AM3352: USB data receiving process receives data of unexpected size

K.Z

Part Number: AM3352

Tool/software:

Hi All,

We are currently developing a product using a TI CPU (AM3352BZCZ60).

However, in the receiving process of USB data, there is a phenomenon that an unexpected receive data size
We would appreciate it if you could provide us with some guidance.

[Situation]

Execute USB0 data transfer (Bulk out transfer EP1) from PC
↓
Interruption during transfer
↓
Execute USB0 data transfer (Bulk out transfer EP1) from PC again.
・
・
Repeat (Bulk In transfer EP3 is also performed in parallel.)

When the above procedure is repeated, at the first data transfer of data transfer resumption,
USB0 interrupt process, EP1's receive data size is about 16 kbyte,
In rare cases, a value of about 16 kbytes (e.g. 16,237 bytes [0x3F6D]) may be reported.
[Usually, the maximum size of 512 bytes or less is notified].

As long as we check the USB communication, there seems to be no problem with the transferred data,
There may be a problem with the interruption process of USB data reception.
Thank you for your cooperation.

Please contact us so we can send you the file in a private message.

Best Regards,

Ito

over 1 year ago

0 Bin Liu over 1 year ago

TI__Guru**** 170531 points

Hi Ito-san,

What software do you run on AM3352, Linux or RTOS?

0 K.Z over 1 year ago in reply to Bin Liu

Expert 2670 points

Hi Liu-san

My Customer is using TI-RTOS

Best Regards,

Ito

0 Bin Liu over 1 year ago in reply to K.Z

TI__Guru**** 170531 points

I am routing your query to our RTOS expert for comments.

0 Jianzhong Xu over 1 year ago in reply to Bin Liu

TI__Guru 59445 points

Hello Ito-san,

We can no longer support TI-RTOS based SW development for AM335x. Please refer to this announcement.

Regards,

Jianzhong

0 K.Z over 1 year ago in reply to Jianzhong Xu

Expert 2670 points

Hi Jianzhong,

Thank you for your reply.

So could you please give me your opinion excluding TI-RTOS?

We believe that this problem is caused by over or underflow of the receive buffer.

[Other conditions are listed below]

・The average data size at the time of successful reception: 512 bytes (4 bytes several times at the beginning of communication, then 512 bytes are received).
・No error processing is currently in place for abnormal received data size.
・The received data is transferred by EDMA with the same size as it is, and EDMA error interruptions often occur.

Please let us know if there is any way to check the error contents.

Best Regards,

Ito

0 Jianzhong Xu over 1 year ago in reply to K.Z

TI__Guru 59445 points

Hi Ito-san,

Let me loop in our HW team for further support.

Regards,

Jianzhong

0 Kallikuppa Sreenivasa over 1 year ago in reply to Jianzhong Xu

TI__Guru**** 196459 points

Hi Ito-san,

From the hardware point i view, can you confirm if these issues are observed on a specific board or multiple boards on the same design.

Is this something that has been tested on the EVM.

Can you confirm all the required decaps and the USB interface has no hardware issues that could be a probable cause for this issue.

Regards,

Sreenivasa

0 K.Z over 1 year ago in reply to Kallikuppa Sreenivasa

Expert 2670 points

Hi Sreenivasa,

Thank you for your reply.

Kallikuppa Sreenivasa said:
From the hardware point i view, can you confirm if these issues are observed on a specific board or multiple boards on the same design.

The problem occurs in multiple boards.

Kallikuppa Sreenivasa said:
Is this something that has been tested on the EVM.

Since the phenomenon is occurring in the operation of the boards, it cannot be confirmed on the EVM
Please let me know if there are any advantages.

Kallikuppa Sreenivasa said:
Can you confirm all the required decaps and the USB interface has no hardware issues that could be a probable cause for this issue.

What is the operation of decapsulation?

Best Regards,

Ito

0 Kallikuppa Sreenivasa over 1 year ago in reply to K.Z

TI__Guru**** 196459 points

Hello Ito,

K.Z said:
Can you confirm all the required decaps and the USB interface has no hardware issues that could be a probable cause for this issue.

What is the operation of decapsulation?

The caps recommended in the data sheet are recommended to be mounted for proper operation of the processor.

Regards,

0 K.Z over 1 year ago in reply to Kallikuppa Sreenivasa

Expert 2670 points

Hi Sreenivasa,

Do you have any reference material for caps?

Best Regards,

Ito

0 K.Z over 1 year ago in reply to Kallikuppa Sreenivasa

Expert 2670 points

Hi Sreenivasa

I don't know much about encapsulation, but can you tell me the specific operation?
How is proper de-encapsulation related to the problem of large amounts of data being sent at once?

Best Regards,

ito

0 Kallikuppa Sreenivasa over 1 year ago in reply to K.Z

TI__Guru**** 196459 points

Hello Ito,

Thank you.

We have seen multiple customer board debug issues being resolved based on the review of the recommended caps.

Please refer to the below section of the data sheet and help confirm you have the required caps in place.

5.9 External Capacitors

7.7.2 mDDR(LPDDR), DDR2, DDR3, DDR3L Memory Interface

Regards,

Sreenivasa

0 K.Z over 1 year ago in reply to Kallikuppa Sreenivasa

Expert 2670 points

Hi Sreenivasa

Thank you for your reply.

What is de-encapsulation?
Is it to de-encapsulate encapsulated packets?

I don't know how to deactivate it and how to check if it has been deactivated, please give me detailed deactivation instructions.

5.9 External Capacitors
7.7.2 mDDR(LPDDR), DDR2, DDR3, DDR3L Memory Interface
→I'll check this one.

Best regards,
ito

0 Kallikuppa Sreenivasa over 1 year ago in reply to K.Z

TI__Guru**** 196459 points

Hello Ito,

Thank you.

K.Z said:
What is de-encapsulation?
Is it to de-encapsulate encapsulated packets?

Point me to the section on the thread where this is mentioned.

I need the context for answering the query.

Regards,

Sreenivasa

0 K.Z over 1 year ago in reply to Kallikuppa Sreenivasa

Expert 2670 points

Hi Sreenivasa

Thank you for your reply.

Kallikuppa Sreenivasa said:
Can you confirm all the required decaps and the USB interface has no hardware issues that could be a probable cause for this issue.

About this statement.
What does Decaps mean and how do I confirm it?

Best Regards,

ito

0 Kallikuppa Sreenivasa over 1 year ago in reply to K.Z

TI__Guru**** 196459 points

Hello Ito,

Thank you.

Decap is the decoupling capacitors that are recommended for the processor.

Kallikuppa Sreenivasa said:
5.9 External Capacitors

7.7.2 mDDR(LPDDR), DDR2, DDR3, DDR3L Memory Interface

Refer below sections of the data sheet.

Regards,

Sreenivasa

0 K.Z over 1 year ago in reply to Kallikuppa Sreenivasa

Expert 2670 points

Hi Sreenivasa

Understood. Thank you.

[7.7.2 mDDR(LPDDR), DDR2, DDR3, DDR3L Memory Interface]
What exactly should I check in this section?

Best Regards,

ito

0 Kallikuppa Sreenivasa over 1 year ago in reply to K.Z

TI__Guru**** 196459 points

Hello Ito,

Thank you.

Recommendations for the capacitor values have been provided in this section.

You need to compare the schematics and confirm the recommended values are added on the board.

Regards,

Sreenivasa

0 K.Z over 1 year ago in reply to Kallikuppa Sreenivasa

Expert 2670 points

Hi Sreenivasa

Thank you for your reply.

I checked and the following values are different from the datasheet.

VDDS_SRAM_MPU_BB
Data seat : 10.01μF
Board : 0.01μF

VDDS_SRAM_CORE_BG
Data seat : 10.01μF
Board : 0.01μF

According to the HW designer, the values here are the same as the EVM.
Is there any other part of the HW that could be the cause or register information that should be set?

Best Regards,

Ito

0 Kallikuppa Sreenivasa over 1 year ago in reply to K.Z

TI__Guru**** 196459 points

Hello Ito,

Thank you for the inputs.

I would suggest you try adding the caps with shortest connections and do a quick test.

Regards,

Sreenivasa

0 K.Z over 1 year ago in reply to Kallikuppa Sreenivasa

Expert 2670 points

Hi Sreenivasa

Thank you for your reply.

Add caps for a total of 10.01 μF and check.
I would like to explain to the customer the background of this check.
What problems can be expected due to the incorrect capacitance of VDDS_SRAM_MPU_BB and VDDS_SRAM_CORE_BG?

Best Regards,

Ito

0 Kallikuppa Sreenivasa over 1 year ago in reply to K.Z

TI__Guru**** 196459 points

Hello Ito,

Thank you,

I am not sure if this is a problem due to the caps.

When the right caps are not installed, we have seen unexpected behavior.

Regards,

Sreenivasa

+1 Kallikuppa Sreenivasa over 1 year ago in reply to Kallikuppa Sreenivasa

TI__Guru**** 196459 points

Hello Ito,

Here are some additional inputs i received:

This appears to me to be a driver issue. When they say “process interrupted” I can’t tell if they mean a physical CPU interrupt or some kind of single-threaded “I need to do something else for a bit” routine.

Bulk transfer MAX_PACKET (size) for USB2.0 is 512 bytes. If they are seeing 16KB get transferred as they say they do, it almost looks to be a driver issue.

As this is TI-RTOS, no help will be forthcoming on that front.

Customer could rent/borrow/buy a protocol analyzer and verify to be sure.

Regards,

Sreenivasa

Processors

Processors forum

AM3352: USB data receiving process receives data of unexpected size