USB 3 Repeater cable (SN65LVPE502 redriver)

Hello Simon,

A USB 3.0 host or device has actually two different physical connections: one for USB3.0 and one for USB 2.0

If you have a USB 2.0 Host and you connect a USB 3.0 device, the connection will be made always over the USB 2.0 physical wires and the USB 3.0 signal wires will be unused.

The SN65LVPE502 only re-drives the USB 3.0 differential signal pairs, it behaves like a pass-through for the USB 2.0 signals.

Go to the below link for more information.

Regards.

http://e2e.ti.com/support/interface/digital_interface/f/130/t/163721.aspx

Hello Elias,

thanks for your reply!

If the USB-3 physical wires are not connected (e.g. USB-2 host <-> USB-3 device, without redriver), it work's without problems.  The connection is established through UBS-2 physical wires like you described.

But when we use the "active cable" with the redriver, the fall back to USB-2 does not work. The problem seems to be the "far end receiver termination detection" (USB-3 spec Fig. 9-1), that the USB-3 device uses to detect if there is a receiver at the USB-3 physical wires. The redriver does not set the input termination to HiZ in this configuration and the device thinks there is a USB-3 receiver and does not fall back to USB-2.

Configuration not working:

USB-2 Host <----------------------USB-2 physical wires ---> USB-3 Device
                                      redriver <--USB-3 phy. wires------>

Regards,

Simon

Simon,

The LVPE502 is actually detecting the USB 3.0 device because you do have a USB 3.0 device attached, that is why it is not setting its RX TX lines to Hi-Z.

Anyway, your host is USB 2.0, so your host is not connected to the LVPE502 at all, I don't understand why your host would "look" for the Hi-Z of the LVPE502.

Can you elaborate more on your application, do you have the cable's schematic, diagrams, etc? What is your USB 3.0 device?

Regards.

Thanks Elias,

ELIAS VILLEGAS said:

The LVPE502 is actually detecting the USB 3.0 device because you do have a USB 3.0 device attached, that is why it is not setting its RX TX lines to Hi-Z.

Anyway, your host is USB 2.0, so your host is not connected to the LVPE502 at all, I don't understand why your host would "look" for the Hi-Z of the LVPE502.

No, it seems that the device "looks" for the Hi-Z of the LVPE502 and does not fall back to USB-2. And because the host does not support USB-3, no connection is established.

ELIAS VILLEGAS said:

Can you elaborate more on your application, do you have the cable's schematic, diagrams, etc? What is your USB 3.0 device?

Our idea is to ship the "active cable" with the LVPE502 with our device (Cypress FX3) because we need more than 3m cable length (in our device there is already about 1.5m cable). But we also need to be compatible with USB-2.0 hosts.

Simon,

Do you have the schematic of the active cable?

The LVPE502 should set its RX and TX to Hi-z when any of its two sides is not connected, probably your cable is terminating the RX terminals of the LVPE502.

Also, what is the part number of the device?

Regards.

Elias,

I sent you the schematic by email. The device is a Cypress FX3 (http://www.cypress.com/?id=3526) and we also tested with an USB-3.0 harddrive from Western Digital (WDBACY5000ASL).

Simon

Hello Simon,

I've just double-checked and the LVPE502 enters to RX.Detect when there is no device detected on any of its channels.

Check that the USB 3.0 type-A connector does not have any termination on the SS lines, also, your USB 3.0 device should fall back to USB 2.0 when there is no far-end  receiver termination or when the link training was unsuccessful, so even if the LVPE502's receiver is not in Hi-Z the devices should connect over USB 2.0, can you check this with your device.

Regards.

Hi Elias,

thank you for your commitment. I made some further tests:

1) To make sure that the type-A connector does not have any termination on the SS lines, I cut the SS lines at the redriver PCB (RX and TX, host direction). But the problem is still there (LVPE502 receiver not Hi-Z!). I also measured the type-A connector and found no termination on the SS lines.

2) I forced the LVPE502 to sleep mode (Pin RX_EN -> GND), now the device is detected (-> LVPE502 receiver Hi-Z, device falls back to USB-2).

3) I cut only the TX lines in device direction (LVPE502 -> device), now the device is also detected (-> LVPE502 receiver Hi-Z!).

It seems that the LVPE502 does not set it's receivers to Hi-Z when just one receiver (USB-3.0 device in this case) is connected.

I also don't understand why the devices do not fall back to USB-2.0 even if the LVPE502 receiver is not in Hi-Z mode. I tested another USB-3.0 device (USB stick) and the behaviour is the same. We will have a closer look at the device behaviour of the Cypress FX3 next week, but I have not much hope that the problem lies there because other devices also do not work.

Regards,

Simon

Hello Simon,

Do you have the "CP" version of the LVPE502?

Because I think you are seeing a known problem described on the errata, this problem is fixed on the "A" version.

You can see the errata at http://www.ti.com/lit/er/sllz070/sllz070.pdf

Regards.