F28M35H52C: Unable to Program C28 side - Error 1015

Hi, 

Your query has been forwarded to an expert, pls expect a response by tomorrow.

Best Regards

Siddharth

Hi Eric,

Thank you for your question. I'll need some more time to look over the hardware schematics, but I'll get back to you tomorrow. 

Thanks,

Charles

Thank you, I look forward to your assistance with this. Let me know if you need anything else.

Hi Eric,

Some software/hardware questions I have for you:

1) When connecting the XDS100v2 debugger and connecting to CCS, are you manually trying to connect the target configuration (.ccxml) file for the device, to see if it is in the locked state (without clicking the debug button)? I assume you are using the JP1 & JP2 wires to do this.

2) Have you made sure that the TRSTn pin is not in a floating state? It could be affecting the CPU.  The schematic is a bit blurry, what side is this pin on?

3) Would programming the C28 side first make a difference?

Thanks,

Charles

Charles

1) Sorry I'm not very experienced with CCS. I would assume I am connecting manually. I am not connecting via debug. While in CCS I have a tab with the *.ccxml file name open and I can click through the tabs to change the TCLK and some other settings. This is the location I do the 'Test Connection' test. I then save this file and open it in UniFlash. UniFlash is what I use to program the chip. JP1 & JP2 are connected during this processes.

2) I have a 2.2K pull down (R24) on the TRSTn pin. This is between the output of the digital isolator and the TI chip. I also have a pulldown resistor (R21) between the FTDI chip and the digital isolator that has normally been DNP. I have tried a pull down on it as well without success.

3) I have erased the M3 side and attempted to program the C28 first. That too was unsuccessful.

Hi Eric,

I see that the C28x flashing sequence did not even start since you are having trouble connecting to it.  Hence, this is not a flash programming issue.

Assuming that C28x is not locked by security:

1) When trying to connect to the C28x, what is the state of the M3?  Is it in halt state?  Or is it executing?  If executing, try halting it.

2) Try programming the C28x after keeping the device boot pins in boot from serial peripheral mode and see if that helps.

3) Is device powered via USB port of the laptop?  If yes, can you instead try powering the board from an external supply?

4) Do you see any toggle on the XRSn pin when C28x fails to connect?

Thanks and regards,

Vamsi

Vamsi, thank you for your assistance

1) I have tried with the M3 executing and while erased. Same result

2) Same error when booting in Boot Mode 2 (PG7 tied high)

3) The CPU is powered (+3V3) by an external power supply. The Debug3V3 is powered by the laptop. It just powers the FTDI chip and the digital isolators. I have put a scope on it and haven't noticed any power dips while trying to program.

4) With the M3 erased and the ARS & XSR separated and floating I appear to be getting a WDR pulse on the XRS. It is @ 25uSec every 1mSec. I see random pulses on the ARS. They are 25uSec as well but happen in an inconsistent manner. With the M3 executing I will see some pulses of the XRS but I believe it is the code preforming a reset since the C28 is not running. I also get an LED flash at the same time that is an indication of a software reset. I do not see any XRS activity corresponding to an attempt to connect.

Regards, Eric

Hi Eric,

1) and 4): I mean to try when M3 is halted - not when executing an erased flash (causes ITRAP) or application. 

I will assign this to a board expert to see if they have any further suggestion.  

Thanks and regards,
Vamsi

Vamsi, I am not familiar with how to do that. What is the procedure to Halt the M3? I will give it a try. I've only had to just program these. I haven't been involved with any development or debug.

Eric,

Would it be possible to ZIP the images at full resolution and attach them to the post vs inline?  I can't zoom in effectively and having trouble reading everything.

Just to clarify, did the previous HW with the 4 channel isolator chip/XDS100 work OK?  I think the answer is yes, but wanted to make sure I understand that. 

To Vamsi's last point he was asking to connect to the device/M3 core  via Code Composer (via the XDS100 connection), halt the M3 and then see if the XDS100 can connect to the C28x core.

This video should explain how to do that(it is using CCSv5, but that shouldn't matter), in case you have only been using UniFlash to interact with the device. https://training.ti.com/concerto-f28m3x-microcontroller-training-part-9-code-composer-studio-emulation

Best,

Matthew

Here is a pdf of the schematic parts. Hopefully this is better.

The previous chip did work without issue.

Thank you for the instruction on connecting to the core. I will try it shortly and report the results.

TI Schematic.pdf

I was able to start the debug session and connect to the M3. I suspended it (assume the same as HALT). This would pause the operation of the chip as the blinking LEDs would stop. I could load a program to it. All seems good.

I tried to connect to the C28 and get the same -1015 error with or without the M3 suspended.

Eric,

Can you probe EMU0/EMU1 at the local C2000 pin and note the value when you are trying to connect, or even just when the board is powered with no JTAG connected?

If these are treated the same even with different isolators I'm not sure why there would be a difference, but I know that standard ARM does not use these pins and C28x does(sampled on the rising edge of TRSTn).  Maybe more a hunch/clue why ARM is OK and C28x is not.

We want to see both as high/VDDIO level when trying to connect(for normal operation). 

Best,

Matthew

Found that the previously used ISO7241 digital isolators had internal 1Meg pull up resistors on the inputs. The ISO7341 used on this build (could not find any ISO7241 instock) do not have internal pull ups on the inputs therefore the outputs 2Y (EMU0) and 3Y (EMU1) of the SN74HC03 were never able to go to a logic 1. Going to replace the SN74HC03 with a SN74HC00 so the output are push pull instead of open drain.

Eric,

Glad we could help, I saw the NAND chip, but I didn't catch that it was OC config with no pulls. 

Let us know if you have anymore questions.

Best,

Matthew