TUSB4041I: USB downstream port occasionally DP/DM crossover

Hello Tian Zhang,

Please describe the failure when the DP/DM crossover happens?  The only way to change the polarity on a downstream port is by writing the registers of the hub through EEPROM or SMBUS/I2C host over SDA/SCL.

Regards,

JMMN

Hello，thanks for your reply，

1.In 50% of all products, DP/DM crossover occurs when they are powered on for the first time

And another question like this

2.The CPU fails to identify the TUSB4041 chip ID

Please share a schematic.  How is "DP/DM crossover" determined?

It can be seen from the schematic diagram that we did not connect EPROM or SMBUS/I2C host over SDA/SCL, but when we actually used it, we found that if the downstream end cannot be used, it can be used by crossing DP/DM. Caused by?

Hello,

Can you confirm my understanding?

• EEPROM / SMBUS is not normally used.
• 50% of the time on power up the CPU cannot find the hub.
• Crossing DP/DM fixes the issue when the CPU cannot find the hub.

What registers are you writing when crossing DP/DM?

Does this problem only occur on the upstream port?

Can the ARM be crossing DP/DM?

Regards,

JMMN

Hello,

1、EEPROM / SMBUS is indeed not connected;

2、The CPU always can find the hub，but 50% of the time on power up the hub downstream port occasionally DP/DM crossover;

3、Crossing DP/DM fixes the issue when the the hub downstream port DP/DM crossover;

4、I did not write any value to the register when the the hub downstream port  DP/DM crossover;

5、This problem only occur on the downstream port;

6、The ARM's DP/DM has never crossed DP/DM.

Thank you for your reply！

Can you explain how "Crossing DP/DM fixes the issue when the the hub downstream port DP/DM crossover;" - how is DP/DM being crossed to fix the issue?

I stripped the cable of the mouse and exposed the DP/DM. When the mouse couldn't be used, I connected the DP of the mouse to the hub downstream port DM, and the DM to the hub downstream port DP, and found that the mouse can be used.

Hi,

That is very strange, without changing the register setting of the hub, it is not possible for DP/DM to swap.

Are you able to see this behavior with anything other than that mouse?

Can you share a DP/DM scope capture?

Regards,

JMMN

Hello,

1.Can you provide a way to test the register value of the hub? Let's confirm whether it is the default value.

2.U disk will also have this abnormal phenomenon.

3.Do you need DP/DM waveforms?

Hi,

You can read back the registers of the hub using an I2C host or SMBUS host, it cannot be done over USB.

Please provide the DP/DM waveforms, that would be very helpful.

Regards,

JMMN

Hi,

We had read back the registers of the hub using an I2C host or SMBUS host, but nothing.I can try again.

The waveform is as shown below:

There is a question about the above waveform, why does it enumerate twice when the USB host enumerates the TUSB4041?

Hi Tian,

So for reading back the registers, you can connect the hub as usual with SMBUSz set high and no EEPROM attached, then once the hub is active, you can use a SMBUS or I2C host to read back the registers at this address:

The waveforms you shared look correct.  The 2nd reset is typically caused by the software.  The host might enumerate a port, then do a soft reset to it which will cause a 2nd enumeration.

Regards,

JMMN