Describe the bug

In addition to #70794 (BLE TX buffers should provide for ISO header), there are more bugs in the ISO central sample:

• The sample is not correctly synchronized with the radio clock. Therefore a race condition emerges that ISO sequence numbers will no longer match L2 ISO payload number after a few iterations.
• The sample does not consider rounding errors on nRF boards whose LF clock's period cannot be represented as an integer in microseconds.
• The sample does not take into account that the L2 in the controller systematically skips the second ISO interval (aka payload count/event number) to provide for initial synchronization (i.e. the sequence of events is 1, 3, 4, 5... and not 1, 2, 3, 4)

This is relevant to all nRF platforms that use a 32k sleep clock.

To Reproduce
Steps to reproduce the behavior:

• see the behavior described in Bluetooth isochronous channels central_iso / peripheral_iso example not working reliable on nrf52840dk – packets not received/lost  #70794

Expected behavior
With proper synchronization the sample should be able to run for hours without skipping payload numbers even if CONFIG_BT_CTLR_ISOAL_PSN_IGNORE=n.

Impact
The sample does not provide a proper template for perfectly synchronized transmission.

Logs and console output
see #70794

Environment (please complete the following information):
nRF platforms with 32k sleep clock

Additional context
I have a fixed sample that I can provide as a PR. Just need some time to make it upstream-ready.