TABLE OF CONTENTS Several years ago, a senior engineer from a well-known U.S.-based OEM visited us in Singapore to "try us out." What his employer shared with the company we were already doing business with was that both were massive corporations that had had their humble (but spectacular) beginnings in garages. But this engineer also produced a strange wish list make me the quietest 65W Flyback ever.
How quiet? A standard noise and ripple measurement setup is shown in Figure 11-1. Between the power supply and the load (in this case a board with appropriate high-wattage resistors) lies a fairly long cable harness. Since the power supply is open-frame, the cables leaving the power supply always manage to pick up a great deal of the noise being emitted from it (via radiation or conduction). And since the scope probe is to be hooked up at the load end, a small 0.1 ?F disc ceramic capacitor is ( almost) always allowed to be present at the point of measurement (as it would be in real-world systems, too).
I said "almost," because the visiting engineer had just thrown us an interesting challenge (in between all the caviar my Boss was busy stuffing him with). He wanted the "usual" maximum of 50mV noise and ripple on the 5V output, measured with a full 100MHz bandwidth analog scope (with automatic cursors that include...
