MSP430FR2355: Get the same ADC value for a certain interval.

Hello Ito,

No, this is not normal behavior.  Not sure why every 64 cycles. Can customer share portion of their code that shows the ADC setup and measurement (including any ISR)?

What is connected to the pin? Does it happen with a short?

Hi Dennis

We are unable to share parts of the code (including ISRs) that show ADC setup and measurements,
but we will send you the customer's report.

Can you answer some of the questions in the report?

Best Regards,

Ito

Hi Keith,

Thank you for your reply.

I will tell my clients what you taught me.

Please confirm the report I sent to Dennis.

Best Regards,

Ito

Hi Dennis

Customer requires immediate attention.
Please reply as soon as possible.

Best Regsrds, 

Ito

Hi Ito,

Yes, I understand.

Since the customer is not able to share their code, can they provide a general description of how they are making these measurements?

Is the same voltage applied to A1 and A3 at same time?

What is the customer using for the ADC reference?

Hi Dennis,

Thank you for your reply.

We are currently seeking answers from our clients to the questions we have received.

According to the customer, the same error seems to occur when connected to either A1 or A3.

Please answer the following questions as soon as possible.

1、If there is an abnormality in one part of the SAR comparison process, can it be found in the outgoing inspection?

2、Is there any other ADC products other than microcontrollers in which the above phenomenon has occurred?

Best Regards,

Ito

Hi Dennis

Thank you for your reply.

A1: 0 to 2.5V
A3: 0 to 2.5V
The ADC reference is
Positive: External reference voltage 2.5V
Negative: External reference voltage 0V (Vss)

Inputs to A1 and A3 are made separately, so they are never made at the same time.

Best Regards,

Ito

Hi Ito,

I do apologize for the delay in responding.

I have not been able to reproduce this behavior.

Has the customer seen any more devices with this issue?

Question for the customer - when performing this test, do they have the debugger connected and running in debug mode,  stopping at a break point in the ADC ISR after each conversion then read the ADCMEM0 register?  If not, can the customer describe in more detail how they change input voltage?

Here is what I think you are saying:

A ramp is applied to the ADC input.

For results that are expected to have bits 3 or 4 of the conversion result be a '1', the result is in error. Only one such grouping is shown but that seems to be what you are saying

This would indicate a defect in the ADC. (TI seems to mostly use charge redistribution in their SAR ADCs.) This is the sort of defect that could be missed during testing unless a full ramp is applied.

Hi David,

From the customer's description on how they are testing the ADC, it sounds like they are running in debug mode with a breakpoint in the ADC ISR, where they can examine the results for each step in the input voltages.  This shouldn't ordinarily impact the ADC conversion results.

Ito, can the customer apply a slow ramping voltage and sample the ADC quickly and output data to UART where it can be sent to PC so it can be plotted?

Also, can you confirm the customer sees this behavior in only one device, correct?

Hi Dennis,

The plot file was sent to a private message.
This behavior occurs in the microcontroller hardware.
This behavior occurred on only one out of several hundred in production.

What causes this behavior?
Is it caused by static electricity, overvoltage, heat, dropping, vibration, etc.?

Best Regards,

Ito

Hi Ito,

Have the customer try setting  the ADCSHT0 = 1 in ADCCTL0 register.  This will increase sample/hold time and will help eliminate this as possible cause.

Hi Dennis,

ADCSHT0 = 1 in ADCCTL0 register, but the same error occurs.
This error is not caused by software.
It is accidental, occurring in one out of several hundred units.
It is caused by hardware.
And what is the cause of this hardware error?
If you do not know the hardware, could you please change the assignment?

Please respond as soon as possible.

Best Regards,

Ito

Hi Ito,

I have requested help from our Shanghai team on this topic.

Hi Ito

>> This behavior occurred on only one out of several hundred in production.

May I confirm these several hundred MCU is the same lot? Are they the same PCB?

which soldering method customer used? May I know the soldering temperature curve?

Can customer consider do the replace test to isolate if this issue is related with THIS DEVICE (only one out of several hundred in production)?

Thanks

Hi Li

May I confirm these several hundred MCU is the same lot? Are they the same PCB?
→Same lot, PCB.

Which soldering method customer used? May I know the soldering temperature curve?
→The soldering method is reflow. Temperature curve have been sent in a private message.

Can customer consider do the replace test to isolate if this issue is related with THIS DEVICE (only one out of several hundred in production)?
→Since there is only one substrate available for experimentation, it is not possible to replace.

Best Reagrds,

Ito

Hi Shanghai team 

Please let me know current progress.

Is there any other data you want?

Best Regards,

Ito

Hi Shanghai team 

Please let me know current progress.

Is there any other data you want?

Best Regards,

Ito

Hi Ito

I am in hospital from last week.

May I confirm these several hundred MCU is the same lot? Are they the same PCB?
→Same lot, PCB.

--> Yes: I understood this error occured in one out of several hundred units.

Which soldering method customer used? May I know the soldering temperature curve?
→The soldering method is reflow. Temperature curve have been sent in a private message.

--> I have recieved the temperature curve Temperature.pdf. I checked this curve according to QFN and SON PCB Attachment. I think the temperature curve is OK on the reflow soldering

Can customer consider do the replace test to isolate if this issue is related with THIS DEVICE (only one out of several hundred in production)?
→Since there is only one substrate available for experimentation, it is not possible to replace.

--> Because customers have made hundreds of products. Can the customer remove the only one FR2355 has ADC issue from the board and re-solder it to a board that has been proven to be working properly to see if it is a problem with this board

Previous discussion with Dennis: ADCSHT0 = 1 in ADCCTL0 register, but the same error occurs. This error is not caused by software. It is accidental, occurring in one out of several hundred units.

--> I still think extending sample-and-hold time is needed because of the external resistance and capacitance factors.

--> before the possible hardware test, please increase the ADCSHTx.ADCCTL0 to 0011b = 32 ADCCLK cycles and add some "NOP" between ADC conversion to check if the issue can be improved (by software).

Previous discussion with Dennis: The plot file was sent to a private message

--> I can see you have sent some file to Dennis by private message. Could you please send to me for the reference?

Thanks

Hi Li,

Please take great care of yourself.

>>Can the customer remove the only one FR2355 has ADC issue from the board and re-solder it to a board that has been proven to be working properly to see if it is a problem with this board

→Customer is concerned that removing the only error MCU may change the state of the MCU. They do not want to touch the error board.
They do not want to do SWAP testing for these reasons.

>> before the possible hardware test, please increase the ADCSHTx.ADCCTL0 to 0011b = 32 ADCCLK cycles and add some "NOP" between ADC conversion to check if the issue can be improved (by software).

→The customer increased the sample & hold time, but all confirmed the same error.

>> I can see you have sent some file to Dennis by private message. Could you please send to me for the reference?

→I'm  sorry. I sent the plot to your private message.

Now customers are concerned that these errors may be caused by their own manufacturing processes.
What precautions can be taken to prevent errors, such as static electricity, overvoltage, heat, dropping, vibration, etc.?
We would like to know if the error occurred during the customer's manufacturing process or during the manufacturing process on TI's side.

I understand this is a very difficult issue.

Best Regards,

Ito

Hi Ito

>> Now customers are concerned that these errors may be caused by their own manufacturing processes.

So far, there is only one MCU or board with ADC issue. it is hard to isolate the device error or board error without SWAP test

>> What precautions can be taken to prevent errors, such as static electricity, overvoltage, heat, dropping, vibration, etc.?

On ESD: please refer to device data sheet and MSP430 System-Level ESD Considerations (Rev. B)

On Thermal: Please refer to device data sheet and Semiconductor and IC Package Thermal Metrics.

>> We would like to know if the error occurred during the customer's manufacturing process or during the manufacturing process on TI's side.

Please discuss the field quality team if field quality can help to check customer's manufacturing process.

if not, please discuss with customer to continue do the mass production. if another device which the similar ADC issue can be found. it should be helpful to do the SWAP test to isolate the issue because there will be more than one error MCUs at that time

Thanks!

Hi Li,

Thank you for providing the documentation.
I would like to hear from your field quality team, please provide me with their contact information.

Best Regards,

Ito

Hi Ito

>> I would like to hear from your field quality team

Could you please contact with field team to check the field quality team?

Thanks!