AM3352: Not operate normally in 0°C

hello Kanae,

Thank you for the query.

Kanae said:

There is a problem with a customer's board using AM3352BZCZ100 that does not operate normally in  0°C environment.

Can you please explain the behaviour?

Do it stop working or are you restarting the board at 0 degree and it does not work.

What is the boot configuration you are using. Have you tried a different boot if available.

Have you confirmed that the clocks are operating normally.

Regards,

Sreenivasa 

Hi Sreenivasa,

Thank you for your support.
Here are my customer comments to your questions.

Sreenivasa said:
Do it stop working or are you restarting the board at 0 degree and it does not work.
=>The following procedure is used to check.
 (1) Check the operation at room temperature (approx. 23°C) and confirm that it operates normally
      (LCD drawing starts and speech is transmitted through the IC).
 (2) Put the device in a thermostatic bath, set the temperature to 0°C, and expose it to the temperature for about 3 hours.
 (3) Startup in the thermostatic bath → did not start normally (LCD did not start drawing and IC did not speak through the IC)
 (4) Disassemble in the thermostatic bath and confirm that the power supply around the CPU is within the operating range of the CPU
 (5) Confirm that the power supply that operates EN by the CPU is not activated, and determine that the CPU is not activated normally.
 (6) Return to room temperature and confirm that it operates normally
 (7) The problem was reproduced when the power supply was turned on after cooling the CPU only by applying a rapid cooling spray at room temperature.
 (8) Based on the above, it was determined that the CPU was the cause of the non-startup.

The reproduction rate is 100% for boards that stop operation.
The problem is reproduced even if the CPU is cooled at room temperature with a quench spray.

Sreenivasa said:
What is the boot configuration you are using. Have you tried a different boot if available.
=>XIP is the highest priority.
    SYSBOOT pin is set as described below. It is not possible to change another boot.

SYSBOOT[15:14]=10b
SYSBOOT[13:12]=00b
SYSBOOT[11:10]=00b
SYSBOOT[9]=0b
SYSBOOT[8]=1b
SYSBOOT[7:6]=01b
SYSBOOT[5]=1b
SYSBOOT[4:0]=11011b

Sreenivasa said:
Have you confirmed that the clocks are operating normally.
=>I have not measured the actual waveforms.
    The problem was reproduced when the CPU was cooled at room temperature with a rapid cooling spray.
    Based on the above, it is assumed that the clock (external crystal oscillator) is operating normally.

Best Regards,
Kanae

Hello Kanae,

Thank you.

Kanae said:

(8) Based on the above, it is determined that the CPU is not the cause of the problem.

Can you please verify the above statement. 

Also did customer try testing at temperature a little higher than 0 deg Ex 5 Deg C ?

Regards,

Sreenivasa

Hi Sreenivasa,

Thank you for your reply.

I have mistaken to write about (8), and fixed it in my post to
"Based on the above, it was determined that the CPU was the cause of the non-startup."

I suppose that my customer did not try testing at temperature a little higher than 0 deg, but I will check it.

Best Regards,
Kanae

Hello Kanae,

Thank you for updating the reply.

Kanae said:

I suppose that my customer did not try testing at temperature a little higher than 0 deg, but I will check it.

Yes please.

Let me review the inputs and check with the team,

Regards,

Sreenivasa

Hi Sreenivasa,

Thank you for your support.

Here are additional Confirmation Details.

1.Clock.
The faulty board is not on hand due to an in-house X-ray survey, so it is not possible to re-measure the clock.
The external clock is 25MHz of SG210stf.
(SG-210STF | Clock Oscillator (MHz) | Crystal Oscillator; epsondevice.com)

2. Test temperature
We have not tested at a temperature slightly higher than 0°C.
The test is performed in two patterns at 0°C and room temperature (about 23°C).

Best Regards,
Kanae

Hello Kanae,

Could you please help us understand if this is a new design or a design that has been in production.

Can you please check the software version customer uses and if they made any changes to the PLL configuration.

Regards,

Sreenivasa

Hi Sreenivasa,

This is the final prototype before mass production of a new design.
I am still checking the software version the customer is using and if any changes have been made to the PLL configuration.

Additionally, our customer suspects that the AM3352 is a defect, as the same issue was found when the AM3352 was placed on a different board.

Best Regards,
Kanae

Hello Kanae,

Thank you.

Kanae said:

I am still checking the software version the customer is using and if any changes have been made to the PLL configuration.

Noted.

Kanae said:

Additionally, our customer suspects that the AM3352 is a defect, as the same issue was found when the AM3352 was placed on a different board.

We do not have these issues reported by any other customer and so it may be a challenge to conclude.

Regards,

Sreenivasa

Hi Sreenivasa,

Thank you for your support.

Here are reply from the customer.

- Software and version information
→We recognize that this is a question about the build environment. If the answer is different, please point out the difference.

 - TI RTOS SDK v.6.03.00.106
 - Build environment Texas Instruments CCS v10.2.0.00009
 - Compiler Linaro GNU v7.3.1

*The above is Non-OS, or if you are using SDK, I am still checking the version of it. (Kanae comment)

- PLL configuration changes
→The PLL-related register settings are as follows. The following is a list of PLL-related register settings.

Best Regards,
Kanae

Hello Kanae,

Thank you.

Based on the inputs i assume customer made changes to the PLL and you are checking on the SDK or the software platform customer is using.

Regards,

Sreenivasa

Hi Sreenivasa,

I have added the OS information and register values corrected in red text to my previous post.

Best Regards,
Kanae

Hello Kanae,

Thank you.

the expert is out of office this week.

I will follow-up and update when i have some inputs.

Regards,

Sreenivasa

Hi Sreenivasa,

Thank you for your support.
I will inform my customer that the answer will come next week or later.

Kanae said:

Additionally, our customer suspects that the AM3352 is a defect, as the same issue was found when the AM3352 was placed on a different board.

Regarding your comment,  my customer would like to get any feedback on potential factors, etc.,
even if it is difficult to reach the conclusion.

Best Regards,
Kanae

Hello Kanae,

Thank you.

Kanae said:

Regarding your comment,  my customer would like to get any feedback on potential factors, etc.,
even if it is difficult to reach the conclusion.

Could you please elaborate what you mean by potential factors?

Did you mean possible reasons for failure ?

Regards,

Sreenivasa

Hi Sreenivasa,

Yes, It means possible reasons for failure.

Best Regards,
Kanae

Hello Kanae,

Thank you for confirming.

Let me check internally with the expert and update you.

Regards,

Sreenivasa

Hello Kanae,

I am following with the expert internally.

I wanted to check is customer has the required terminations and decaps as per the data sheet recommendations.

Regards,

Sreenivasa

Hi Sreenivasa,

Thank you for your support.

I will check with the customer if he can provide schematics and report back to you via private message.

Best Regards,
Kanae

Hello Kanae,

Thank you for the note.

Regards,

Sreenivasa

Hello Kanae,

I had a  look at the schematics.

The cap values seems to be in line with the data sheet recommendations.

Couple of recommendation

Could customer add a series resistor 22R on the oscillator output, reduce the DDR_reset pulldown to 1.5,,2K and do a quick check.

Also, checking if customer performed some tests to determine the team at which the failures are observed - Ex 5 or 10 deg.

Regards,

Sreenivasa

Hi Sreenivasa,

Thank you for your reply.

Unfortunately, my customer is not able to verify the test at  5 or 10 deg, etc.,
as they have already removed the defective device from the board to analyze it.

Anyway, I will share your view with them.

Best Regards,
Kanae

Hello Kanae,

Thank you.

The reproduction rate is 100% for boards that stop operation.

Any thoughts on the quantity that was tested.

Regards,

Sreenivasa

Hi Sreenivasa,

Thank you for your support.
Customer initially reported that the problem occurred in 1 out of 6 units tested
in a low-temperature environment, and they tested an additional 13 units,
resulting in the problem occurring in 1 out of 19 units.

I will explain to our customer to understand that it is very difficult to answer the question
of identifying the cause of the problem based on the current verification results only.

Best Regards,
Kanae

Hello Kanae,

Thank you.

Agree with you.

Regards,

Sreenivasa