Single-Stage vs Two-Stage Architecture: What the Choice Changes
Every isolated LED driver must do two jobs: draw current from the mains with a clean power factor, and deliver a regulated output to the LEDs. The architecture question is whether one converter does both or whether two converters are placed in series. A single-stage design merges the power factor correction and the output regulation into one switching cell, while a two-stage design uses a dedicated power factor stage feeding a separate DC-DC converter. The choice affects ripple, hold-up time, efficiency, cost and the behavior of the driver when dimmed, and it is one of the first decisions a designer makes.
In a single-stage design the power factor cell also regulates the output, so the bus capacitor is small and the control loop carries the whole burden. The advantage is component count and cost, and for low-power products the efficiency can be competitive because energy is converted once. The disadvantage appears at the output: the bus carries the 100 or 120 Hz envelope of the rectified mains, and suppressing it requires a large capacitor or a very fast loop. Hold-up time is also short, because there is little stored energy to carry the output through a mains dip.
A two-stage design separates the problems. The power factor stage produces a regulated high-voltage bus, typically 380 to 400 volts, with a power factor above 0.95 and a bus that is largely free of the line-frequency envelope. The second stage, usually a flyback or an LLC converter, turns that bus into the 24 volt output with its own control loop, and it can be optimized for ripple, transient response and isolation without also satisfying the harmonic limits. The cost is more components, more losses and more board space.
For a dimmable product the two-stage arrangement has a practical advantage. Because the output stage sees a stiff bus, the output ripple stays low across the dimming range, and the control loop can be tuned for a clean transient response at any output level. In a single-stage design the loop gain changes with the operating point, and ripple that was acceptable at full output can become a visible modulation at low output, which is exactly where flicker complaints originate.
Efficiency follows the power level. At low output, typically below thirty watts, a single-stage design can beat a two-stage one because the second conversion always costs something. At higher power the power factor stage becomes mandatory for harmonic compliance, and adding a second stage improves the output quality enough to justify the loss. The crossover depends on the topology and the switching devices, and it is a design decision rather than a universal rule.
The PV100 is a 24 volt constant-voltage driver in a plastic housing, aimed at the low-to-mid power range where a compact, low-cost architecture matters. The architecture behind a product like this determines how it behaves when dimmed, how much hold-up it offers and how the output looks on a camera. When comparing drivers, the useful question is not which architecture is printed in the brochure but how the product performs at the dimming levels the project actually uses.
One more consequence is weight and size. The two-stage design spreads heat across two switching sections, which can reduce the hot spot under the transformer, but it also needs a larger board and a taller housing. In a plastic case with a fixed form factor, the layout determines whether the product meets its case temperature target, and that is a thermal decision made at the same time as the architecture.
Inside the Flyback Transformer: Energy Storage, Leakage and Construction
Lumen Maintenance: L70, L90 and the TM-21 Extrapolation
Related Article