Difference between synchronous and conventional dc-dc boost converter?

Question

What is the difference between synchronous and conventional dc-dc boost converter? which one is best and why?

Answer 1

Synchronous rectification refers to the practice of using an active element such as a MOSFET switched at the appropriate instants in place of a diode. Here's a simplified example of a boost converter:

^{simulate this circuit – Schematic created using CircuitLab}

Not shown here is the logic to switch the transistors. In the synchronous case it's important that M2 and M3 are never both on at the same time, as this would short C2 to ground, possibly damaging the transistors, and severely decreasing efficiency.

The advantage is the conduction losses through a MOSFET (M3) may be less than through a diode (D1). Consider the voltage across D1 will be about 0.6V, or perhaps 0.2V if current is low enough that a Schottky diode is feasible. However, a properly selected MOSFET in many circumstances may have a V_DS even lower, and thus lower losses.

The disadvantage is higher complexity and cost.

Answer 2

The distinction is actually about rectification on the output/secondary side of the converter. Synchronous rectification means that a transistor (normally a MOSFET) is used to control the current flow for minimum loss.

Asynchronous rectification refers to the use of one or more passive diodes to control the current flow.

The latter is simpler, and quite effective at low power levels. At higher power levels (to a point), the lower V_DS of a full-on MOSFET results in lower losses.

At very high power levels (high current levels), BJTs, IGBTs or other types of devices may have lower losses than MOSFETs.

Answer 3

The term synchronous rectification (abbreviated sync rect) implies the replacement of a diode by a controlled switch, usually a MOSFET. This transistor can be self-driven, meaning its \$V_{GS}\$ is naturally self-generated by the converter (an auxiliary winding for instance) or it requires an extra driving circuit to generate the appropriate signals. In continuous conduction mode (CCM) where the inductor current never returns to 0 within a switching cycle, the control of the sync rect can be complex to avoid shoot-through (both switches are on during a short period of time). This is even more complicated with isolated designs since the main switch lies in the primary side while the controlled switch is in the (isolated) secondary side: a second signal is usually needed (but not always) to ensure minimal shoot-through.

In discontinuous conduction mode (DCM), the diode naturally turns off (the inductor current cancels withing a switching cycle) and the control of the sync rect is greatly simplified. Some converters such as those operated in boundary mode (at the boundary between CCM and DCM), lends themselves well to sync rect control because they can't enter CCM.

Sync rect controllers observe the drain-source voltage of the sync rect MOSFET. As its body diode conducts first, the drain-source voltage spontaneously drops and later on, the MOSFET is turned on: turn-on losses are small considering the zero-volt switching (ZVS) action. Drive losses also benefit from this ZVS action as the Miller effect is greatly reduced. When the current decreases in the MOSFET, the voltage tends to go down as well and as it passes a given threshold, it signals the control chip to turn the MOSFET off. This is a difficult phase because you want to extend the conduction duration as long as you can (to benefit from the low \$r_{DS(on)}\$ but, on the other end, you need to avoid shoot-through losses. Finally, parasitic inductance corrupts the signals by extra offset and ringing implying rigorous PCB layout rules (shortest paths, smallest areas etc.) to avoid false turn-off.

The conduction losses of a diode depend on its dynamic resistance \$r_d\$ and its threshold voltage \$V_{T0}\$ (\$\approx\$0.6 V for Si and \$\approx\$0.4 V for a Schottky). The resistive term is sensitive to the squared rms current while the dc source is affected by a the average current: \$P_d=r_dI_{rms}^2+V_{T0}I_{avg}\approx V_fI_{avg}\$ where \$V_f\$ is the diode total drop at a given operating point (extracted from the I-V curves). You thus realize that the diode dissipation is related to the average current flowing in its junction (considering a low ripple current of course).

For a MOSFET, the conduction losses are classically given by \$P_d=r_{DS(on)}I_{rms}^2\$ where the term \$r_{DS(on)}\$ represents the on-resistance at the highest junction temperature (100 °C or so). This where you realize it is sometime hard to beat the classical diodes, especially considering a MOSFET contribution sensitive to the ac ripple of the current while a diode is less. It can be the reason while you often see paralleled MOSFETs to better perform than a simple diode: paralleling MOSFETs help reducing the heat and can avoid a heatsink while there is not much you can do for a diode beside selecting a bigger die or adding more aluminum around it.

Finally, the adoption of sync rect also depends on the diode voltage drop contribution to the overall output. A 0.4-V drop is large for a 3.3-V or 5-V output but does not affect efficiency the same way for a 24-V output for instance. Sync rect will make better sense in the first example unless you face size constraints which impose you to design without heatsink.

Answer 4

Instead of crafting my own wording in a broken English, a simple Google search for {difference between synchronous and conventional dc-dc boost converter} reveals the hit on top, from Rohm Tech Web,

it reads,

As shown in the figure, the difference lies in the fact that whereas in the nonsynchronous rectifying type, the low side switch is composed of a diode, in the synchronous rectifying type, as same as the S1, the switch is a transistor.