What Is Discharge Cycle?

Had a call last Tuesday with a startup doing their first handheld device. PM asks me how many times their users can charge the battery. I said it depends. He wanted a number. I told him 500, maybe, under perfect conditions that will never exist in the field.

That's the problem with discharge cycles. The concept is dead simple but applying it to real products gets complicated fast.

Your battery doesn't count plug-ins. It counts capacity throughput. Drain 100% of nameplate capacity and that's one cycle. Do it in one shot or spread it across a week of partial discharges - same result.

The 18650 sitting in your flashlight right now might be on cycle 47 or cycle 212. Unless you've been tracking amp-hours religiously, you have no idea. The cell sure doesn't know. There's no little counter inside. Your BMS estimates it based on coulomb counting, and coulomb counting drifts. I've pulled packs apart where the reported cycle count was off by 30% compared to actual calendar aging.

Samsung SDI published some data back in 2019 showing their 21700 cells hitting 83% capacity retention at cycle 300 under their standard test protocol. Panasonic NCR cells from the same era were more like 78% at 300 cycles. These numbers came from climate chambers at 23°C with CC-CV charging at 0.5C and discharge at 1C to 2.5V cutoff. Your mileage will vary because your use case isn't a climate chamber.

Everyone in this industry talks about depth of discharge like it's some kind of magic lever. Shallow cycles extend life, deep cycles kill cells faster. True enough. But the relationship isn't what most people expect.

I ran a side project in 2021 trying to quantify this for a medical device customer. We took 200 cells from the same production lot, split them into groups, and cycled them at different DoD levels for eight months. The 100% DoD group cratered around cycle 380. The 60% DoD group was still at 88% capacity at the equivalent of 900 full cycles. The 30% DoD group barely showed any degradation at all after a year.

But here's what the datasheets won't tell you. Calendar aging was eating into all the groups at roughly the same rate regardless of cycling. By month ten, even the cells we barely touched had lost 4-5% capacity just sitting there at 50% state of charge in a 25°C room. The cells cycled hard were getting hit from both directions - cycle aging plus calendar aging stacking on top of each other.

So when your customer asks why their two-year-old device doesn't last as long even though they "barely used it" - that's why.

Discharge Cycle

I've got a box of cells in my garage from a golf cart battery pack that spent three summers in a shed in Tucson. Owner swore he only used the cart maybe twice a month. Pack was rated for 2000 cycles. It died at 160. Cells looked fine externally. Cracked one open and the jelly roll had brown spots on the separator. Electrolyte had partially decomposed.

Heat kills lithium cells. Not fast like a dead short. Slow, like leaving butter out on the counter.

The Arrhenius relationship gives you roughly 2x degradation rate for every 10°C above 25°C. But that's an average across chemistries. NMC cells I've tested fell off faster. LFP held up better. The pouch cells we use for wearables are more sensitive than cylindrical cells with the same chemistry because the thermal mass is lower and they spike hotter under load.

Cold discharge is annoying but not destructive. You lose capacity temporarily. It comes back when the cell warms up. I've discharged cells at -15°C that only gave 61% of rated capacity, then let them warm to room temp and got the full 3.2Ah on the next discharge. Cold charging is the killer - lithium plating on the anode. That damage is permanent. Most BMS lock out charging below 0°C for exactly this reason.

Every discharge moves lithium ions from the graphite anode to the cathode. Every charge shoves them back. Do this enough times and things start breaking down.

The SEI layer on the anode thickens. That layer is supposed to be there - it protects the graphite from the electrolyte. But it grows with every cycle, and growth means it's consuming lithium that could otherwise store energy. This is why capacity fades even when nothing dramatic happens.

Cathode particles crack. NMC and NCA cathodes have a crystal structure that expands and contracts as lithium moves in and out. Micro-cracks form. Surface area increases. Side reactions accelerate. I've seen cross-sections of cathode particles from high-cycle cells that look like shattered glass under the microscope compared to fresh cells from the same batch.

The electrolyte breaks down. Trace water causes problems. High voltage accelerates oxidation. You end up with gas generation, increased impedance, and eventually the cell swells or vents.

None of this happens evenly. The cells in your series string that run hotter or started with slightly lower capacity age faster. After a year or two, your "matched" pack isn't matched anymore. The weak cell limits the whole string on discharge and gets overworked trying to keep up. Death spiral from there.

IEC 62660 and the various UL standards give you a baseline. Pass those and you can ship product. But I've watched cells ace certification testing and then show up in field returns six months later with swollen pouches.

The certification tests are designed to be repeatable across labs, not to simulate real use. Nobody uses their phone at exactly 0.5C constant current. Nobody charges their power tool at a perfect CC-CV profile at 25°C.

If you're serious about understanding how your pack will perform, you need to build a test protocol that matches your actual user. Profile your expected discharge current. Include the rest periods. Hit the temperature extremes your product will see. Run enough cells to get statistical confidence - and that means at least 15-20 per test condition, not the 3-5 that most startups try to get away with.

The cells that fail early in life testing almost always had manufacturing defects. Burrs on the electrode, contamination, bad welds on the tabs. Those will show up in the first 50 cycles. The cells that fail late - those are the ones telling you something about your design.

You can calculate expected cycle life if you know your user profile. A phone that gets charged daily, maybe drains to 20% on average before plugging in - that's roughly 0.8 cycles per day. Over a two-year expected life, you're looking at 580 cycles. If your cells are rated for 500 cycles to 80% capacity, you're going to start seeing complaints around month 18.

This is why Apple and Samsung oversize their packs relative to advertised capacity. The "100%" you see on your phone isn't 100% of what the cells can actually hold. They keep a buffer at the top and bottom to reduce stress on the cells. Your phone might have 4500mAh of physical capacity but the software only lets you use 4000mAh of it.

Same idea works for any battery product. You can either hit your cycle life target with better cells - which cost more - or you can hit it by putting in extra capacity and running the cells easier. The second option usually wins on total cost when you factor in warranty reserves.

Discharge Cycle

Discharge cycle is just a way to quantify wear. Like odometer miles on a car, but the relationship between miles and remaining life is squishier. Two packs at 300 cycles can have completely different remaining capacity depending on how they got there.

The cell manufacturers will give you their numbers. Your job is figuring out how those numbers translate to your specific product and your specific users. There's no formula for that. You test, you ship, you watch the field data, and you adjust.