TUSB8044A: using the TUSB8044A HUB, the problem of plugging and unplugging peripherals may not be recognized.

Hi Benson:

    I can't see schematic clearly, Can you send schematic in PDF file?

    The printer is super speed, high speed or full speed  device? 

Regards

Brian

hi Brain

Thank you for your prompt reply!

Sorry, I don’t know how to upload a PDF file. I have already replied with the PDF via email.

Thanks again for the support!

Printer information：

Interface: Hi-Speed USB (USB 2.0)

pro.sony/.../up-d898md

did you send PDF fie to my email? I can't find it.

or you can accept my friendship request and send with messaage.

Regards

Brian

I have accepted your invitation and sent you the sch PDF

 Thank you again for your support!

Hi Brian

Did you successfully receive the SCH pdf I sent?

After several days of testing, the problem still exists. Please check whether the SCH is correct and let me know.

Thank you so much!

Hi Benson:

    Can customer remove pulldown resistor on all PWR_EN pin? I didn't see other issues in schematic.

Also can customer take waveform for 1.1v/3.3v/GRST?

Regards

Brian

hi brian

1) remove pulldown resistor on all PWR_EN pin

--I'll remove it later and test again. the IC has a default pull-down, and resistor is indeed repeated.

2) about power-up timing

Test chart

3) My question:

Sch design is based on instructions in the "TUSB8044AEVM User's guide"，VDD1.1V is also after VDD3.3V,

but datasheet  requirements vdd1.1V must before VDD3.3V, Is there any problem?

thanks a lot!

from this waveform, power sequence should be fine. Does customer has USB analyzer?

1. After several days of investigation, we found that there are differences in the performance of the Linux and the Windows system.Our host is a Linux system.

we suspect that multiple plug-and-unplug operations may cause mounting abnormalities or USB IO blocking?

Is there any relevant technical documentation that I can refer to?

2. There is another problem. When the USB input port is quickly plugged and unplugged, the response of the host will be slowed down.

For example, if you plug and unplug quickly 5 times, it will take about 40 seconds for the host to respond.

Is there any way to reduce the response of fast plugging and unplugging? , try to reduce the filter capacitance of host vbus?

3. In addition, as mentioned in the previous article, SCH's VDD33 does not meet the requirements of the datasheet before VDD11.

Does it need to be modified?

Hi Benson:

   for Q1: what difference you saw between Windows and Linux? Windows has no issues?

  Q2: 40 seconds did too slow, do you guys have USB analyzer to take USB trace . if not, can you take waveform on DP/DM and Vbus?

  Q3: based on you waveform, it should be ok even if 3.3v is on earlier than 1.1v.

Regards

Brian

Hi Brian

Thank you for your patient support!

Q1: The probability of USB not being recognized when connected to a PC is significantly reduced.
Q2:Sorry, we don’t have USB analyzer

Today，The mouse was not recognized once. At this time, the oscilloscope measured that the clock was not working.

The waveform is as shown below（Waveform burrs are due to measurement reasons and can be ignored.）

Please tell me, what are the conditions for the HUB clock to work?
Under what circumstances will the CLK exception in the picture above occur?

Crystal stop could be USB2 suspend happened when DP/DM idle more than 3ms.

It can be monitored on pin 45, but you need to enable stsOutputEn first.

what are the conditions for the HUB clock to work?

  Hub clock should work after hub is powered on and upstream port is connected to host.

Regards

Brian

hi Brain

It is certain that the HUB is power on and the VBUS signal can be detected. In this case, can we speculate that the DP/DM abnormality of the UFP port caused the CLK abnormality?

Judging from the analysis we have done, there is likely to be a problem with the Host's USB port, so we need to know how to monitor the status of the UFP port.
We have not configured the external interface (I2C, SMBUS) and E2PROM. How can we monitor the status of upsteam HS?

thanks again!

For Windows, you can use  usbtreeview tool to monitor upstream port connection.

You can search  usbtreeview tool from internet.

Regards

Brian

Hi Benson:

    Any update?

Regards

brian

Hi Brian

Unfortunately, the problem still exists. We found through the "usbtreeview" that sometimes the host only enumerates 1 hub (normally usb3.1 there are 2 hub), and the device decognized will fail at this time;

We captured the UFP(host-hub)  enumeration waveformas follows:
1. Waveform when hub enumeration is ok

2. Waveform when hub enumeration fails

A) When enumeration fails, the hub has detected the host access and has pulled up DP, but the host has not given a reset signal;

B)  At this time, we detect the status of the hub IC: reset pin50、pin48 voltage both normal, and pin62 clock works normally;

Please advise, can this phenomenon rule out the problem on the hub chip side, and do we need to focus the analysis on the host side?

You are such a good man, thank you for your continued support!

hi Brain

We have a small amount of hub production needs. We need to make the PCB boards in the next few days. Please help us review the sch again. If there is anything that needs to be modified, please tell me in time
I will send sch PDF to you through friends.

thank you very much!

I got sch and will review by today.

Regards

Brian

For schematic, what's the meaning of green circle? also can you use 5v to 1.1v instead of 3.3v to 1.1v?

Regards

Brian

From failed waveform, what is DP voltage ?

did you try different host or PC to see any improvement?

Do you have TUSB8044A EVM to try?

Regards

Brian

Hi Brian

1) The green circuit is an optimization point compared to the previous version and can be ignored.  What is sent to you is the latest sch

In addition, Please help to check carefully the correct of IO default states of the state machine.   pin37-pin42、pin45

2) We will consider converting 5V to 1.1V；But, as the information earlier, the reset take effect until 10ms after 1.1V and 3.3V stabilize.

3) From failed waveform, DP voltage is 3.07V

4) As the information provided earlier, using a PC will improve things. And we don't have an EVM board

I would like to know if the enumeration failure waveform can to determine whether it is a problem on the host side? It has nothing to do with the hub ic end?

Thank you so much!

Hi Benson:

for Q1: default setting for pin37-42 and 45 is correct, device is in I2C mode.

Q2: I looked at above supply waveform  and it looks ok.

Q3: recommended to order TUSB8044A EVM which will help debug.

   From the enumeration failure waveform, it did shows host didn't respond correctly, Was host Vbus on during failure mode?

Regards

Brian