Trouble Connecting to F2809 After Flashing

Hi Randy,

As we already established your clock is fine. I'd also check supply voltage especially on dedicated FLASH supply voltage pin.

I assume that the code is correct, and you've tried even with simple examples like basic blinking led code, and you arrive at the same result?

I'd just like to know if you can connect to programmed DSP after you power cycle it or never again?

Regards, Mitja

Hi Mitja,

We did look at the 3.3v and it seems fine. The problem with checking other code is that we 

only have 4 boards - I've already bricked two of them with this problem, and the hardware

guys need the other two. So without being able to reconnect (and therefore reflash),

program changes and any other debugging is out of the question.

I can try looking at the FLASH voltage pin. 

I've tried power cycling, unconnecting the USB, restarting the PC, reseating the connectors.

No matter any of these, I still cannot connect. 

Thanks for your help.

--Randy

PS: Also since I first posted yesterday, we have verified that the same debugger hardware (jtag emulator, cables, etc)

work fine with another board, and that this board still has problems with another debugger. 

Robert Chyla, of Signum, also gave me a lot of help yesterday. We verified using the "emudiag" utility signum

has on their site that there is a problem at the scan chain level, i.e., the diagnostic could not even get a pattern

cycling correctly from TDI through to TDO. 

I am really quite puzzled by this problem. 

--Randy

PSS: Pardon me if I wasn't clear, but there are other versions of our program that run fine on other boards. It just seems that I've

come up with a unique build that's hosing something up at the emulator level or somesuch. That's very possible because I

had peeled away a lot of the functionality from our normal build just to do a simple test. But I would hope that it would be 

impossible to generate a build (short of locking up the system via the CSM) that prevented the emulator from connecting!

--Randy

Conjecture: would leaving the system in the EALLOW state cause this sort of problem?

--Randy

Randy Yates said:

Conjecture: would leaving the system in the EALLOW state cause this sort of problem?

It leaves device registers open, but unless you have a pointer running wild you should be fine.

In any case it is hard to answer this question as not all the registers are documented by TI. Especially the ones dealing with JTAG emulator connection.

By my experience when the device is locked one can still connect to the device through JTAG, but the device will return garbage when reading from secure areas.

You can still read correct data from unsecured areas (M0, M1 being the case).

Have you tried powering up the device with XRS pin held down and with the emulator connected and connecting the CCS to the device?

In any case if you do find the cause of your problems please report (as much details as one can these days with NDAs at every corner).

Regards, Mitja

This obviously helped recover piccolo device:

http://e2e.ti.com/support/microcontrollers/tms320c2000_32-bit_real-time_mcus/f/171/p/21170/81806.aspx#81806

Regards, Mitja

Hello Randy,

In addition to the recommendation provided by Mitja above, another recommendation is to configure the device to "Boot to RAM".  This will help to understand if the software has left the device in a state which disables any further connection by JTAG.

If you are able to connect to the device in this mode, proceed to load a simple code which blinks an LED.  

Regards,

Devin

OK, I'm back to a working system. To be brutally honest, I'm not

positive what happened, but my theory is that there were two

independent issues, neither of which had anything to do with the 2809

software, but whose confluence along with other factors made it look

so:

1. One of the two boards I was trying to use presumably has a H/W

problem. Connecting the board to a known-good emulator still resulted

in a bad, unconnectable board. Indeed the boards require a

modification involving the emulator signals.

2. I either had a driver conflict / configuration problem in my

installation of CCS or somehow had gotten my installation corrupted. I

reinstalled CCS 3.3 today, along with the updates. At the same time, I

noticed in the Windows XP hardware device manager that a TI XDS driver

was installed as well as the signum device driver. I removed the XDS

driver. After the upgrade and removal, things started working with the

known-good board.

Thanks Mitja and Devin for your help and support. Sorry for the "false

alarm."

--Randy