AM6412: USB cannot be detected

Hi Tashrio-san,

I am routing your query to our USB hardware expert for comments.

Hello  Tashrio-san,

I will need additional inputs including the schematics to review and comment.

Has customer followed the EVM or is this custom implementation.

Regards,

Sreenivasa 

Hi Sreenivasa,

I will send you the schematics offline.

Thanks and regards,
Koichiro Tashiro

Hello  Tashrio-san,

Thank you and Noted.

Regards,

Sreenivasa

Hi Sreenivasa,

I got additional question from the customer.
What is USB0_VBUS input threshold (VIH, VIL)?
Datasheet just says it is USB 2.0 spec compliant.


USB 2.0 spec. DC Electrical Characterization seems based on 5V VBUS, but AM64x Design guideline requires an external resister divider.



Thanks and regards,
Koichiro Tashiro

Hello  Tashrio-san,

Thank you for the schematics.

Can you check with customer the length of the USB signals and if Rcalib is placed near the connector or the SOC.

What is the VBUS voltage customer is connecting -  based on the voltage divider used without the zener, the voltage is expected to be in the range of 4.75 to 5.25V 

Regards,

Sreenivasa

Hello  Tashrio-san,

Koichiro Tashiro said:

I got additional question from the customer.
What is USB0_VBUS input threshold (VIH, VIL)?
Datasheet just says it is USB 2.0 spec compliant.

The VBUS voltage divider is 2:1

The VBUS input range is 4.75V to 5.25V .

1/3 X 5V is the expected nominal value.

Regards,

Sreenivasa

Hi Sreenivasa,

Kallikuppa Sreenivasa said:

Can you check with customer the length of the USB signals and if Rcalib is placed near the connector or the SOC.

Let me check it with the customer.

Kallikuppa Sreenivasa said:

1/3 X 5V is the expected nominal value.

As VBUS input range is 4.75V(min.) to 5.25V(max.), usb0_vbus input range should be 1.58V(min.) to 1.75V(max.) ?

Thanks and regards,
Koichiro Tasihro

Hello  Tashrio-san,

Thank you.

Please red below section.

9.2.3 USB VBUS Design Guidelines
The USB 3.1 specification allows the VBUS voltage to be as high as 5.5 V for normal operation, and as high as
20 V when the Power Delivery addendum is supported. Some automotive applications require a max voltage to
be 30 V.
The device requires the VBUS signal voltage be scaled down using an external resistor divider (as shown in
the Figure 9-4), which limits the voltage applied to the actual device pin (USB0_VBUS). The tolerance of these
external resistors should be equal to or less than 1%, and the leakage current of Zener diode at 5 V should be
less than 100 nA.

In customer use case , customer did not use the Zener. So the input range could be limited.

Regards,

Sreenivasa

Hi Sreenivasa,

The customer already checked that part in datasheet, but it does not directly answer to their question below.
They are not asking VBUS voltage range in general, but AM64x USB0_VBUS VIH, VIL.

Koichiro Tashiro said:

What is USB0_VBUS input threshold (VIH, VIL)?

Could you provide USB0_VBUS VIH, VIL?

Thanks and regards,
Koichiro Tashiro

Hello  Tashrio-san,

Thank you.

Please refer below explanation. 

• The USB VBUS 5V typically has a range of 4.75V to 5.25V.

  The USB PHY only asserts VBUSVALID if the voltage on VBUS is >=4.4V at the time of measurement.

• The USBn_VBUS pin is an analog input pin that should be connected to the VBUS signal as shown in USB VBUS Design Guidelines section of the datasheet.

  The circuit that monitors VBUS only checks for specific thresholds which are defined by the USB specification and provides the status to the USB controller via internal signals. Providing USB electrical characteristics is beyond the scope of the datasheet. We expect customers to reference the appropriate USB specifications provided by the USB Implementors Forum (USBIF). Please refer to the documents found on usb.org. 

  Additional information in case the USB is configured as host (USBx_VBUS connection is optional)

  Software doesn't read/know the exact threshold. The USB controller generates interrupts when VBUSVALID changes state. By the way, Linux doesn't use this VBUSVALID interrupt, since software doesn't have to do anything when short-circuit happens. The Linux driver is not checking the status of VBUS to determine if there is a fault condition. In that case, you should connect the fault output of the VBUS load switch to a GPIO and configure the GPIO to generate an interrupt that indicates there has been an over-current condition.

  I will have to check with the USB expert if you need additional inputs.

  Regards,

  Sreenivasa

Hi Sreenivasa,

Thanks for your explanation. The customer still wants to know "specific thresholds" in below comment.
Based on previous discussion above, the input voltage for USB0_VBUS must be over 1.58V for proper operation?

The circuit that monitors VBUS only checks for specific thresholds which are defined by the USB specification and provides the status to the USB controller via internal signals. 

Thanks and regards,
Koichiro Tashiro

Hello  Tashrio-san,

Thank you.

Let me check with the USB update and update you.

regards,

Sreenivasa

Hi Sreenivasa,

Do you have any updates to this item?

The customer has two additional questions.
- Are there any software configuration available to change D+/D- voltage level smaller such as 0.4V?
- Please provide an internal circuit of USB_VBUS.

Thanks and regards,
Koichiro Tashiro

Hello  Tashrio-san,

Thank you.

Koichiro Tashiro said:

Are there any software configuration available to change D+/D- voltage level smaller such as 0.4V?

We do not have provision to configure the voltage level

Koichiro Tashiro said:

Please provide an internal circuit of USB_VBUS.

This is not something that i believe is available. 

Just to be sure, I will check the device expert.

Regards,

Sreenivasa

Hi Sreenivasa,

Kallikuppa Sreenivasa said:

Just to be sure, I will check the device expert.

Did you get any inputs from the expert?

Thanks and regards,
Koichiro Tashiro

I meant both USB0_VBUS threshold and an internal circuit.

Thanks and regards,
Koichiro Tashiro

Hello  Tashrio-san,

Thank you.

Let me follow-up and update you.

As i said the answer is this request may not be supported but i am reconfirming with the expert.

Regards,

Sreenivasa

Hello  Tashrio-san,

Our device expert confirmed that we may not be able to support the above request.

Regards,

Sreenivasa

Hi Sreenivasa,

The customer has one more question.
Is there any register(bit field) available for CPU to check if VBUS is detected or not?
The customer is using Linux, so Linux console command is also acceptable.

Thanks and regards,
Koichiro Tashiro

Hello  Tashrio-san,

Thank you.

I will assign the query to an USB software expert.

Regards,

Sreenivasa

Hi Bin,

Thanks for the patch. The customer tried it and found the VBUS was properly detected both on failing boards and normal boards.
For next step, they wants to check USB ID status and how D+/D- are configured on AM64x side.
Could you tell me how to check them?

Thanks and regards,
Koichiro Tashiro

Hi Bin,

Thanks for your quick reply. The customer will check it.
Could you also tell me how to check D+/D- status in linux?

Thanks and regards,
Koichiro Tashiro

Hi Tashiro-san,

D+/D- are USB data lines, which changes rapidly according to the transferred USB packets, there is no such thing of "D+/D- status" to report in Linux.

Hi Bin,

Thanks for your support on this.
The customer thinks the issue is related to the device failure and asked for failure analysis.
This item is now closed.

Thanks and regards,
Koichiro Tashiro

Hi Tashiro-san,

Thanks for the update. I am closing this thread.