So, what is the right modulation scheme for automotive cockpit applications such as navigation and audio and infotainment systems? Does it indeed matter? Designers often face a choice between SMPS ICs with two fundamental schemes: Pulse Width Modulation (PWM) or Pulse Frequency Modulation (PFM). The classic PWM utilises a constant period and a variable on time, while PFM utilises a constant on-time and a variable off-time.

Fig. 1 – For PWM TK is constant. For PFM aK is constant. ɛK is typically zero for both approaches
PWM’s constant period gives a fixed frequency. That fixed frequency gives a known fundamental and identifiable harmonics. The trade-off, relative to PFM though, is lower efficiency as periods of extended off-time and lower frequencies, which apply for PFM, don’t apply for PWM. While PFM offers benefits of lower quiescent current, at low duty cycle and frequency, there is a catch. PFM’s variable off-time equates to a variable period, per cycle; in other words variable frequency. A fundamental frequency of an unknown quantity is liable to be difficult to resolve in terms of EMC, catching out the unaware (although in some circumstances it can help with spectral spreading). This matters in the automotive cockpit but it matters particularly near a radio tuner. So, in this environment the use of an SMPS IC utilising PWM, at a frequency where the harmonics will avoid reception bands, in the system on-state is a necessity to ensure reception integrity. Use of PFM will only lead to EMC and audio quality headaches.
What of pulse-skipping though? Sometimes control schemes that utilise pulse-skipping are encountered, often known as hysteretic, burst-mode or bang-bang control. These utilise a different control scheme (with no error amplifier to integrate the output voltage error) but result in pulse-skipping behaviour. Pulse-skipping implies a variable period, so control of the fundamental frequency with respect to EMC and audio reception becomes a concern again. This doesn’t just apply to pure pulse-skipping schemes, such as hysteretic control. PWM topologies typically reach a low duty cycle limit, determined by their minimum on-time, beyond which they will either skip switching cycles or fold-back their frequency to extend that low duty operation. Some PWM schemes can offer a sufficiently small minimum on-time that such pulse skipping is not encountered within the usual automotive battery range. The ON Semiconductor NCV890xxx family of 2 MHz buck converters for example, will allow 18V:3.3V conversion at 2 MHz, without pulse skipping or frequency fold-back. Always check the minimum on and off times!
So, does the choice of modulation scheme for automotive cockpit SMPS matter? Yes! While PWM is preferable, it’s a case of buyer beware. Not all PWM schemes are born equal.