TUSB8041: 3.1 not showing in any OS, 2.1 is fine, power sequence ?

Can you send schematic for review?

Regards

Brian

Hello Brian, which format would you like ? Screenshots ?

pdf

TUSB8041_IMPLEMENTATION.pdf

Here is it, thanks for your help

let me review and back to you soon.

Regards

brian

TUSB8041_ROUTING.pdf

For what it's worth, here you can check the routing as well...

1: upstream port is connected to USB connector J4? If so, AC cap on RX should be 330nf.

2: all USB down stream ports are connected to  J2? you need to place 100nf on all downstream TX lanes.

3: what is USB1-TX/RX?

4: Cap on GRSTz is too big, 1uf should be enough.

Regards

Brian

Hello Brian, thanks a lot for your answers. Here is what I can input:

1: Upstream port is J4 indeed. I put 100nF AC caps on both lines, as shown on the inserted pic. I don't remember seeing 330nF on the datasheet for upstream port implementation ? Did it change since I implemented it ?

2: On downstream ports, for a matter of design and form factor, I had to put the AC caps on all lines on the motherboard, the closest possible to the device-side USB connectors. Is it alright ?

3: USB1 is not related to the hub on this design, it is the passthrough USB port from Up Squared to my motherboard.

4: GRSTz is implemented with TLV803S reset IC, which from rememberance requires a 10uF decoupling cap on power line.


I will try to implement AC caps on upstream lines with 330nF parameter, and only on RX line this time. Although, again for a question of form factor, I can't quite get them close to the IC. Is this a problem ?

I was also wondering if PCB fiber weaves orientation could have an effect on my problem ?

As mention page 3 of this TI guideline: https://www.ti.com/lit/an/spraar7i/spraar7i.pdf?ts=1658832474362&ref_url=https%253A%252F%252Fwww.google.com%252F

The weaves should be angled from the traces...

Hello Brian, good morning.

I have processed to the following tests on my prototypes:

1: Added 2x 100nF caps on top of the ones already present on upstream RX lines for a total of 300nF, and let 100nF on TX lines. Result: same problem.
2: Kept 300nF on upstream RX, removed caps and bridged on TX lines. Result: same problem.

3: Removed all caps on all lines and bridged. Result: same problem.

The hub is still detected as Compliance mode in Windows, and seems to be in a retry sequence, as it continuously appears and disappears from device manage. It doesn't identify, doesn't enumerate, only says there was a problem with this USB device.

I don't know what to do. For the test, I ordered new prototypes with modified AC caps values, and re-routed upstream RX lines so there is no chance of cross-talk with other lines. I'm waiting for delivery, but after today's tests, I highly doubt it will change anything.

there is 100nf cap on the TX side from the other side of USB connector, so either remove 100nf on RX side of upstream or leave it with 330nf.

TX of upstream is still 100nf.

the compliance mode could be other reason. But AC cap needs to be correct first.

Regards

brian

Hello Brian, thanks for that answer, it made me realize something I haven't thought about yet:

The USB upstream connector is actually a male connector on my PCB, which goes into a female port on the Up Squared board... Could it be that I inverted RX and TX pairs ? I know for a fact that in standard USB-C male to male cables, one RX goes to TX on the other side, which means the inversion is made by the cable... As I'm not implementing TUSB8041 as a standalone hub, I wonder if I shouldn't invert both lines... 

Would you by any chance have a scheme/example where TUSB is implemented as a hub with a cable going out ? Or with a female port ?

About the caps, I will receive the new prototypes next week with corrected caps values, will let you know about it.

Thanks again for your time

Johnny

yes, USB TX/RX inversion is made by cable. I fyou have customized USB connector or cable, you need to be careful to make sure, TX of the host  is connnected to RX of the device..

Regards

brian