Due to the way the USB 2.0 protocol works, the 1480A capture must be started before the device under test is connected. This is the reason:

When a HS-capable device connects to a USB 2.0 host, it first connects as a FS device. The reason for this is that the device may be connected to a USB 1.1 host controller (which doesn't support HS). The device then does a 'device chirp', where it issues KJ signaling to the host. This allows a HS host to know that the connected device is HS-capable.

If the host is also a HS host, it will respond with KJKJKJ signaling, telling the device that the  host is also HS capable. At this point, both the host and device knows that the other side is HS-capable so both switch their transceivers into HS mode.

The 1480A is tracking this exact sequence. If the 1480A does not see the chirp sequences then it would not know to switch its transceiver into HS mode. Therefore, the 1480A capture must be started before a HS device is connected. If the 1480A capture is started after the chirp sequences have already been exchanged on the bus under test, there is no way the 1480A can know whether FS or HS is being used.

Note that the default 1480A transceiver mode is FS. Therefore, the 1480A will capture FS packets even if the 1480A capture is started *after* the device under test is connected. This sometimes also works for LS mode, depending on the traffic being communicated.

Most USB 2.0 analyzers have the same requirement as described above.