TUSB2E11: USB handshake failure

Hi Jon,

Could you help provide the schematic for both these designs, as well as a register dump from both devices?

Do both boards work at LS/FS? Or does the TUSB2E11 board not work at all?

The TUSB2E221 and TUSB2E11 are very similar in terms of functionality internally, so if one device works, the other should as well. My guess currently is that there may be a difference in configuration causing this issue.

Thanks,

Ryan

Hi Ryan,

The TUSB2E11 do not work at all, It may seem that the host think a LS device is connected but the handshake fails and eventually it does not recognize it. I'm not able to read the registers because the I2C bus is connected to the SoC which i need the USB connection to communicate with. I need it to work as a USB redriver with no setup required. I bypassed the onboard redriver and connected my SoC to an external TUSB2E221QFNEVM eval board and that worked fine out of the box. I suspect there are something with the inband RAP setup from the SoC that fails. Can you confirm that the TUSB2E111 and TUSB2E221 are equal in that regard? To test it properly, I'm making my own external motherboard to test and measure several re-drivers. Among those, I will design in the TUSB2E2211005VBWR but I cant find specifications for the WQFN-FCRLF (VBW) package on the TI site - the link doesn't work. Can you help me with that?

Thanks!

Jon

Hi Jon,

Jon Gustavsen said:

The TUSB2E11 do not work at all, It may seem that the host think a LS device is connected but the handshake fails and eventually it does not recognize it. I'm not able to read the registers because the I2C bus is connected to the SoC which i need the USB connection to communicate with.

It sounds like this is likely a design issue then with your setup. Like you pointed out, the eUSB lines of the repeater should be used to configure the device as a host or a device, so if that isn't happening, that may be an issue.

Again, if you could provide the schematic for this design, I would appreciate it, that way I can review it and see if there are any issues there.

Jon Gustavsen said:

I need it to work as a USB redriver with no setup required.

When you say it needs to work as a USB redriver, could you help clarify? Do you mean just the eUSB2 to USB2 functionality with the equalization and pre-emphasis that comes with the device? Or do you want this to act natively with eUSB2 on both sides?

Jon Gustavsen said:

I suspect there are something with the inband RAP setup from the SoC that fails. Can you confirm that the TUSB2E111 and TUSB2E221 are equal in that regard?

Yes, both these devices should configure through the RAP the exact same way.

Jon Gustavsen said:

Among those, I will design in the TUSB2E2211005VBWR but I cant find specifications for the WQFN-FCRLF (VBW) package on the TI site - the link doesn't work. Can you help me with that?

We are working on having this available on TI.com. At the moment, I can provide you the EVM schematics for the TUSB2E221QFN, which hopefully should provide what you need for layout and schematic. Here is what I have for the package dimensions: 4230575_datasheet.pdf

Thanks,

Ryan

Hi Ryan

See attached sch snip, as you see - it's with a NXP chip - same problem there. I then modified my board with the TUSB2E111 chip.

Yesterday I received TUSB2E221 chips, and modified the design with it - same symptom. Only thing that works is to shortcut eDP+ to DP+ and eDP- to DP- on the my board and use the TI eval kit between my board and my laptop host.

Both power rails are rated for 150mA, should be sufficient. Left side goes straight to SoC eUSB, right side to USB connector. RST line comes from SoC out of reset signal.

I've checked connectivity, polarity and power - all good. Haven't found any noise

I can confirm I want a eUSB to USB device

Initial handshake measured on TI evalkit DP+ and DP- line

This is according to the standard and this setup work

And similar when measuring DP+ and DP- on my board with chip mounted

It starts correctly with device (redriver) pulls D+ high to indicate FS or HS capability

Next, host connects termination resistors - also correct as I understand it

But then, the device shall hold the D- high for >1ms to indicate HS - but something happens and it fails

On my host computer I see it recognizes something on the USB line but gets a descriptor error 

I suspect there are something on the e side, next I will measure and compare the handshake there.

But it is strange that the eval kit works fine out of the box? Any difference is a bit more resistance on the eUSB lines because there a short cable from my board to the eval kit. Do you think that may affect the signal in some way that make a difference?

Anyways, it seems the problem is on my board with the SoC so I think this case may be closed.

Thanks!

Jon

Hi Jon,

I don't see them in the schematic you sent, how are the GPIO pins configured? Are you using I2C for configuring these repeaters or just leaving these pins floating?

Our TUSB2E221 does not share the same pinout as the PTN3222. How exactly are you modifying your board to use our device?

If you are seeing this issue on both NXP and TI devices, that would indicate to me that there is an issue with the design as opposed to the devices. Have you tried different PHY configurations via I2C to see how that improves performance? You should also be able to read registers 0x20 and 0x60 to ensure the repeater is being configured or not. Is RESETB being pulled-up externally?

Would it be possible to provide the layout for review to see if there are any issues there? A block diagram of your system may be good as well.

Thanks,

Ryan