Power Supply Design 527 What about the output ripple with this approach? I have just measured a power supply using the exact circuit of Figure 18.3b, with 2200 uF reservoirs, and I found -79 dBu (87 uV rms) on the +17 V output rail, and -74 dBu (155 uV rms) on the 17 V rail, which is satisfyingly low for inexpensive regulators, and should be adequate for almost all purposes; note that these figures include regulator noise as well as ripple. The load current was 110 mA. If you are plagued by ripple troubles, the usual reason is a rail decoupling capacitor that is belying its name by coupling rail ripple into a sensitive part of the ground system, and the cure is to correct the grounding rather than design an expensive ultra-low ripple PSU. Note that doubling the reservoir capacitance to 4400 uF only improved the figures to -80 dBu and -76 dBu respectively; just increasing reservoir size is not a cost-effective way to reduce the output ripple. 18.4 Using Variable-Voltage Regulators It is of course also possible to make a ±17 V supply by using variable output voltage IC regulators such as the LM317/337. These maintain a small voltage (usually 1.2 V) between the OUTPUT and ADJ (shown in figures as GND) pins, and are used with a resistor divider to set the output voltage. The quiescent current flowing out of the ADJ pin is a couple of orders of magnitude lower than for the 78/79 series, at around 55 uA, and so a