Your Pack's Ah Rating Is Only Real at a Stated Discharge Rate

An amp-hour rating is a conditional value, and it is only true at a stated discharge rate, a stated temperature, down to a stated end-of-discharge voltage — and, if you intend to warrant it, at a stated point in the pack's life. The same discipline runs through the whole electrical performance section. Continuous and peak C-rate limits are two separate numbers, and the peak is meaningless without a duration attached to it. A cycle-life claim is only comparable between suppliers when depth of discharge, charge rate, test temperature, and the end-of-life capacity threshold are all named. A voltage window is incomplete without minimum, nominal, and maximum values plus the accuracy the charger is expected to hold. Every one of those qualifiers belongs on the same line as the value it modifies, because that is the only form a supplier can size cells against and quote from without adding margin nobody asked for.

I have been the guy at the table with a spec I wrote myself, watching a supplier's application engineer read it slowly and then ask me a question I could not answer. My line said "Capacity: 10 Ah." His question was one word: at what rate. It wasn't a trap. He needed it, because cell selection and interconnect sizing both came out of that single line, and I had handed him a number with no conditions on it. So he did what any supplier does with a number like that — he picked the interpretation that covered him, and the pack that came back was heavier and more expensive than the product needed. Nothing was wrong with the cells. The spec was short a phrase.

Everything below is how I write the electrical section now: four parameters, and the conditions that have to travel with each one.

Capacity is a family of numbers wearing one label

A single amp-hour figure is a summary of a curve, and that curve moves with the conditions. Ask for capacity at a gentle rate and at room temperature and you get something close to the nameplate. Ask for the same amp-hours at the rate your product actually pulls, or at the cold end of your operating range, and you get less. I don't need to explain that at the cell level to know it happens — I have watched the same pack report different numbers across a table of discharge conditions, and so has every supplier you will talk to.

So the capacity line in my spec now reads as one sentence with three conditions in it: the discharge rate in amps or in C, the ambient temperature, and the end-of-discharge voltage. The cutoff matters as much as the rate. A pack taken down to a lower cutoff gives up more amp-hours, and if that is not what your product's low-voltage shutdown actually does, the number on the sheet is fiction.

Then there is the question nobody writes down: warranted at beginning of life, or at a stated cycle count. Those two produce different designs and different prices. If your product needs its rated runtime in year three, say so, because a supplier who reads an unqualified amp-hour number assumes the easiest possible reading and quotes the smallest pack that meets it. The gap shows up two years into the program, in the return rate.

Peak needs a stopwatch, not an adjective

Continuous discharge and peak discharge are separate spec lines, and the peak line has a clock on it. A ten-second peak and a sixty-second peak are not the same requirement, and a supplier reading "peak 3C" with no duration will size for whatever duration their last program used. Write the duration in seconds. Write how often the peak can repeat, if it repeats — twice a minute is a different thermal and interconnect problem than once an hour. Write the temperature the peak is rated at, and the state-of-charge window it applies over, because a peak that is only available above 50% state of charge is a peak you cannot count on at the end of a shift.

If your pack sees charge pulses — regenerative braking, a rapid top-off between shifts — say that too. Charge peaks do not appear on a discharge C-rate line, and no supplier will assume them for you.

Cycle life is a test report, not a promise

Two suppliers can both quote "2,000 cycles" and be describing two different products. Before those numbers are comparable, four conditions have to be named: the depth of discharge used per cycle, the charge rate, the test temperature, and the end-of-life threshold — 80% of rated capacity is the common one, 70% is not the same claim. Add a fifth if you want to be precise about it: whether the test ran at cell level or on a complete pack with the BMS and the thermal path in the loop. Pack-level is the number your customer actually experiences.

The useful habit here is to read a cycle-life claim backwards. When a supplier says 2,000 cycles, I ask which four conditions produced it, and the answer tells me whether the price difference between two quotes is a bargain or a different product wearing the same label.

The voltage window, and the sag limit that saves rework

Minimum, nominal, and maximum pack voltage, all three, all stated. Nominal is the one people argue about because it is a label rather than a measurement, so I state it as the defined reference voltage and let minimum and maximum do the real work. Then I add the line most customers forget: the accuracy the charger is expected to hold. If the pack maximum is a hard limit and the charger is allowed plus or minus 1%, the supplier needs to know that in order to leave headroom in the design. Say nothing, and they will assume the charger is perfect.

The one I push hardest for is a voltage-sag or internal-resistance limit measured under a defined current pulse. Something like: maximum sag of X volts under a ten-second pulse at the stated peak current, at a stated state of charge, at a stated temperature. That line does two things. It tells the supplier what the interconnect — busbars, tabs, connector, weld quality — actually has to deliver, so they can size it instead of over-building it. And it gives you something measurable to reject against at first article, which a vague "low resistance" claim never does. In my experience this is the most commonly missing line in a customer's electrical section and the one suppliers are happiest to see, because it turns a guess into a requirement.

What it looks like on the page

These go into a four-column table, not a paragraph of prose: parameter, value, condition, and warranted-at. The condition column is the one that matters, and it never says "standard," "typical," or "normal conditions." It says numbers. The values below are placeholders to show the shape of the thing; yours come out of your product's duty cycle.

Then the part that saves the most time: when a quote comes back, I check that the response echoes the conditions, not just the values. A response that says "up to 10 Ah" is not an answer — it is a range, and you cannot design to a range or inspect against one. Send it back and ask for the condition. Suppliers who can turn that around quickly are usually the ones who can answer the harder questions later in the program.

I will admit the tradeoff, because it is real. Over-specifying conditions your product will never see costs money — you are buying margin you do not need, and a stricter condition can push a supplier toward a more expensive cell or a heavier interconnect. The judgment call is to write the conditions your duty cycle actually produces, not the worst case you can imagine. Under-specifying is still the more expensive mistake, but only because the bill arrives later and in a different budget line.

So when you re-read your electrical section, take each value one at a time and ask four questions of it: at what rate, at what temperature, down to what voltage, and at what point in the pack's life. If a number cannot answer all four, it is not a specification yet. It is a wish with units attached.