TMDXDOCK28035

maybe it helps unplugging all other ftdi devices

like USB-serial

Anders,

I've seen issues starting a debug session when the CLA is include on CCS 4, but I haven't seen the issue on CCS 3.3.

I'm going to move this to the CCS forum so they can comment.

-Lori

As a side note, the workaround for CCS 4 I've seen is the following:

In C/C++ view

1. Do not use "Debug Launch" icon.

2. Go to "Target Configurations" tab.

3. Right click on configuration for F28035. (If one does not exist, you will need to create one by going to Target-->New Target Configuration)

4. Choose "Launch selected configuration."

In (resulting) Debug view

5. Connect to target by Target-->Connect Target

6. Load code by Target-->Load Program (A window will come up - browse to the .out file for the project you are loading.)

At this point, you will be at "main," and you can continue from there.

Anders,

Please provide the information requested here: http://tiexpressdsp.com/index.php/Troubleshooting_CCS#Information_for_Support

Thanks

ki

Customer have removed an RS232 -> USB converter and now the board works with CCS 3.3. The customer would like to use CCS v4.1 but still have errors (see attached files).

The customer would like to use the UART -> USB converter during development, so is there a workaround available that makes the customer able to use the converter and the EVM board?

1. Screenshot of error message Already provided
2. CCS Version: 4.1
3. CCS Installation directory: C:\Program Files\Texas Instruments\ccsv4\
4. Setup Configuration .CCXML file for configuration being used (found in the "<user name>\user\CCSTargetConfigurations" directory):
5. Setup Configuration ccBoard0.dat file (found in \ccsv4\DebugServer\bin\win32\BrdDat)(Please make sure it is saved after failure.) (optional):
   1. If using Spectrum Digital emulators, “sdopts.cfg” file located in Windows\System32 and is generated by sdconfig:
6. Any GEL files used for each CPU (optional, mandatory if customized by user):  none 
7. If Simulator - Simulator board name  none 
8. Chip Part Number and revision: TMS320F28035
9. Target Board Manufacturer: TI
10. Target board part #: F28035XDS100USB
11. Target board revision #:  1.0 
12. Target connector used (ex: 14 pin TI JTAG): We know
13. If customer board, please provide schematic, chips in the scan chain for the JTAG scan chain (up to and including resistors, FETS, etc. that are in the scan chain): We have the schematic
14. Picture of target board (optional): We have the pricture
15. Host Operating System:  win
16. Host Operating System Version:   xp 
17. Host Operating System Service pack:  sp 2 
18. Host Operating System bits (32-bit or 64-bit)  32 
19. Emulator Board Manufacturer: On target board
20. Emulator model # (ex: XDS560, XDS100v1, XDS100v2, 510USB+, etc.): We know
21. Emulator connector used (ex: 14 pin TI JTAG): We know
22. Emulator adapter used (ex: TI 14 pin emulator to 20 pin ARM target, 14 to 14 pin isolation adapter, etc.): We know

Any news so far from the CCS - Team regarding F28035 and CCS4.1.1?

I also cannot connect to a F28035 with CCS4.1.1. I tried to run CCS4.1.1 plus F28035 experimenter kit - without success (see attachment). I tried to use different configurations, all without success. When I try to "launch selected configuration" an error message pops up ant then the whole CCS4.1.1 session collapses!!! This is unaceptable!  Even if we get an debugger error message, the IDE should behave reasonable. 

The very same hardware is running without any problems under CCS3.3.83.19, so why not under CCS4.1.1?

Please note that I have also tested CCS4.1.1 with a F28335 experimenter kit, a F38335 with the peripheral explorer kit and the F28027 Piccolo USB stick - all working under CCS4.1.1 with no problems - ONLY the F28035 configuration shows this misbehave!

Regards

Anders,

Your ccBoard0.dat file is identical to mine. I even used your *.ccxml file. And I am able to connect to the Experimenter's Kit - Piccolo F28035 with my XDS100v1 from both Spectrum Digital and Blackhawk. Which company is your XDS100v1 from?

ki

Anders Lange said:

CCS Version: 4.1

Are you using 4.1.0 or 4.1.1? Select 'Help->About Code Composer Studio' to get the version number.

Frank,

If you are getting a crash, you should be generating crash dump files and jvm error logs. Could you zip them up and attach them here?

http://tiexpressdsp.com/index.php/Troubleshooting_CCS#CCS_Diagnostic_Logs

Thanks

ki

Ki-Soo Lee said:

Anders,

Your ccBoard0.dat file is identical to mine. I even used your *.ccxml file. And I am able to connect to the Experimenter's Kit - Piccolo F28035 with my XDS100v1 from both Spectrum Digital and Blackhawk. Which company is your XDS100v1 from?

ki

Scratch that. Rereading your post, I see you are using the on-board emulation. I am trying to get a setup and see if that will work for me too. From what I understand there may be some jumper setting or something to set that USB port for emulation.

I am seeing the same symptom: Every time I try to Debug using an F28035 Experimenter

Kit with CCS4.1.1, CCS terminates, saying that the JVM has exited. OS is Vista Home (SP2).

At one point CCS offered to send a dump file to TI, but then Eudora couldn't find the file to

attach, and afterwards I wasn't able to find it either.

This customer is using Windows XP and do not have a crash. The CCS 4.1 will not connect to the EVM. When the customer is using CCS 3.3 it works

Version of CCS is 4.1.1.00014

KI-Soo,

see attachment.  OS is Vista 32bit. No problems with CCS3.3.83.19, XDS100 and F28035 at the same PC.

Regards

Hi Ki-Soo

Customer is using XP, you should have received the information by email.

CCS v3.3 is working, but CCS ver 4.1 is not...

Hi Anders,

        This seems really odd. Can you tell me the exact hardware configuration of the USB Docking station? What jumpers are placed where? Also, can you send a screen shot of the customer's device manager window? I'd like to make sure that the XDS100 emulator is connected and recognized by Windows correctly. Thanks!

Charlie

Hello folks,

I have yet been able to reproduce the connectivity and crashing issues with the 28035. I have successfully connected with the onboard XDS100v1 and also with external XDS100v1 emulation. Thank you for the various support files (ccxml, board.dat, dmp files) but we still have not been able to figure out why people are having problems. Without a reproducable test case we cannot properly investigate the issue. Please make sure that the board is properly recognized by Windows if using th onboard emulation as Charlie suggested. Another thing that can help is please send a Debug Server Log:

http://tiexpressdsp.com/index.php/Troubleshooting_CCS#Debug_Server_Logging

Thanks

ki

Hi,

I'm attaching my debug log.

Best regards,

Geof

Ki and Charlie,

I have attached (end of pdf) a screenshot of the vista device manager. It clearly recognizes the XDS100. No jumpers are set at the onboard XDS100. The board is powered by USB.

The same hardware with the F28035 is running with no problems under CCS3.3.83.19.

I have also tested different hardware ( Piccolo 28027 USBstick, Peripheral explorer board 28335) - no problems with CCS4.1.1

AND, I have also replaced  the F28035 controlcard by a F28335 controlcard at the same experimenter kit (with onboard XDS100)  - it works both under 4.1.1 and 3.3! 

Could it be that the parallel CLA causes some issues under the CCS4.1.1 target configuration?

Regards

The customer i am supporting, has the EVM working under CCS 3.3

This tells me that windows is working with the EVM

Here's my dbgView log.

I also succeed if I replace the F28035 controlcard with an F28027 controlcard.

To keep all the information on the same thread, I'm posting this debug server log file on behalf of Anders.

For those using the onboard emulation, do you have an external power supply attached as well? Or, is it the board completely USB powered?

If the board is completely USB powered, would it be possible to connect an external power supply and see if this makes any difference for the connect issues you see?

Andy,

it does not make a difference.  Both, external and USB only powered board lead to the 28035 - CCS4.1.1 crash.  As you can probably guess, I did this test a couple of weeks ago as one of the first steps to find out what's wrong.

I would recommend that we stop speculations about possible hardware issues here. Some of us tested the same hardware under CCS3.3. => successful. We changed the controlcards 28035 against 28335 and 28027 => all successful. Of course I also changed the 28035 card against a 2nd 28035, I also used a 2nd XDS100 Experimenter Kit => CCS4.1.1. crash.

We also inspected the windows device manager, XDS 100 is recognized.

So let us focus on software issues: In one of my earlier posts I mentioned that the 28035 is the only device in this family with the CLA on board. Under CCS3.3 we do use two different CCS-sessions during debug, one for the core of the 28035 and another one for the CLA. Could it be that the CCS4.1.1. target driver has a problem with the two active units inside the 28035?   Is there anybody, who has ever established a succesful parallel debug for the F28035 and the CLA under CCS4.1.1?

Thanks

Frank

Is this related? Symptom looks similar, and it's also Vista.

http://e2e.ti.com/support/development_tools/code_composer_studio/f/81/t/42251.aspx

Petrei seems to have found a missing library dependency, but I don't know enough to follow this trail.

I've done more experimenting...

This problem is related to the combination of CCSv4 AND the on-board XDS100v1 emulation on the DOCK.

• It is not related to 28035 -- I can show the failure with 2808 control card plugged in.  So it is not the CLA.
• It is not related to the USB power versus external supply power.  I have tried both. 
• It is not related to CCSv4 alone -- I can get it to work with an external XDS100v1.
• It is not related to XDS100v1 alone -- I can get the external XDS to work.
• It was not present in CCSv3 -- I can get CCSv3 to work fine.
• It is not related to the specific versions of CCSv4 -- I've tried 4.0.2, 4.1.0, 4.1.1 so far.
• And it fails on any of the TI laptops -- I have now tried it on several.

Can we get someone in CCS-land to try this setup (the 28035 Experimenters Kit)?  The failure can be exhibited simply by doing:

• Create XDS100v1 Configuration with F28035 selected
• Launch Configuration
• Connect Target  (this will fail).

Thanks very much.

Best regards,
Geof 

Hi.

Attached is a second log file from the customer

Hi,

That fixed the problem!  

So it looks like we are shipping the TMXDOCK's with improperly programmed FTDI chips.  

Thanks very much for your help with this problem.

Best regards,
Geof

Geof, Anders:  It does NOT work in my case.

I reprogrammed the FTDI chip of the onboard XDS100 with Steven's XML-File and the FT_PROG utility => results are the same. F28035 plus XDS100 Docking station does NOT work under CCS4.1.1 - it was and is still working under CCS3.3. 

I was somehow expecting this, because where is the logical reason in replacing the FTDI - binary code, if the hardware was already succesful working under a CCS3.3 environment?

There was another thread in the C2000 forum, which discussed improperly programmed FTDI chips but this is and was definitely not the case for the board(s) which are at my desk.  

Regards, Frank

PS: Geof, I don't think that the TMXDOCK's are shipped with improperly programmed FTDI chips, they are working under CCS3.3 but not under CCS4.1.1

Hi,

My board worked with CCS 3.3 before the firmware upgrade.  So I didn't mean that they were unprogrammed before.  It appears to me that the boards are shipped (sometimes) with firmware that only supports v3.3.  Once I did the upgrade, my board worked with CCS v4.

I also admit that I didn't go back and recheck whether or not the "reprogrammed" board still works with CCSv3.3.

But I can say that I had the case where hardware worked with 3.3, and didn't work with 4.x until I reprogrammed the board.

It's a confusing issue.

Best regards,

Geof

Frank,

I agree there is more than one issue going on.  There was a report a while back that the "Debug Launch" was giving issue.  Could you try this sequence just to rule it out?

1. Do not use "Debug Launch" icon.

2. Go to "Target Configurations" tab.

3. Right click on configuration for F28035. (If one does not exist, you will need to create one by going to Target-->New Target Configuration)

4. Choose "Launch selected configuration."

In (resulting) Debug view

5. Connect to target by Target-->Connect Target

6. Load code by Target-->Load Program (A window will come up - browse to the .out file for the project you are loading.)

-Lori

I just got feedback from my customer, and everything is working after reprogramming the firmware.

Lori,

I can rule out the sequence above, I tried that one as well. CCS4.1.1 crashes  after 4. "Launch selected configuration".

I still suspect the CLA (pure speculation) driver under ccs4.1.1.

Those of you who have a running 28035 under CCS4.1.1:  Do you also debug the CLA? Do you open a 2nd debug session in parallel to the debug session of the main processor? How does it work?   There is a nice description from Lori how to debug the F28035 and the CLA  in parallel under CCS3.3.. Is there a similar video/pdf for CCS 4.1.1?

Regards

Frank

Frank Bormann said:

I can rule out the sequence above, I tried that one as well. CCS4.1.1 crashes  after 4. "Launch selected configuration".

Ok thanks for trying it.  It was worth a shot.

Frank Bormann said:

Those of you who have a running 28035 under CCS4.1.1:  Do you also debug the CLA? Do you open a 2nd debug session in parallel to the debug session of the main processor? How does it work?   There is a nice description from Lori how to debug the F28035 and the CLA  in parallel under CCS3.3.. Is there a similar video/pdf for CCS 4.1.1?

Yes, I am able to start a debug session using a CLA configuration, connect to the main CPU and then connect to the CLA and debug.  I don't have a video showing this yet.  I spent some time yesterday trying to reproduce your error without any luck :(  I share your frustration. 

Just to make sure - are you connecting to the main CPU first, before connecting to the CLA?  It won't work the other way around as the CLA is a slave to the 28x.

-Lori

Frank -

I wanted to follow-up on this.  I don't have any new sugguestions but I wanted to see if this was still an issue?

-Lori

Lori,

the situation on my laptop is still the same, CCS411 crash with F28035 and XDS100. XDS100 is fine on CCS3.3.83.19 (all type of controlcards).   I cannot spent time this week on this..

Frank 

I have the same problem with the HVMotorCtrl+PFC Kit (TMS320F28035). CCS 3.3 is able to connect with the processor via xds100-f28035, but CCS 4.12 is not. CCS 4.12 gives "Error initializing emulator" popup.

Attached is my dbgview log file.

Thanks Michael. Could you also provide the information requested here?

http://processors.wiki.ti.com/index.php/Troubleshooting_CCS#Information_for_Support

This will help our engineers with their investigations

Thanks

ki

Ki,

I will try my best...

Problem:

I have been trying to make connection between CCS and target (HVMotorCtrl+PFC Kit, with F28035 controlCARD Release 1.0) via the built-in xds100v1USB connection:

• With the latest CCS 3.3MCU edition, I am able to connect to the target
• With CCS 4.1.2.00027, I am not able to connect to the target using the "Debug Active Project", or "Launch TI Debugger" pulldown menu. CCS shows the "Error initializing emulator" popup, as if the controller card is not there or there is not power. Eclipse does not crash, and I can build the project, re-start the debug server, and same error will occur

The operating environment is 32-bit Windows Vista with SP2.

CCS 4.1.2.00027 was installed on D:\Program Files\Texas Instruments\ccsv4.

The debug server log was send in the last post.

Let me know what else you need...

Regards,

Michael