CC2640: - During firmware build chip becomes totally unresponsive

Jacob Hinson

Part Number: CC2640

The CC2640 sometimes is no longer responding.

Video is here: https://www.youtube.com/watch?v=-8415I2EOpk

I can not explain it, because I do not touch the PCB, it's sitting next to me and then it stops responding.

I can flash the chip 2-50 times, and then it stops responding. I can not clean it, it can not be reset, it's broken. Then I change it to a spare chip and it works

We have had this feedback""

[GN] - Actually we saw a similar problem on one board here, we have no idea why it happens but some Google searching shows other people have had similar problems, it might be down to the fact that the firmware is only partly programmed. Aaron found on one site that they suggested if you remove the 24MHz crystal; so the partly loaded firmware cannot start; then you might be able to mass erase the device and then refit the 24MHz crystal but I don't think we tried that here."

And have tried it but for some reason, it keeps happening and we cannot just flash the chip and use the boards

its making firmware development impossible

Any ideas why this may be happening?

Thanks

Jacob

over 7 years ago

0 YiKai Chen over 7 years ago

Guru 735695 points

Do you use custom PCB? This is usually caused by bad JTAG connection so I suggest you to check connection first.

0 Fredrik K over 7 years ago in reply to YiKai Chen

TI__Guru** 106705 points

Hi Jacob,

Can you try using Flash Programmer 2? Does it properly identify the chip? Can you do a "Forced Mass Erase"?

What is the state of the chip in the failing case? What are the voltages on VDDR and the DCOUPL pin? What is the current consumption?

Can you share a photo of your board connected to the debugger?

Cheers,
Fredrik

0 Jacob Hinson over 7 years ago in reply to Fredrik K

Prodigy 190 points

1. Power supply 3.25V

VDDR 1.66V

DCOUPL 1.26V

Consumption ~9 mA (including LDO)

2. How it happens: in CCS I start debugging, the chip is successfully flashed, I press Start. But debugging does not stop at my breakpoints. I stop debugging, I try to start again and get an error.

3. Flash Programmer 2 sees the chip, but can neither erase\read it nor make Mass Erase - see video

www.youtube.com/watch

0 Jacob Hinson over 7 years ago in reply to YiKai Chen

Prodigy 190 points

See below

0 YiKai Chen over 7 years ago in reply to Jacob Hinson

Guru 735695 points

Nothing shows in your reply.

0 Jacob Hinson over 7 years ago in reply to YiKai Chen

Prodigy 190 points

Yea its custom - here is an image

0 Jacob Hinson over 7 years ago in reply to Fredrik K

Prodigy 190 points

image is below on the thread

0 YiKai Chen over 7 years ago in reply to Jacob Hinson

Guru 735695 points

I suppose you also need to connect RESET_N between XDS110 and CC2640.

0 Jacob Hinson over 7 years ago in reply to YiKai Chen

Prodigy 190 points

I connected connect RESET_N between XDS110 and CC2640. There is no separate RESET pin on the board, I soldered the wire to R1)
The result is the same

Thanks YK Chen

0 YiKai Chen over 7 years ago in reply to Jacob Hinson

Guru 735695 points

I also suggest you to give power from external power supply instead of XDS110.

0 Jacob Hinson over 7 years ago in reply to YiKai Chen

Prodigy 190 points

Also tried - powered by batteries :)

0 YiKai Chen over 7 years ago in reply to Jacob Hinson

Guru 735695 points

If you use Flash Programmer 2 to download hex file, does it work?

0 Jacob Hinson over 7 years ago in reply to YiKai Chen

Prodigy 190 points

I will try

0 Jacob Hinson over 7 years ago in reply to Jacob Hinson

Prodigy 190 points

no, its doesn't work - see my answer #3 and video here:

3. Flash Programmer 2 sees the chip, but can neither erase\read it nor make Mass Erase - see video www.youtube.com/watch

0 YiKai Chen over 7 years ago in reply to Jacob Hinson

Guru 735695 points

I can see your video. The link doesn’t work.

0 Jacob Hinson over 7 years ago in reply to YiKai Chen

Prodigy 190 points

link is here

www.youtube.com/watch

and here

https://youtu.be/-8415I2EOpk

Maybe you need a VPN if you are in china?

Jacob

0 YiKai Chen over 7 years ago in reply to Jacob Hinson

Guru 735695 points

Try to refer to processors.wiki.ti.com/.../CC26xx_HW_Troubleshooting to do some HW debug.

0 Jacob Hinson over 7 years ago in reply to YiKai Chen

Prodigy 190 points

We are testing the run CMD_PING as a final option, everything else is correct including the voltages - please let me know if you have any other suggestions?

I can see other threds about this with no real answer and im scared about it!

Thanks

Jacob

0 Jacob Hinson over 7 years ago in reply to YiKai Chen

Prodigy 190 points

Yes, we tried CMD_PING there is no answer from the chip, we even connected it directly with JTAG and there is still no answer.

Do you have any other suggestions?

0 Glenn Needham over 7 years ago in reply to YiKai Chen

Prodigy 50 points

Hi, Just to add more information to that my colleague has already given..........

We are using the compact JTAG so only 2 connections are required to the debugger. Reset is pulled up with a resistor. We are using an external power supply and an onboard 3V3 regulator.

We can load firmware and debug a few times then on one random occasion the chip become unresponsive.

0 Markel Robregado over 7 years ago in reply to Jacob Hinson

Guru 86765 points

Hi Jacob,

Try Yikai's suggestion below. After you have done the connections below, try program it using SmartRF Flash Programmer 2.

"I suppose you also need to connect RESET_N between XDS110 and CC2640."
"I also suggest you to give power from external power supply instead of XDS110."

-kel

0 Fredrik K over 7 years ago in reply to Glenn Needham

TI__Guru** 106705 points

Hi Glenn and Jacob,

As already mentioned in the thread, reset must also be connected to the debugger due to how the device boots up in debug and programming mode.

Have you done any modifications to CCFG in your SW project? Does the problem also happen if you use an unmodified project? Can you control the device from Smart RF Studio?

Can you share your HW design files?

Cheers,
Fredrik

0 Jacob Hinson over 7 years ago in reply to Markel Robregado

Prodigy 190 points

Both have been done

0 YiKai Chen over 7 years ago in reply to Jacob Hinson

Guru 735695 points

Do you have another custom board to test the same thing?

0 Jacob Hinson over 7 years ago in reply to YiKai Chen

Prodigy 190 points

Yes we have 3 and all are giving the same issue

0 Fredrik K over 7 years ago in reply to Fredrik K

TI__Guru** 106705 points

Can you try one more thing: power up the board, do a manual reset by pressing the reset button on the LaunchPad, try connecting with Flash Programmer 2.

0 Jacob Hinson over 7 years ago in reply to Fredrik K

Prodigy 190 points

1. I did not make any modifications to CCFG

2. Yes, it happened when I flashed the unmodified project from TI (when we were looking for the reason why it's impossible to connect with the chip on the Bluetooth, I used the project from TI simple_peripheral)

3. I used the RESET line only to try to repair the chip. Debugging did without it - do you think this issue can happen because debugging was done without the RESET line?

HW - i will message you direct

0 Jacob Hinson over 7 years ago in reply to Fredrik K

Prodigy 190 points

We tried it after the chips had died - the chip did not work again using this method

0 YiKai Chen over 7 years ago in reply to Jacob Hinson

Guru 735695 points

So, you cannot do mass erase from Flash Programmer 2 to all of your 3 custom board? By the way, if you have 4th board without download anything, can you download FW successfully into it with XDS110 ?

0 Jacob Hinson over 7 years ago in reply to YiKai Chen

Prodigy 190 points

We were using this acftion before but had no RESET set-up, so we think it may be (as per fredricks message above) that the issue was caused by not having RESET attached while debugging?

0 YiKai Chen over 7 years ago in reply to Jacob Hinson

Guru 735695 points

I couldn’t understand your descriptions and question. Can you elaborate?

0 Fredrik K over 7 years ago in reply to Jacob Hinson

TI__Guru** 106705 points

Not having the reset signal could have put you in the state you describe, BUT it should be resolved simply by connecting the signal. It would not permanently lock the device.

Your symptoms really look like the JTAG interface and the ability to do mass erase has been locked in CCFG.

0 Glenn Needham over 7 years ago in reply to Fredrik K

Prodigy 50 points

Hi, just to update on this.......
For reference this applies to the cJTAG (2 wire) debug interface.

We had great difficulty getting this to work reliably, we could program the CC2640 several times and debug then on one occasion the chip would get put into an unresponsive state. Possibly due to the config being programmed incorrectly and disabling JTAG access (although this has not been confirmed). Our setup was a custom target board connected by wires to the Launchpad JTAG outputs using the Launchpad as a debugger/programmer. However, we had seen the same behaviour on normal debuggers.

We have now made some changes and things look much more reliable, the changes we made were:
1. Connect the reset signal of the target to the JTAG output on the launchpad, this shouldn't be needed as 2 wire debug interface should not use or need the reset
2. Make the wires between the target and the launchpad much shorter - basically they want to be as short as you can possibly make them
3. In the debugger settings in the IDE change the JTAG clock rate from the default of 2.5MHz to 1Mhz

With all of these changes we now seem to be able to debug and program reliably, time will tell if this continues to be the case. My guess is that changing the JTAG clock frequency alone might have been sufficient but we haven't proven that yet.

Also note.... if the CC2640 is already in this 'locked' state then these changes WILL NOT bring it back to life, they only help prevent it happening in the first place

Hope this helps someone!!!

0 Fredrik K over 7 years ago in reply to Glenn Needham

TI__Guru** 106705 points

Hi Glenn,

Good to hear that you were able to resolve the problem. Do let us know should it reappear.

You are right that normally reset is not required, but for the CC26xx devices the emulator uses the signal to trigger a halt in boot. Without reset, the device may potentially start executing code which disables the JTAG before the JTAG can actually connect. Should this have been the case with your problem, connecting the reset signal would have resolved it.

Regards,
Fredrik

Bluetooth®︎

Bluetooth forum

CC2640: - During firmware build chip becomes totally unresponsive