Power budgets: will your battery last the demo?

TL;DR:

• A power budget lists what each part draws, adds it up, and divides your battery's usable energy by the total.

• Convert everything to watts (P = V × I) so parts at different voltages can be added; account for converter losses.

• Include peaks and a margin. Motors can draw several times their running current when starting or stalled.

Your robot works perfectly at your desk. Ten minutes into the demo it slows, then resets. Almost always the cause is the same: nobody added up how much power the robot needs, so the battery was undersized, or the voltage sagged under a load nobody counted. A power budget is the cure, and it needs only addition and a bit of unit care.

Sum the currents, then the energy

Electrical power is P = V × I, in watts when V is in volts and I is in amps. A battery stores energy, usually written as capacity in amp-hours (Ah) or milliamp-hours (mAh). Energy in watt-hours is Wh = V × Ah. Runtime is energy divided by power: hours = Wh / W.

Currents can only be added directly when parts run from the same voltage. A robot usually has a motor supply at battery voltage and a regulated 5 V for logic. Converting everything to watts avoids the confusion.

Worked example: adding up currents for motors, sensors and a computer

The robot runs from a 2-cell lithium polymer (2S LiPo) pack, nominal 7.4 V, capacity 2000 mAh. Its loads, measured or taken from datasheets, are as follows (the numbers are an example).

1. The 5 V loads total 0.75 + 2.5 = 3.25 W. A regulator is not perfect. If a switching regulator has an efficiency of 0.85, the battery has to supply 3.25 / 0.85 ≈ 3.82 W for them.

2. Total from the battery: 8.88 + 3.82 = 12.70 W.

3. Battery energy: 7.4 V × 2.0 Ah = 14.8 Wh.

4. Do not use all of it. Take 0.8 of the capacity as usable, since cells should not be run flat: 14.8 × 0.8 = 11.84 Wh.

5. Runtime: 11.84 / 12.70 ≈ 0.93 h, about 56 minutes.

If the demo is 30 minutes, the margin is 56 / 30 ≈ 1.9 times. That is comfortable. If the demo were 60 minutes, the budget says no, even before the next problem.

Peaks and sag

Averages hide peaks. A motor draws far more current when it starts or when its wheel is blocked than when it runs. If each motor can draw 2 A at stall, then two stalled motors draw 4 A. The pack must be able to deliver that without its voltage collapsing, and the wires, connectors and motor driver must be rated for it. Battery packs have a maximum continuous discharge, often given as a C rating: a 2000 mAh pack rated at 10C could supply 10 × 2.0 = 20 A. Check the actual rating from the manufacturer.

A battery also has internal resistance. Under a heavy load the terminal voltage sags, V = V_open − I × R_internal. If a pack with 0.1 Ω internal resistance supplies 4 A, it sags by 4 × 0.1 = 0.4 V. A computer's regulator that cannot cope with that dip may reset, which looks like a random crash. Fix it with a bigger pack, a lower-current design, or a separate battery for the logic.

Margins

• Plan for the worst realistic case, not the average: slopes, carpet and pushing against something draw more.

• Capacity drops as batteries age and in the cold.

• Leave at least a third of the budget as spare. In our example, 56 minutes for a 30 minute demo meets that.

• Use a battery monitor or low-voltage alarm so the robot stops itself before a pack is over-discharged.

LiPo packs: Use a charger made for the pack's cell count and chemistry, charge on a non-flammable surface where you can see it, never leave it charging unattended, and never charge a swollen, punctured or hot pack. Do not discharge below the minimum voltage in the pack's documentation. Fuse the battery lead, and do not short the terminals. Store and carry packs in a fire-resistant bag. Never use mains power in the robot.

Where it goes wrong

• Mixed units. Adding amps at 5 V to amps at 7.4 V gives nonsense. Convert to watts.

• Datasheet maximums used as typical. Use measured values where you can.

• Forgotten parts. LEDs, fans, radios, servos holding a position and a camera all draw current.

• Marketing capacity. Read the real figure and check it with a discharge test.

• Skipping the regulator loss. Linear regulators waste the voltage difference as heat: dropping 7.4 V to 5 V at 0.65 A wastes (7.4 − 5) × 0.65 ≈ 1.56 W.

How real robots use it

Teams keep a power table next to the wiring diagram and update it with every new part. They log real battery voltage and current during tests to compare against the table. The same arithmetic sizes the wiring, and what you learn when you build, as in soldering basics for robot builders, keeps connections sound at those currents. Data wires between parts are covered in I2C, SPI and UART.

Try it yourself

Pick a battery-powered gadget with a stated capacity and a way to measure current, such as a USB power meter. Measure its draw in idle and in use, compute the expected runtime from the capacity, and then run it flat to compare. The difference between the two numbers is your margin.

Frequently asked questions

Can I just look at the mAh rating?

Only when comparing packs of the same voltage. A 2000 mAh pack at 7.4 V holds twice the energy of a 2000 mAh pack at 3.7 V, so energy in watt-hours is the fair comparison.

Why measure instead of trusting datasheets?

Datasheet figures are for set conditions. Your wiring, load and temperature change the numbers, and a measurement tells you what your robot really does.