TMS320F28377D: 28377,DSP does not run after power-on

Have you been able to connect to the device through JTAG and Code Composer to see if the device is unlocked/unprogrammed?  From the description it sounds like the flash programming was unsuccessful and the device also got locked.  This can happen if the flashing process is interrupted or if there is a power rail issue during programmation.  Please confirm.

Best,

Matthew

Hi Matthew,

I know the lock issue exist among all GEN2 device, but the issue has been solved in GEN3, am I right?

Angela,

I had missed that point that this was Gen3 device. 

You are correct that accidental lock during normal programmation should not happen with the architecture of the flash and CSM password locations.

Has the customer been able to connect with CCS to see what the flash contents contain on the bad devices?  Is behavior any different with CCS and/or Uniflash?

Are there any error messages from C2Prog when the device fails to program?

Best,

Matthew

Here is the error information:

Hi Matthew,

They can programing successfully through c2prog when pull down the boot. But fail to program through JTAG. Here is the error information.

Has the customer been able to connect with CCS to see what the flash contents contain on the bad devices? 

What should we check the flash contents? 

Hi Matthew,

Here is the further information:

1. The customer irradiated the chip with X-ray and found that GPIO51 52 53 54 has a clear point of failure, corresponding to SPI interface and SCI interface. If there is a test device plus probe in these areas during the production testing phase, and if there are probe points found elsewhere around the area? Customer wants to confirm if the failure is caused externally.

2.They found that the external crystal does not vibrate, and the internal crystal works fine. Customer tested the crystal input foot impedance is abnormal.  What causes the crystal input foot abnormal? There were also several cases of the same fault condition, the crystal input had no apparent impedance difference.
3. The DSP is noticeably hot after power-up and the input power foot impedance to ground is abnormal, which is basically an assumption that the DSP has been punctured. If this breakdown is due to an external power supply anomaly or other EOS issue?

Hi Matthew,

Currently, there are two forms of failure:
1, short circuit of power, the chip is hot, cannot program: customers agree that there is the possibility of overvoltage damage
2, be able to program, the internal crystal works, the external crystal can not work, the crystal pin to ground resistance is around 300 ohms: they want to know what will cause this failure?

Thank you for your help!

Angela,

I'm going to do some more research on my end and will get back to you.  If the over-voltage was high enough and of enough duration there could be damage across multiple pins and not just local to where the overvoltage may have occurred.

Best,

Matthew

Angela,

Wanted to follow up if the customer has any new or additional information on the issue?

Best,
Matthew

Hi Matthew,

Here is the new information:

The customer uses external oscillator through 123pin-X1 and 121pin-X2, and they found that 7pcs board’s oscillator is not working properly on their production project.

The customer checked their PCB, they found that the external oscillator were not damaged(They used this oscillator in other boards, the oscillator can work), but that the impedance of F28377D‘s 123pin was significantly reduced.

The have done the F28377D A-B-A test, the failure followed the F28377D rather than the oscillator.

At present, all the damaged boards have been returned by the market during a period of time, and the phenomenon is that the crystal does not vibrate. They  measured pin 123 pin-X1 has significantly reduced the resistance to GND, from 360 Ω to 1K; this is probably the main cause of no vibration, they want to know why this resistance has become smaller?

Angela,

Based on the data we have thus far, we need to start a return process for the failing devices, so that we can evaluate the devices with our test engineering team.  Please ask the customer to follow the appropriate procedure for the failing devices as outlined here:

https://www.ti.com/support-quality/additional-information/customer-returns.html

Let me know if there are further questions.

Best,

Matthew

Thanks, Matthew. They will follow the procedure. No further question.