4 Semi-Distributed Architecture

In semi-distributed architecture, multiple gate drivers share power from a single isolated bias power supply device. This architecture enables the cost saving of a separate device for each gate driver, but at the same time, failure in one device can stop the power supply to multiple gate drivers.

Within the semi-distributed architecture, there are two different possibilities. Figure 4-1 shows the first type of semi-distributed architecture. In this architecture, high-side gate drivers are supplied using three separate isolated bias power supply devices, for example using three integrated DC/DC modules. On the other side, all low-side gate drivers are supplied using a single isolated bias supply device. Some of the possible designs are LLC resonant (UCC25800-Q1), PSR-flyback (LM518x-Q1, LM2518x-Q1, LM515x-Q1, LM34xxx-Q1) and push-pull (SN650x-Q1) topology-based devices, which can deliver the required power to multiple gate drivers making these devices an excellent choice for semi-distributed architectures. While supplying multiple gate drivers using one device, the secondary side of the transformer (Pre:Sec = 1:3) has multiple windings, as an example - three secondary windings for three gate drivers.

Figure 4-1 Semi-Distributed Architecture at Low Side Only

Figure 4-2 shows the second type of semi-distributed architecture where both high-side and low-side gate drivers are supplied using two single isolated bias supply circuits. In each of these circuits (high side and low side), there is one transformer with multiple secondary windings. This architecture requires only two isolated bias power devices, which can be a cost-effective option. Unlike the low side, which has common ground, the high-side voltage reference is floating. Therefore, due to cross-coupling and so forth, the stability of a multi-winding transformer at the high side is challenging.

Figure 4-2 Semi-Distributed Architecture at High Side and Low Side