TUSB8020B: Crystal oscillator doesn't start, again

we suspect soldering issue, can you re-soldering the failing unit and test again?

Done but didn't help.
The reset and power logic is functioning allright.

Is a USB host connected to the hub and VBUS is high?

Regards,
JMMN

Yes

Booth options with board powered by VBUS and Hub powered by target is tested.

The circuit get the chrystal oscillating in booth configutations for 60% of the boards.

1. I have got one out of 5 failing boards to work after dismount/remount the same circuit.
    Bad connection for the ground-tab isn't the only problem with the circuit.

2. Variation of resistive (680k, 820k, 1M, 2.2M, None) and captive (5.6p, 12p 18p, 22p) load to the chrystal
    (CTS 445C25D24M00000, 24MHz ±20ppm Crystal 18pF 40 Ohms) gives no result to the port.

3. Variation of the reset delay (2ms, 10ms, manual) gives no difference.

Is there any other IO's involved in the internal oscillating circuitry than 3V3, GND, GRST, XI and XO?
You asked about USB_VBUS-pin and if HOST is connected - does that really make any difference in this case?

Fredrik,

The clock will not run if there is no upstream connection (USB host, plus USB_VBUS high). Also, the clock will stop when the host puts the hub in a low power state. The host will typically do this if there are no devices connected downstream of the hub.

Regards,
JMMN

Hi Fredrik,

Any update on the issue?