With hard disk controllers, it was an interesting problem, particularly when one used a crystal oscillator to drive the transmit circuit on an HDC, while using a PLL containing a VCO to lock to the received data stream. The transmit oscillator drove lots of synchronous circuitry within the controller IC, which then imposed that clock on Gnd and Vcc, therefore requiring a lower output impedance on the VCO steering voltage, charge pump, and even the phase detector so that the Vcc-to-Gnd noise didn't mess with the filter on the steering voltage. A lock-detector helped by switching the time-constant of the filter once the PLL was locked. This was tricky, though.

RE