpose, using the updated routine on the OS startup.
This -apparently useless - information made us find today a way of changing permanently a basic TI-Nspire prototype in a TI-Nspire CAS !
Once again, TI-Planet achieves what was impossible !

What was so different between my TI-Nspire prototypes and the commercial versions ?
Let's have a further look using a software that I made :

TI-Nspire identify themselves with two values:

• an id of the model, stored in the NAND
• a bit stored in the ASIC chip, giving again the type - basic or CAS - of the model

The model ID stored in NAND specifies entirely the TI-Nspire model:

• 0C for TI-Nspire CAS
• 0D for TI-Nspire Lab Station Cradle
• 0E for TI-Nspire
• 0F for TI-Nspire CX CAS
• 10 for TI-Nspire CX
• 11 for TI-Nspire CM CAS
• 12 for TI-Nspire CM

The model type, store in the ASIC chip, respecifies if the TI-Nspire is CAS or basic:

• 0 for a basic TI-Nspire
• 1 for a TI-Nspire CAS

Permanently modding a TI-Nspire model into another one sharing the same hardware can be achieved by changing these two values.
It is possible for the id stored in the NAND, which is publicly documented. We already did something like that for modifying the Boot2, the diagnostic software or even for reinstalling previous versions of the OS.
However, the ASIC is that fat undocumented chip from TI.
Unfortunately, even if we had documentation, we don't even know if this chip is flashable...

Let's get back to my basic TI-Nspire prototypes, identified as TI-Nspire CAS by the computer since OS 3.1.
There must be something looking more like "CAS" on them... Everything looks normal on the About... screen, which shows an id starting with 0E :
But if we use the software previously used...
... we finally find the difference ! The ASIC chip doesn't have the value giving the type of the model, which defaults here to all bits high. On these prototypes, changing only the id in the NAND should change them in CAS models. Let's do it using the tool menu!

The calculators can then be rebooted: they will remove the basic OS and wait for a new CAS OS to be sent.

Here's a video with such a modded TI-Nspire prototype on which you'll see that the mod is permanent as I reboot the calculator:

It took us time, but people who were scammed into buying TI-Nspire prototypes online finally get something far more interesting than production Nspire!



Source with access to fullsize images and to the tool download page:
http://tiplanet.org/forum/viewtopic.php?t=10331&lang=en

Many thanks to Bisam and SilverOne for helping me translating this big news into english.
Many thanks to Bsl, Goplat and Jimbauwens, who made me able to develop this tool after exchanging about dumping/flashing for several years now.

The latest TILP-II 1.17 beta brings a feature we've been waiting for for more than 15 years: TI-80 support!

Did you think TI-80 didn't have a link port?
They have one...

As I discovered some years ago, the TI-80 VSC teacher model has the usual mini-Jack link port.
Normal TI-80 apparently don't have the link port, but the needed circuit is present.
You would just need to solder missing components.

The link port was not meant to transfer programs, but screenshots using the TI-Graph Link 80 software I've never been able to find anywhere. That software might be lost forever.



The TI-80 doesn't use a z80 CPU, but some proprietary 16-bits ASIC from Toshiba @ 980KHz, without any public datasheet.

Christmas 2009, as dumping through software was impossible, I hard-dumped the 64KB ROM chips from TI-80 3.0 and 4.0:
I only got 32KB of useful data from each chip, the same data being repeated in the higher ROM addresses.

Strangely, the dump was incomplete as some system strings were obviously missing. But this, was the beginning of everything!




So where was the missing data?
With only 3 chips, excluding the ROM and the RAM, the only possibility left was the proprietary ASIC, which of course couldn't be hard-dumped.



Randy Crompton studied the partially dumped systems, documented the used assembly language, and developped a dumper he could store in the calculator through some exploits, like the dumping method for the TI-81.

The dumper revealed 16KB missing data, hidden in the ASIC chip. So strangely, the TI-80 is the only graphing TI calculator whose system is stored on 2 different chips.

Summer 2011, Randy acquired a TI-80 VSC, added some TI-80 support in his local copy of TiLP, being able to get the 16KB missing data for the 4.0 system, and to develop an emulator.



So today, at last, two years and a half after my hard-dumping, TiLP replaces the lost TI-Graph Link 80 software thanks to the integration of Randy's code. You can now finally connect your TI-80 to the computer to get screenshots again, and maybe more in the future if a kernel is developped for the TI-80.





Source:
https://tiplanet.org/forum/viewtopic.php?t=10256



Link:
http://lpg.ticalc.org/prj_tilp/beta/setup.exe