The step from laboratory coin-cell studies to full energy storage system qualification has turned battery aging test into the longest and most expensive stage of any ESS development programme. Cells, modules and packs must be cycled thousands of times under precisely controlled current, temperature and depth of discharge before capacity fade, resistance growth and thermal behaviour can be trusted. Within that test chain the discharge load bank is no longer a passive dump resistor. It is the instrument that decides how faithfully a cycle profile is reproduced, how much energy the laboratory has to buy and throw away, and how many channels one test hall can operate at the same time.
This article looks at what a real energy storage aging test demands from a discharge load bank, then presents the four models in our portfolio that between them cover the widest part of that work — from a 100 A module bench to a 1500 V pack-level stand with full energy recovery.
Why discharge, not charge, is the hard half of the cycle
Charging a battery is a controlled-source problem: a DC source pushes current into the pack and the energy goes somewhere useful. Discharging is a controlled-sink problem, and the energy has to go somewhere physical. In a conventional resistive load bank it becomes heat inside the cabinet, which means the thermal design, the airflow and the room cooling all become part of the test's operating cost. In a regenerative load bank it is converted back to AC and returned to the grid, which changes the economics of a multi-thousand-cycle campaign completely.
Aging test also differs from a simple capacity check in three ways that shape the equipment choice. First, duration: a continuous cycling campaign may run 6000 hours or more, so duty cycle and derating matter more than peak specification. Second, profile fidelity: capacity fade and DC resistance growth are computed from the difference between charge and discharge curves, so current accuracy and the sampling rate at which the load bank can follow a dynamic profile directly determine the resolution of the degradation model. Third, safety: an aged cell is a more dangerous cell, and the load bank must fail in a way that protects the pack, the chamber and the operator.
What to specify before you compare models
Six parameters decide almost everything. The voltage window must span from the fully charged pack voltage down to the test cut-off — and if you intend to characterise to 0 V for transport-safety work, the load bank must be able to hold regulation at very low voltage, which many resistive designs cannot. The current range must cover both the 0.5 C continuous aging profile and the 1 C or 2 C pulses used in HPPC and DCIR measurement, and the pulse capability is usually what limits a design rather than the continuous rating. Regulation mode matters: constant current, constant power, constant resistance and constant voltage must all be available, because IEC 62619, IEC 63056, UL 1973 and GB/T 36276 test sequences each call for a different one. Accuracy of ±0.2 % to ±0.5 % of full scale is the practical floor for degradation work. Cooling and duty class determine whether the bank survives a 24/7 campaign, and the communication interface — CAN, RS485/Modbus RTU, Modbus TCP or Ethernet with a documented SCPI-style command set — determines whether it will integrate with your existing cycler software or force you to write a new test executive.
1. LB-R Series air-cooled resistive DC load bank
Our workhorse for conventional aging benches. One cabinet covers 0–1000 V and 0–1000 A in four current ranges, with 100 kW to 500 kW of dissipation per enclosure and forced-air cooling sized for a 50 °C ambient so that it can be installed beside a chamber rather than inside a climate-controlled room. Constant current, constant power, constant resistance and constant voltage modes are all available, with ±0.3 % of full-scale current accuracy and a 5 kHz control loop that follows stepped profiles without overshoot. Over-voltage, over-current, over-temperature and reverse-polarity protection are hardware-implemented, not software-limited, and the DC contactor opens on loss of communication rather than holding the last setpoint. RS485/Modbus RTU and CAN are standard. It is the lowest-cost way to run a large number of parallel aging channels, and if your duty is intermittent or your electricity is cheap, it remains the rational choice.
2. LB-RG Series regenerative discharge load bank
For continuous cycling, the energy question dominates. A 500 kW resistive bank running a 70 % duty profile converts roughly 8.4 MWh into heat every day, and the HVAC plant then spends more energy removing it. The LB-RG returns that discharge energy to the grid instead, at 95 % conversion efficiency with THDi below 3 % and unity power factor, so the discharge half of the cycle is largely self-funding. It spans 0–1500 V and up to ±1000 A per cabinet, from 250 kW to 1 MW, and — importantly for aging work — it is bidirectional, so the same unit can charge and discharge through a single connection and a single calibration. That single-connection topology removes a class of cabling errors and contactor sequencing faults that appear in two-quadrant benches over long campaigns. Grid-feedback protection, anti-islanding and a documented interface to the facility's power management system are built in, because a regenerative bank is a grid-tied inverter and has to be treated as one.
3. LB-HV Series high-voltage pack-level load bank
Pack-level aging of 1000 V and 1500 V racks is a different problem from module testing: currents are lower but the stack voltage, the stored energy and the consequences of a fault are much higher. The LB-HV covers 200–1500 V and 0–600 A with 300 kW of dissipation, designed around the safety architecture that pack work requires — live insulation monitoring on both poles, arc-flash detection, a pre-charge circuit, redundant contactor feedback and an interlock loop that can be wired to the chamber door. It carries a dedicated CAN port that talks directly to the pack BMS, so cell voltages, module temperatures and balancing states are logged on the same timebase as the load bank's current and voltage traces. That single timebase is what makes pack-level degradation analysis possible: without it, you cannot separate a weak module from a weak interconnect. Constant power mode is the default for rack-level aging because it emulates the DC bus behaviour of a real PCS.
4. LB-M Series modular rack-mount load bank
Module and cell-string aging is a channel-density problem, not a power problem. The LB-M puts a 0–150 V, 0–300 A, 6 kW or 12 kW channel into a 4U 19-inch enclosure, and up to eight modules can be paralleled for higher current or operated independently as eight separate test channels from one controller. Because each channel has its own current transducer and its own calibration certificate, a bench built from LB-M modules can run a 96-channel module aging matrix in a single rack without the cross-channel drift that appears when channels share a shunt. Front-panel isolation and touch-safe terminals make it practical to keep the rack in an open laboratory rather than a guarded test cell. This is the model most customers choose when the aging programme is about statistical life distribution across many small samples rather than about a handful of large packs.
Choosing between them
A simple rule works well in practice. If your annual discharge energy is below roughly 500 MWh and your benches run intermittently, the LB-R pays for itself and its simplicity is an asset. Above that threshold, or wherever cycling runs continuously, the LB-RG's recovery of discharge energy typically returns its price premium within two to three years of continuous operation. Pack-level work at 1000 V and above should go to the LB-HV on safety grounds alone, regardless of energy economics. And if what you need is hundreds of independent low-voltage channels rather than high power on a few, the LB-M is the only one of the four designed for that shape of test.
Two further points are worth raising in a selection review. First, calibration: ask for the interval, the traceability chain and whether the unit can be calibrated in place, because a load bank out of calibration invalidates months of accumulated aging data. Second, mechanical integration: an aged prismatic cell swells, and if you are correlating electrical degradation with mechanical strain, the bench should carry load cells under the compression fixture measuring swelling force alongside the electrical channel. We supply that instrumentation as an option on all four models, sampled on the same timebase as the current and voltage data, because the correlation between swelling force and capacity fade is one of the more useful early indicators of a cell approaching end of life.
Frequently asked questions
Q1: Can I build an aging bench from contactor-switched resistor grids instead of a proper load bank? You can, and many laboratories start that way. The limits appear quickly: resistor steps give coarse current control, there is no constant-power or constant-resistance regulation, and the switching transients confuse the DCIR calculation. It is acceptable for a rough capacity check, not for a multi-thousand-cycle degradation study.
Q2: How accurate does the load bank need to be for DCIR and capacity fade measurement? Plan for ±0.2 % to ±0.5 % of full scale on current and voltage, with at least 10 samples per second logged on a common timebase. Resistance growth in a healthy cell over the first 500 cycles can be only a few percent, so measurement error of 1 % will mask the trend you are trying to measure.
Q3: Is a regenerative load bank worth the extra cost? It depends on duty. Below about 500 MWh of annual discharge energy, the payback is slow. On a continuously cycling bench, recovering 95 % of discharge energy instead of dissipating it typically recovers the price difference in two to three years, before counting the reduced HVAC load.
Q4: Can these load banks discharge a pack all the way to 0 V? The LB-R and LB-RG regulate down to a few volts and will follow a controlled run-down below that, but deep discharge to 0 V is best handled by a dedicated deep-discharge channel; tell us the cut-off voltage at the quotation stage rather than discovering the limit during commissioning.
Q5: Do the load banks integrate with our existing cycler software? Yes. All four publish a documented command set over Modbus RTU, Modbus TCP, CAN or Ethernet, and can be driven as a slave device from a third-party test executive. Drivers for the common laboratory platforms are available, and a raw register map is supplied for anything else.
