Bidirectional DC Power Supply, Explained

What makes a supply bidirectional

A conventional DC power supply only pushes current outward: it sources energy into a load. A bidirectional supply adds the opposite capability — it can also pull current back from the device under test, behaving as an electronic load. Because both functions live in the same output stage, the instrument crosses from sourcing to sinking continuously, with no relay click and no dead band at zero current.

This matters whenever the device under test stores or returns energy. A battery being cycled charges (the supply sources) and then discharges (the supply sinks). A solar inverter pushes power back toward its DC port during certain tests. A motor drive on a dynamometer regenerates during braking. In each case a one-directional supply forces you to bolt on a separate electronic load and stitch the two together; a bidirectional unit does both in one control loop.

On platforms such as the N35500 the same channel covers 0–2250 V and up to 42 kW in a 3U chassis, with channels paralleled in master/master configuration to reach megawatt scale. Ultra Power Systems builds the N35500 platform referenced throughout this site.

Regenerative, not dissipative

When a bidirectional supply sinks power, it has to do something with the energy it absorbs. A traditional electronic load burns it as heat in banks of transistors and resistors, then expels that heat with fans or chilled water. A regenerative supply instead converts the absorbed DC back into clean AC and returns it to the building mains.

• Efficiency: recovery reaches up to 93%, so most of the energy you would otherwise pay to cool is recycled.
• Thermal load: far less waste heat means smaller HVAC, lower ambient rise, and denser racks.
• Running cost: on long endurance cycles the electricity saved is significant, especially at tens of kilowatts.

The trade-off is added power-electronics complexity and grid-quality requirements, covered in the regenerative-energy-recovery article.

Programmable control and operating modes

"Programmable" means the output is set and read over a digital bus rather than by front-panel knobs. These supplies expose constant-voltage (CV), constant-current (CC), constant-power (CP) and constant-resistance (CR) modes, selectable and switchable under test-sequence control. Typical figures include 0.02% of full-scale setting accuracy and command-to-output response of 5 ms or better.

Interfaces span LAN, RS232, RS485 and CAN, speaking SCPI, Modbus-RTU and CANopen so the unit drops into existing automated test equipment. On top of the raw modes, firmware can emulate complex sources: a battery with realistic state-of-charge behaviour, or a photovoltaic array following an I-V curve.

Where to go next

Each subsystem below has its own deep-dive:

• How a bidirectional DC power supply works — the source/sink control loop in detail.
• Regenerative energy recovery — how energy returns to the grid and what it saves.
• Four-quadrant operation — the voltage/current quadrants and what each enables.
• Battery emulation and solar array simulation — emulating real sources.

For platform specifics or a quote you can reach the manufacturer at Ultra Power Systems.

Frequently asked questions