Blue Screen with Blinky Project

Do you use windows 7 or higher?

I have seen that the safety MCU demo accidentally have some issue in the registry to automatically call the nowFlash with a window 7 machine.

But you can launch nowFlash and program it yourself.

To save the time to download the code to flash, please:

1. Connect to the device in CCS

2.Click tools -> On Chip flash

3. in the onchip flash window, click "Necessary sectors Only" and then " Remember my settings"

By doing this, the ccs will not erase and program all the 3Megabyte flash. Save time, work better.

Let me know if you have more questions about the blinky led example.

Regards,

Haixiao

Connecting to the device must be done in debug mode, that is : the debug stop button is active again and
I'm not going to hit it now !
Under tools there is no entry On Chip Flash. Maybe because the target is not connected.
"Connect Target" ends up in :
Warning:
(Error -2062 @ 0x34BC)
Unable to halt device. Reset the device, and retry the operation. If error persists, confirm configuration,
power-cycle the board, and/or try more reliable JTAG settings (e.g. lower TCLK).
(Release 5.0.429.0)

I tried nowFlash. It stops with a manufacturer-ID-error even when I run the blank check only.

Sounds like there are some problem in the JTAG path.

Can you publish the DOS screen shot of the NowFlash? (Please check the "keep window" to keep the information)

Regards,

Haixiao

Thanks for the hint "keep window" as there is no help joint to the help button and no helpfile in the folder.

The JTAG-LED is lit twice for about one second while you can hear a very soft humming from the board.

You see the white field in nowFlash ? This appears when you focus another window and come back to

nowFlash. NowFlash must then be killed by windows, you can't close it normally.

Regards

Detlef

Please change the CPU Family from CORTEX_RXX to CORTEX_RXX_LE, this device (RMxxx) is a little endian device.

If you have already programmed big endian code to it, the MCU might be damanged.

Try to recover it using following method:

1. Press and hold on the RST button. (there are two push buttons on the USB board, one is marked as RST, another one is marked as PORRST).

2. run nowFlash, choose "Erase" only

3. click execute button on the nowFlash win. wait until the follow window shows up (two or three dots '.' after the "connect")

4. Release the RST button.

5. wait...   If error message show up and the error message say sth like 'frequency match', adjust the frequency in the nowFlash win and click the execute button again (do not press the RST or PORRST).

If other message show up, re-try step 1-5 a couple times.

6. If the flash is sucessfully erased, then you should be able to connect to the MCU and program the new code.

Regards,

Haixiao

Haixiao,

thanks for this, I can use the demo again (except the temperature module, which is fixed at 25°. Don't care)

Now we are where we started.
The bluescreen might be cured: I setup a default debug target and could use
the stop button without a crash. But it's not finally tested.

Still CCS is "Unable to halt device" :

Can you publish your  target configuration window? Here it shows ARM9 in the pic you posted. Our device is a Cortex R4F core.

Regards,

Haixiao

Haixiao,

it's like hell!

I've been logging the results of my actions in CCS here in this reply and then again : BlueScreen.

I set both target options (Blinky and Default) to the XDS100v2 and RM48L950. I had to do it twice, for at CCS-Restart,

the default config had been reset to anything.

The first download failed due to a memory failure (I can't explain exactly which one because I don't dare start this ugly thing now!)

The second worked, but single stepping stopped at the end of the loader (why not at "main" ?).

I closed the debug session and started a new one.

CCS tried to erase and load again, so I hit the break button. Bad boy ! CCS looped until BlueScreen.

Detlef, Haixiao,

Is the board that you are working on really "EVMDRX45X" ?

If so,  from what I can tell the processor on the board would be Jacinto not a TMS570.   Is there a TMS570 on this board

as a secondary processor?

If the board only has a Jacinto 1 DDR device, then it would make sense that it is identified as an ARM9 since that part should be

ARM926.  But then I'd also not expect any of the flash programming tools to work or the Hercules safety demos, since it would be a very different

product family.  (the ARM926 class is a predecessor to the "A" series cortex and usually is running an OS like Linux or QNX)..

Best Regards,

Anthony

Anthony,


thanks for joining in.

First of all I'm new to TI, CCS, nowFlash, Hercules. All these commotions originate here.

But: When I purchase a development kit I expect it to be complete and a good introduction to the topic especially for TI-newbies.
How comes that the Hercules demos don't find the automatically installed nowFlash?
Why isn't there any help in nowFlash, even more why should I use nowFlash in a demo at all?
If you followed the discussion, Haixiao worried about me having killed the MCU. With a demo program !!!

Every now and then CCS cannot work correctly with the board and causes a BlueScreen.
That's like 20 years ago when windows crashed once a day.
You're right, I first falsely checked the EVMDMRX45X - board. But with the correct Blinky-target selection
RM48RL950 there had been no change. Only after changing also the default target there had been a little improvement.
Now CCS doesn't crash with the stop button no more, it crashes only while communicating with the board.
Why isn't this thing set correctly for the first steps?

There also must be a path problem. When I use halcogen the files are nicely sorted inc and src. I can't use this
in a project until all *.h - files are copied into the src folder.
As discussed with Haixiao the use of examples doesn't work because the makefiles won't be found. So I first had to look
for makefiles (couldn't know that they are simply called makefile). But even then they are useless because they
include lines like "CG_TOOL_ROOT := C:/ti/ccsv5/tools/compiler/tms470_4.9.5" which is completely different to
my installation. I even don't have ccsv5 but ccsv4. Do I have to hack makefiles for a first glance at the MCU?

I have to decide about the use of this MCU in a new project and with the DK I just wanted to get a bit familiar with Hercules.
In the last days I only pottered away my time and achieved nothing.


Regards
Detlef

Today I had to update JAVA to Version 7.17. This seemed to improve CCS a bit because I had some contacts to the board, though 

most of the time the connections failed. Is it true that every single stepping action takes about 2 seconds to complete?

Now that CCS finished again with a bluescreen I would like to contact the MCU via JTAG (KEIL + ULINK).

Do you have a schematics available so that I can disconnect the associated FTDI-Pins and connect the MCU to a JTAG pinhead?

Regards

Detlef

Hi Anthony,

Best Regards

Hi Detlef,

Understand;  I'm not sure it'll be any easier for you though to try the Keil toolchain given you have the USB stick.

I'd hate to see you start cutting into the USB stick and then wind up with a kit that's both crashing *and* a brick.   I don't know how comfortable you are w. the fine pitch SMDs that you would need to hack into in order to wire a JTAG header for uLINK into the USB stick.     I do know that while I'd be tempted, I probably wouldn't do this myself.

If you need to go w. the Keil MDK for time reasons, I'd suggest pretty strongly getting one of the other kits (CNCD or HDK) that has a JTAG header on it already.

How about the powered hub suggestion?  I think that could be worth a try unless you're sure there's no issue with power.

Best Regards,

Anthony

Hi Anthony,

for your convenience  :-)  I will check the power issue, though the demo software is able to communicate with the board which is

directly connected to a laptop.

I didn't have a look at the schematics yet but as long as the routes don't change layers underneath the FTDI chip

it shouldn't be a big deal. For me the most important thing is to get away from CCS to Keil so I can rely on

a well known tool while exploring something new (even if I have to purchase a board that fits better than this USB stick).

It's like unless you don't know how to handle a scope you shouldn't start searching signals in a TV set.

Best Regards

Detlef