Build an OEM Battery Pack Power and Energy Budget That Survives the Real Duty Cycle

Start with an electrical boundary and a mission definition

Before asking how many amp-hours to buy, decide what the battery must complete. A mission could be an inspection shift, a sequence of actuator movements, or a defined operating period followed by safe shutdown. Give it a start condition and an end condition. State whether the equipment may enter an energy-saving mode, whether accessories remain energized, and what operation must still be possible after the low-battery warning. Otherwise, two teams can produce different budgets while both believe they have specified the same runtime.

Choose one power boundary for the main calculation. Power measured at the battery connector already includes the downstream converter and equipment losses present during that measurement. Power measured on a regulated electronics rail does not. Record the connector, measurement location and which auxiliaries were connected. Include the battery management system's own consumption separately only when it lies outside the measured boundary. This prevents both optimistic omissions and the common mistake of adding the same converter loss twice.

Also state the battery's starting state: available charge, temperature, age assumption and preceding storage period. A promise to operate for four hours from a fully charged new sample is different from four hours after a week on standby near the end of the agreed service interval. The latter requires additional evidence rather than a larger unexplained safety percentage.

Describe normal operation as modes rather than one average number

Break the mission into modes whose durations and loads can be defended. Separate active work, processing, idle, communications and shutdown. For repeating equipment, capture how many events occur per hour and their possible overlap. An average meter reading is useful for total demand, but it can hide a startup pulse or a long interval in which two large loads run together. Record the timing relationship, not just each component's nameplate rating.

Use measured input power when a representative host is available. When it is not, identify estimates and give their source, such as a converter efficiency curve or a subsystem requirement. Keep uncertain quantities visible so later measurements replace them. The table is an original planning example for a hypothetical four-hour instrument; it is not a customer case or a product rating. All loads are assumed to be measured at the regulated equipment boundary.

Translate mission energy into a transparent planning requirement

For each mode, energy in watt-hours equals power in watts multiplied by time in hours. Add the mode energies only after checking that their durations describe the same mission. If a radio operates during active measurement, its demand must either be included in that mode's measured power or added as a simultaneous load. Adding a separate hour of radio operation to an already complete four-hour timetable would change the mission rather than refine it.

Assume, solely for this example, that the conversion stage is 90 percent efficient across the relevant profile. The estimated battery-terminal requirement is 94 divided by 0.90, or about 104.4 Wh. Suppose the project then reserves an additional 10 Wh at the same boundary for controlled shutdown and a specified communications task. The planning demand becomes about 114.4 Wh. These are arithmetic assumptions, not efficiency or reserve recommendations for every application; a measured efficiency map is better than a single value.

If the supplier and equipment owner agree that only 80 percent of a reference energy figure is available under the chosen combined service conditions, the corresponding reference requirement is 114.4 divided by 0.80, or about 143 Wh. Do not interpret this illustrative factor as a universal discharge limit. Its purpose is to show the direction of the calculation. The factor needs a documented basis and must identify whether it already includes temperature, aging, cutoff and reserve.

Avoid stacking arbitrary derating factors

Cold operation, aging and the equipment's minimum input voltage can all reduce usable energy, but they do not automatically create independent percentage losses. A supplier's discharge result at a stated temperature and cutoff may already include two of those effects. Applying separate generic penalties to the same result can double-count them. Conversely, using a room-temperature cell capacity while ignoring the installed equipment cutoff can overstate mission energy.

Keep a short assumption register alongside the budget. For each adjustment, name the physical effect, the supporting data and the boundary to which it applies. Distinguish an engineering uncertainty allowance from an actual measured reduction. If only preliminary cell data exist, label the output a screening estimate and plan a finished-pack confirmation. The purchasing decision should be based on a reproducible condition set rather than a surprisingly precise runtime calculated from incomplete information.

A nominal watt-hour label is useful for comparing broad product families. It does not describe the entire voltage-time curve delivered to your host. Ask for energy under the proposed discharge profile and equipment endpoint, including the reason the run ended. That connects the commercial reference rating to the mission requirement.

Check power and current independently of total energy

A brief high-power event may consume little energy and still prevent the equipment from operating. In an illustrative case, a 120 W demand lasting two seconds consumes only about 0.067 Wh at that measurement boundary. Yet the pack, protective electronics, wiring and connector must support the associated current without an unacceptable voltage drop or protective interruption. An energy spreadsheet alone will barely reveal this event.

For regulated equipment, estimate current across the permitted battery input range rather than only at nominal voltage. At a fixed input power, lower terminal voltage means higher current. Then replace the estimate with synchronized current and voltage captures under representative conditions. Keysight's electronic-load documentation distinguishes constant-current and constant-power operation; choosing the right load behavior matters when reproducing a host's demand.

Write a separate transient envelope covering amplitude, duration, repetition interval and the operating state in which the event can occur. Include the minimum acceptable voltage at the host connector and the required response if the event cannot be completed. Do not request a protection bypass to make an underspecified peak disappear.

Carry the budget through the installed current path

The relevant voltage is the voltage that reaches the equipment, not only the voltage at an unloaded pack terminal. The approved harness, connector pair, protection path and any service disconnect form part of the installed electrical path. Their effects become more visible during high current. A demonstration using short laboratory cables should therefore not be treated as proof for a different production harness.

Document the installation revision used for validation, including cable length and connector identity. Ask the responsible engineers to assess voltage drop and heating under the continuous and pulsed profile. This is a system review, not a suggestion to replace protective components or choose wire from an article. If host-side capacitance or a startup controller changes, the battery demand may change even when average operating power stays similar.

Separating these boundaries also helps diagnose a failure. A stable pack output with a collapsing host input suggests a different investigation from a pack-side interruption accompanied by a logged protection event. Preserve both traces and the event timing so the supplier can evaluate the actual mechanism.

Turn the spreadsheet into a supplier evidence package

Send the supplier a versioned mission profile, the input-voltage limits, transient captures or estimates, installation drawing and an explicit acceptance endpoint. Request a response that identifies assumptions and deviations. A statement that a pack is suitable is much less useful than a response explaining which requirements are covered by existing evidence and which require a project-specific sample evaluation.

The lithium-ion references linked below are starting points for discussing voltage class, capacity and packaging. They are not interchangeable approvals for the hypothetical mission. Ask for the current approved specification and a matched charger review before selecting a reference. The existing custom-pack specification guide covers the wider RFQ; this article adds a power-and-energy worksheet that can become an attachment to it.

1. Freeze the mission and measurement boundary before comparing quotes.
2. Separate normal energy demand from the transient current envelope.
3. Identify which losses and service conditions are measured and which remain assumptions.
4. Agree the sample profile, stopping conditions and required raw traces with the supplier.
5. Approve the pack and host installation together, preserving revisions and unresolved deviations.

Questions that reveal a weak runtime claim