Part Number: TUSB4020BI
Dear TI Team
We are observing an intermittent issue on a design using the TUSB4020BI as a USB 2.0 hub on our system‑on‑module.
The downstream port in question (PWRCTL1) powers a second USB hub on the carrier board (signal name on our schematic: USB_2_EN). The problem appears only on a few boots and is illustrated in the attached oscilloscope capture (08_lsusb_fail_module_hub.png).
Channel mapping in the screenshot:
- CH1: USB VBUS (pin 9)
- CH2: GRST (pin 11)
- CH3: PWRCTL1 (pin 4)
- CH4: level-shifted PWRCTL1 signal
Observed behavior (failing boot)
During a failing boot, the PWRCTL1 output from the TUSB4020BI shows a “toggling” or “dancing” behaviour:
-
Instead of asserting once and staying high, PWRCTL1 repeatedly toggles high/low during the early boot phase.
-
The reset input to the TUSB4020BI (GRST) looks clean and behaves as expected; we do not see repeated resets from our external reset circuitry.
-
The over‑current signal for this port remains inactive (no over‑current condition observed).
After this “weird” toggling phase, PWRCTL1 eventually stabilizes and remains in the expected state. From that point on, the module hub itself appears to operate normally.
System state from the host’s perspective
Despite the hub apparently recovering and PWRCTL1 stabilizing, the resulting system state (seen from the USB host) is not correct:
-
The TUSB4020BI itself is detected and enumerated by the SoC root port without issues. In Linux, lsusb shows the USB hub.
-
However, the downstream USB hub on the carrier board, which is enabled by PWRCTL1, and all devices behind it are not recognized when the embedded Linux system is up (checked via lsusb).
In contrast, on successful boots we observe PWRCTL1 asserting once and remaining high; in these cases, the downstream carrier‑board hub and all attached devices are enumerated and fully functional.
Question
Is this kind of repeated, self‑initiated toggling of PWRCTL1 a known or intended behavior of the TUSB4020BI under certain internal fault or retry conditions, and could it explain why the hub itself finally works but all devices behind that port are no longer detected by the host OS?
Best regards,
Marvin
