I'm using a SAM9261EK, its in the shipped condition, Basic LCD Project is loaded.
Yesterday i tried to load some other code into the dataflash, I was able to connect using SAM BA 2.9, I loaded some trial example in the IAR compiler, but not sure what address range and all i selected for sending the file.
Now when i am trying to connect using SAM BA i am getting Fatal Error "No Valid Processor ID found".
Tried with disconnecting J21, open/ close J4 ... nothing works.
When i keep J4 open, the same Basic LCD Project still executes.
Tried looking for some recovery options but could not find procedure,
PS: I am using a Serial - USB converter on my laptop, i guess that wont be the issue as I was able to connect once.
Please help as i'm stuck :(

Use JTAG (J-Link/SAM-ICE) for SAM-BA connectivity (XP, SAM-BA prior to 2.10)

Try the USB connectivity.

Check for "ROMBoot >" prompt on serial port (115200 8N1)

Try the BP4 button recovery feature in AT91BootStrap, assuming you implemented it that way. Demonstrated on 9260/9G20/9XE boards.

Use IAR to inject code into SRAM or SDRAM, via JTAG, and run that. Load code to erase NAND/DataFlash. There are probably some 9260 recovery documents that talk about this.

Performance of SAM-BA is particularly problematic under Win7.

Thanks for the quick reply.
Unfortunately I dont have JTAG, so the only option is serial or USB ... when i connect the kit using USB, the PC detects the brd but SAM BA does not show \usb\ARM0 dropdown option.

In the default state ie J4 open & J21 connected 1-2, i get the following serial message
"RomBOOT"
">-- Basic LCD Project 1.4"--

Is there anything that I can do after this via serial ?

Well you really need to be using JTAG if you want to succeed in any development of the SAM9xxx parts. If you set up your IAR projects correctly should should be able to download/debug code directly in SRAM or SDRAM, and benefit from single-stepping and break-pointing afforded by the debugger.

Your problem would appear to be that it is trying to execute semi-valid code in dataflash, For SAM-BA to work it needs to stop at RomBOOT and not execute user code either in the internal flash, or external data flash.

Removing J21 completely and closing J4 and not having any external NOR should permit the board to come up in a way SAM-BA can see it.

You might want to try different versions of SAM-BA, the most recent is 2.10, but 2.7 should be about the same era as the 9261-EK board.

You might want to try different PC's, OS's or serial ports. SAM-BA can be very fussy.

right now J21 is disconnected, so the dataflash / datacard CS is dissabled, also J4 is closed hence there is no valid code .. this should help connecting to SAM BA right ?

is there any recovery procedure for SAM9261, i tried but could not find any material.

From my recollection, it should allow SAM-BA to work, all things being equal. I'd have to dig out my 9261-EK board to experiment.

However, let me reiterate that SAM-BA, the PC application, is not very robust. It tends to fail when you'd most like it to work properly.

This is the document for the 9260-EK, the entire 9xxx range work in a fairly similar fashion, so insight you can draw from it is applicable to your board.
http://atmel.com/dyn/resources/prod_doc ... oc6281.pdf

General advice.
Provide traps doors in your flash based code that permits the device to be erased. Look at how the 9260EK AT91BootStrap uses BP4 to erase the boot blocks in Nand or Data Flash devices.
Provide a means to break, or go into a monitor mode, on the serial port.
Provide a means to enter RomBOOT/SAM-BA mode from you boot strap code.
Test and debug code when it is running in SRAM or SDRAM before burning it into FLASH.