How Discharge Cutoff Voltage Affects Battery Capacity and Cycle Life

Take two cells both sold as 4.0 Ah. One delivers close to its rating at the current your application actually draws; the other falls short — and neither datasheet is lying. That gap is illustrative, not measured, but the mechanism behind it is well documented: the difference usually sits in the discharge cutoff voltage and the test conditions behind the rating.

Most buyers treat the cutoff as a simple stop-discharge threshold and compare only the headline Ah and Wh figures. It’s closer to the operating boundary of the electrochemical system: it sets the depth of discharge and how much lithium is de-intercalated from the anode, and those conditions split a single cell model into very different usable capacity, energy output and fade behaviour. That is why rated capacity, measured capacity and cycle life rarely line up on a datasheet alone.

Rated capacity is a declared value, not a measured constant. IEC 61960-3:2017 defines it as the charge a cell delivers under specified conditions — a set temperature, a set current and a discharge to a set end-of-discharge voltage. Change any one of those, and the same cell reports a different capacity. This article separates the three voltage limits that get conflated in procurement conversations, shows how the cutoff moves usable capacity and energy, connects it to cycle life, and ends with a comparison grid you can put in front of a supplier.

Cutoff Voltage and Rated Capacity

A cutoff voltage is the closed-circuit voltage at which discharge is terminated. That’s the whole definition, and it makes the number a per-product figure rather than a chemistry constant. No international standard hands you one mandatory cutoff per chemistry — the standards define test methods, and the authorised limit is supposed to come from the cell maker. Two cells with the same nominal voltage can therefore carry different specified discharge limits. For procurement, the value that matters is the one printed in the exact cell’s specification sheet, at its stated current and temperature.

Three different limits share the name, though. There’s the cell cutoff, the BMS undervoltage protection, and the device low-voltage shutdown. They’re set by different owners and describe different things: a test and usage limit, a programmable protection trip, and a load-side threshold your application controls. So a pack can stop delivering usable power long before the cell reaches its specified limit. The BMS layer often decides how much of the rated capacity your application ever sees, which is why two packs built from identical cells can still behave like different products.

Lower the cutoff and you add charge, but the added energy is smaller than the added amp-hours, because the extra charge arrives at a falling voltage. Amp-hours measure charge — current over time — so lowering the cutoff simply extends how long the cell keeps delivering it. Watt-hours measure energy, which is that charge weighted by voltage, so the same extra amp-hours are worth less near the bottom of the curve. Drop the cutoff and the Ah gain stays fairly linear while the Wh gain keeps shrinking. A datasheet that reports Ah only hides this.

The reason is the shape of the discharge curve. Many lithium-ion cells hold a relatively stable voltage region and then decline more steeply near the end, though the exact shape varies with chemistry, design, current and temperature. The extra amp-hours live in that steeper region, where energy per unit of charge is lowest — which is why the additional capacity has to be read from the specific cell’s discharge curve rather than carried across datasheets.

The rate effect is large enough to change a purchasing decision. Panasonic rates the NCR18650B at a minimum 3,200 mAh, but that figure is declared at a 0.65 A discharge to a 2.5 V cutoff at 20 °C; at 1C (3,250 mA) to the same 2.5 V cutoff, the datasheet’s discharge-rate curve shows a lower delivered capacity (Panasonic NCR18650B datasheet, retrieved 2026). The point is not one private number but the direction: hold the cutoff and raise the current, and the same cell reports less. That is why two figures can both be truthful and still not be comparable.

Curve shape decides how much the extension is worth. Published characterization shows LFP holding a flat plateau before a sharp fall, while NMC declines progressively through the tail; one study used 2.8 V as a common limit and extended full discharges to 2.4 V for LFP and 2.9 V for NMC at C/5, C/2 and 1C (Sigma-Aldrich characterization of advanced Li-ion cathode materials, 2026). On a flat LFP tail the added charge sits at a nearly constant voltage; on a progressive NMC tail it does not.

What to verify before comparing two capacity figures: the cell model, the cutoff voltages compared, the discharge rate, the temperature, and whether the datasheet reports Ah only or also discharge energy in Wh.

Cutoff Voltage and Cycle Life

A cutoff voltage sets how deep each discharge goes, so it shapes cycle count indirectly. But the two are not interchangeable variables, and how much deeper discharge costs you depends on the cell’s discharge curve and the duty cycle — not on a fixed ratio between cutoff voltage and lifetime.

That makes the literature useful as a warning rather than as a lookup table. Take two LFP studies that point the same way but disagree on size.

The first is a measured cycling programme: 80% depth of discharge gave 3,221 cycles, while 20% depth gave 34,957 (Omar et al., Applied Energy 113, 2014, 2014). Those are real cycles for a specific cell and test condition, and they describe a depth-of-discharge relationship rather than a cutoff-to-lifetime one. The end-of-life criterion and test temperature were not retrievable, so treat the ratio as directional.

The second reached a different number by a different method. It attributed fade mainly to energy processed and modelled about 5,300 cycles to 80% capacity even at 95% depth of discharge (Carnegie Mellon CEIC working paper, 2009) — a projection, not a measured lifetime.

Read together, they say something simple: shallow cycling helps, but the magnitude depends on the cell, the rate and what each test counted as failure. There is no universal cutoff-voltage-to-cycle-life conversion, and any supplier implying one is extrapolating.

Cycle-life claims are only comparable when the retention criterion, rate, temperature and termination condition travel with the number.

A current 21700 model shows how specific the figure gets. The Samsung INR21700-50E specifies a 2.5 V discharge cut-off, 4,900 mAh minimum at 0.2C to 2.5 V, and 4,753 mAh minimum at 1C to the same limit. An independent aging dataset for the same cell uses the full 2.5 V to 4.2 V window and states the same 4.9 Ah nominal capacity (Ponce-Alcántara et al., Scientific Data, 2026). Strip the retention target, charge rate, discharge rate or cutoff from that number and it stops meaning anything — which is why those choices sit inside the broader set of UAV battery design requirements that should be fixed before a custom pack is developed.

Pushing the cutoff far lower has a lifetime cost out of proportion to the capacity gain. Severe over-discharge can damage the anode structure and the SEI passivation layer, and under sufficiently extreme conditions copper dissolution from the current collector and later copper deposition can contribute to internal-short-circuit risk. Studies place copper dissolution at roughly −12% SOC and observed internal shorting only after more than 100 mAh was removed from an unbalanced cell (Guo et al., Scientific Reports 6:30248, 2016; NASA NTRS, 2010). Recognise that as a distinct regime, though. It is not the same as operating within the cell manufacturer’s specified discharge limits, so the short version is simple: use the datasheet cutoff and leave deep over-discharge alone.

Where the cutoff actually sits is a trade-off, not a single best value. The conflict is between maximising single-cycle usable capacity and maximising whole-life cycle life, and there is no universal optimum — only one that fits your application. Deeper discharge can raise the capacity delivered in an individual cycle, but lifetime cannot be inferred from the cutoff alone, since it also depends on the cell design, the rates, temperature, rest periods and the end-of-life criterion.

That plays out differently by use case. A long-life application that cares about cost across many cycles may accept less capacity per cycle to keep more headroom. A short-lived device replaced within a few years may trade some margin for runtime. Either way, base the decision on test results for the selected cell under conditions that resemble the real duty cycle, not a number copied across every product.

How to Compare Battery Specifications for Supplier Audits

Most datasheet comparison errors share one root cause: a numeric capacity value is quoted without the conditions that produced it. A rated Ah or Wh figure is the joint result of three inputs — the voltage window, the discharge rate and the ambient temperature — and none of them can be left implicit.

The two that get dropped most often pull in the same downward direction. A higher current raises polarisation and voltage sag, so less lithium is de-intercalated before the cutoff. In the cold, electrolyte viscosity rises and lithium-ion transport slows, so less capacity arrives before the cutoff. How large either effect is depends on the specific cell and discharge condition. That is exactly why the number has to travel with its conditions rather than with a generic percentage. An operating cutoff and a rated-capacity cutoff are also not the same line, so the two figures may sit in different regions of the curve.

The standards show how narrow the tolerance band is. IEC 62660-1:2018, for example, pins room temperature to 25 °C ± 2 K and requires cells to stabilise before each test (IEC 62660-1:2018, 2018). A datasheet that omits temperature and rate has omitted the two conditions that move the number most.

Use the grid below as a set of questions rather than a scorecard. Each row converts a datasheet line into something a supplier can answer with a figure or a document.

Parameter	Question for the supplier	Why it matters
Exact cell model	Which cell model and which manufacturer datasheet support the rating?	Prevents comparisons across different cells or product variants
Discharge cutoff	At what end-of-discharge voltage was capacity measured?	Establishes the voltage boundary for the result
Discharge rate	What current or C-rate was used?	Determines how much capacity is delivered under load
Test temperature	What ambient temperature and tolerance applied?	Makes environmental conditions comparable
Measured energy	Are Wh values or discharge curves available?	Quantifies usable energy, not just charge
Cycle-life criterion	What retention threshold, rate and cutoff were used?	Makes cycle-count claims interpretable
Pack protection	What are the BMS thresholds, delays and recovery settings?	Connects cell data to pack-level operation

One further standard belongs in the conversation as a credibility marker, provided it is not overread. UN 38.3, 6th revised edition, lists eight tests, including forced discharge (T.8), which requires cells in a fully discharged state. It does not test usable capacity (UN Manual of Tests and Criteria, sub-section 38.3, 6th revised edition). It is a transport safety test, not a performance benchmark.

If a supplier cannot supply the rate and temperature behind a capacity figure, the figure is not comparable. When the load is bursty rather than steady, compare the continuous and peak C-rating alongside the capacity figure, since the same cell delivers less charge at a higher rate.

What Cutoff Voltage Means for UAV Battery Selection

For a UAV battery pack, usable energy depends on how the cell discharge limit, the BMS protection settings and the flight-controller failsafe thresholds interact. These limits do not simply reduce to the lowest voltage value. Their effect depends on pack configuration, where voltage is measured, load sag, protection delays and the action each threshold triggers.

A flight controller typically initiates a return-to-launch, landing or warning rather than cutting power. ArduPilot’s failsafe, for example, can trigger on a pack-voltage threshold with a configurable delay and a configurable action, and its documentation’s 10.5 V figure is a parameter example for a particular battery setup rather than a limit that applies to every UAV pack (ArduPilot battery failsafe documentation, retrieved 2026). The value you use depends on the series count, the measurement method, the trigger delay and the action assigned to the threshold. The BMS is the layer that may actually disconnect the pack once a cell reaches its configured protection threshold. Selection therefore means coordinating these limits around the aircraft’s operating profile, not treating the cell datasheet cutoff as the only relevant boundary.

Where usable flight time is won or lost is in those interactions. The energy actually available during a mission depends on the aircraft’s load profile, voltage sag, failsafe logic and the BMS protection settings. As a pack ages and its internal resistance rises, voltage sag under load may become more pronounced, potentially causing a low-voltage failsafe to trigger earlier. The documentation also notes that the failsafe can use raw or sag-corrected voltage, which matters under load: a pack that sags below threshold during a climb may trigger the failsafe while its rested voltage is higher. Evaluating these interactions is more informative than comparing the thresholds alone.

In practice, that means matching three numbers and ordering them for the series count: the cell’s specified cutoff, the BMS undervoltage trip and recovery, and the autopilot threshold. Review them together under the expected load. The cell-level discharge curves, the BMS undervoltage and recovery settings and the aircraft’s low-voltage response will show whether the pack can deliver the required energy while keeping a sensible margin before any protective action fires.

Cycle-life expectations follow the same logic: reading a cycle count against its retention criterion and rate is the same discipline behind any lithium battery cycle-life test report.

In a UAV pack, usable energy is set by how the cell limit, BMS thresholds and autopilot failsafe interact, not by comparing three voltage numbers.

A Rated Capacity Is Only as Meaningful as Its Test Conditions

The four chapters above point to one engineering judgment: a battery’s rated capacity is only meaningful when the discharge conditions are defined, and its usable energy depends on how those conditions interact with the application. A headline amp-hour figure without a cutoff, a rate and a temperature is not a specification you can design around, and a comparison built on those figures alone will not survive contact with a real duty cycle.

That judgment has three practical consequences:

• Capacity and energy diverge as the cutoff drops, so ask for Wh or a discharge curve rather than Ah alone.

• Cycle life is conditional, so read any cycle count with its retention criterion, rate, temperature and cutoff attached.

• The cell limit, BMS protection and device cutoff must be ordered around your operating profile rather than reduced to a single number.

The next step is a specification conversation rather than a datasheet comparison. Share your target voltage, capacity, continuous and peak current, operating temperature and equipment interface requirements, and we can map them onto a pack design that keeps the cell limit, the BMS thresholds and your device cutoff in the right order. Get in touch with our team to start that conversation.