ULTRAPOWER — Ultra Power Systems

DC Power Supply for EV Battery Testing: How to Choose the Right Source

Charge-discharge cycling of EV cells, modules and full packs needs a four-quadrant DC source that sinks energy as cleanly as it delivers it. This guide maps your test profile to the right voltage class, power band and bidirectional rating so you specify once and avoid costly under-sizing.

Why EV battery testing demands a bidirectional source

Battery validation is inherently two-directional: you charge the device under test, then discharge it, often thousands of cycles deep. A unidirectional supply paired with a resistive load wastes discharge energy as heat and doubles your equipment footprint. A bidirectional, regenerative DC power supply handles both halves of the cycle in one chassis and returns discharge energy to the grid at up to 93% efficiency, slashing electricity and HVAC cost on long cycling campaigns.

The N35500 platform from Ultra Power Systems covers cell, module and pack work from a single architecture, with constant current, constant voltage, constant power and constant resistance (CC/CV/CP/CR) regulation for accurate emulation of real charge and discharge profiles.

Matching voltage class to your device under test

Pick voltage headroom from the top-of-charge of your highest pack, not its nominal rating. With an output span of 0–2250 V, one platform spans 400 V, 800 V and emerging 900 V+ architectures without paralleling extra cabinets.

Device under testTypical nominalRecommended supply class
Single cell / small module3.2–60 VLow-voltage range, high current
48 V / mild-hybrid module48–96 V0–200 V class
400 V EV pack350–470 V0–600 V class
800 V EV pack650–900 V0–1000 V class
Commercial / bus / aerospace1000–2000 V0–2250 V class

Always leave margin: a supply that tops out exactly at your charge ceiling cannot drive the constant-current taper that completes a CC/CV charge.

Power, current and the fast-transient requirement

Size power from your worst-case C-rate. A 100 kWh pack cycled at 1C needs roughly 100 kW; at 2C, 200 kW. A single N35500 unit delivers up to 42 kW in a 3U chassis, and master/master parallel scales the same hardware into the megawatt range for full-pack and battery-lab work.

  • Accuracy: 0.02% F.S. so coulomb-counting and capacity measurements stay trustworthy.
  • Response: ≤5 ms transient response to follow drive-cycle and pulse profiles without overshoot.
  • Regulation modes: CC/CV/CP/CR to emulate chargers, loads and real-world duty.

BOBC and on-board charger testing

For bidirectional on-board charger (BOBC) and vehicle-to-grid validation, your DC source must act as the battery the charger sees, sinking and sourcing on command while comms exercise the control loop. LAN, RS232, RS485 and CAN interfaces with SCPI, Modbus-RTU and CANopen let you fold the supply into an automated test sequence and drive it from your existing ATE framework. Battery emulation via the NS81000 option (7 built-in libraries plus custom curves) lets the charger see a realistic state-of-charge ramp instead of a flat rail.

Ready to scope a system? Request a quote with your voltage class, peak power and channel count, or email Ultra Power Systems.

Frequently asked questions

Do I need a bidirectional supply or can I use a separate load?
For repetitive cycling, a bidirectional regenerative supply is almost always the better buy: it cuts equipment count, returns discharge energy to the grid at up to 93% efficiency, and gives seamless source-to-sink transitions that a separate supply-plus-load pair cannot match across the zero-current crossover.
What voltage class should I buy for an 800 V pack?
Choose a 0–1000 V class so you have headroom above the ~900 V top-of-charge. The N35500 platform also covers this range alongside 400 V packs, so one investment serves multiple programs.
Can the same supply test cells, modules and full packs?
Yes. The platform spans 0–2250 V and scales from a 42 kW 3U unit up to megawatt class via master/master paralleling, so cell cycling, module validation and full-pack work share one architecture and software interface.
How accurate is the measurement for capacity testing?
Output and readback accuracy is 0.02% of full scale, which keeps coulomb-counting, capacity and efficiency calculations within the tolerance most EV battery test standards require.

Get a configured quote

Tell us your voltage, power, application and timeline and get a configured quote from Ultra Power Systems.

Get a Quote Contact the team