Power Tip 61: Watch the conversion ratio on boost converters
Have you ever needed to provide a boosted non-isolated power supply output from a lower-voltage input? The boost converter is a traditional solution. Still, you need to be mindful of limitations on the control IC.
You may be motivated by cost and area considerations to push the power supply operating frequency as high as possible. However, efficiency concerns and controller considerations will limit how high a frequency you can use.
Just as buck power supply controllers have minimum controllable on times, boost controllers have minimum controllable off time. Boosts with wide conversion ratios can create issues when you violate these limits. Consider a boost converter operating in the continuous conduction mode, as shown above. Its duty cycle is
Making some substitutions and solving for maximum operating frequency based on minimum controllable off times,
As an example, a boost that converts 24V to 140V requires a duty factor of 83 percent and an off-time duty factor of 17 percent. In this example, a LM5122 boost controller has a minimum controllable off time of 750nS, which should be guard banded by at least another 250nS. Doing the math, the switching frequency should be limited to 170kHz.
Here is a schematic of a boost converter that was built to provide 140V at two amps from an input voltage of 24V.
This design is interleaved; there are two power stages running 180 degrees out of phase. The current is balanced between the two stages by circuitry in the top controller (master) that sets the input current in each stage with resistive current sensing of the inductor current. The master controller also controls the clock phase and frequency for both stages, as well as softstart and faults.
Running the boost stages 180 degrees out of phase provides a number of advantages. With 280W of output power and a 90+ percent efficient power supply, there is 18W of power dissipations. Two phases provide the opportunity for precise control of currents and hence dissipations in the semiconductors and inductors. It also spreads the heat to facilitate cooling.
Since the phases are running 180 degrees apart, there is input and output ripple current cancellation, which reduces both the peak and the RMS ripple current in the capacitors. Since the inductor currents are out of phase, the effective ripple current frequency is twice the frequency of each phase. Basically, you double the effective switching frequency of the power supply with no impact on its efficiency.
The chart below shows the efficiency of this power supply versus load current. There are three domains to the efficiency curve. The lower domain efficiency is limited by the overhead losses of control and gate drive. As the current is increased, these losses become less dominant, and switching losses become more significant, At higher current levels, the efficiency falls due to increased conduction losses in the FETs and inductors. The efficiency peak could be moved to the right with lower resistance parts.
Boost controllers have a limited conversion ratio set by minimum off time of the controller and the operating frequency. It is important to be mindful of this limit. Violating it will cause pulse skipping, increased output ripple with lower frequency content, and erratic power supply operation. Interleaving boost power stages provides a method of increasing the effective switching frequency. Very high ratios are possible, and this article has provided a design example with a ratio of 6:1.
For more information about this and other power solutions, click here.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
