Lock Down That Noise — Don’t Let It Escape


How can I prevent switchingmode power supply noise from devastating my circuit performance?

RAQ:  Issue 37


With great difficulty but it can be done.

Switching-mode power supplies are inherently the noisiest circuits imaginable. A large current from the supply is being turned on and off at high frequency with very fast dI/dt. There are inevitably large fast voltage and current transients.

The only way to prevent interference to sensitive circuitry in the system is to keep the transients within the converter. We cannot stop switching large currents inside it, but we can, and must, prevent the transient currents and voltages from escaping. Start by grounding all the terminals of the converter at AC.

Capacitors block DC but have low impedance at AC, so they should be ideal for this purpose. In theory, if we place a large capacitor between the converter input and its ground, the input will keep the capacitor charged and the transient currents will flow in the capacitor and not from the power source. A similar output capacitor absorbs transients and sources steady DC.

Unfortunately when we actually build such a system it is common to find much more noise than we can tolerate in the input and output circuits - what can be wrong?

If the capacitor is placed some distance from the converter, the impedance, resistive and inductive, of its connection to the converter will be large enough to prevent it from working properly; if it is chosen badly it will have higher impedance than the basic 1/2πfC formula predicts; and if it shares its ground path with other circuitry the noise in the common ground impedance will be disruptive.

Add to these effects the less important, but still damaging, consequences of external currents induced by fast-changing magnetic and electrostatic fields, and even electromagnetic radiation, from within the converter, and it is obvious that preventing converter noise is not simply a matter of placing a couple of random capacitors on its input and output.

Silencing a DC-DC converter requires systematically finding all the possible paths by which noise can escape from it, and ensuring that they are all locked down. The linked article discusses how this may be done.

Of course before we start we must choose or design the converter itself to have minimal external noise. This is a separate issue which may be discussed in a future RAQ.



James Bryant

James Bryant was a European applications manager at Analog Devices from 1982 to his retirement in 2009 and he still writes and consults for the company. He holds a degree in physics and philosophy from the University of Leeds and is also C.Eng., EurEng., MIET, and an FBIS. In addition to his passion for engineering, James is a radio ham and holds the call sign G4CLF.