Constant-current constant-voltage (CCCV) charging is the default protocol for lithium-ion batteries across nearly every application from consumer electronics to vehicle packs. It works reliably, it is simple to implement, and it has a well-understood safety record. It is also thermally conservative by construction, meaning that it systematically leaves charge speed on the table in conditions where faster charging would be physically safe.
This conservatism is not a design flaw. CCCV was developed as a general-purpose protocol that must work safely across the full range of ambient temperatures, cell ages, and pack configurations that any given cell might encounter. The margin it leaves is the margin required for that generality. The question is whether a pack engineer who knows the specific cell, the specific pack, and the operating environment actually needs all of that margin. In most cases, the answer is no.
What CCCV actually does thermally
The constant-current phase of a CCCV charge runs at a fixed current regardless of the cell's actual thermal state. The current is typically set at 0.5C to 1C for cycle-life-optimized protocols, or up to 2C for faster-charge specifications on modern cells. In either case, the protocol continues at that current until the cell voltage reaches the upper cutoff, then switches to constant voltage and tapers current until it falls below a low-current threshold.
From a thermal perspective, this fixed-current approach has two inefficiencies. First, at low SoC, the cell's internal resistance is typically higher than at mid-SoC (due to the electrolyte transport resistance and solid-state diffusion dynamics in the electrode material at low lithiation levels). This means Joule heating is actually higher at the start of the CC phase than in the middle. A thermally-aware protocol could charge slightly more slowly at the beginning when the resistance is high, then accelerate through the mid-SoC window where heating per unit charge is lower. CCCV does the opposite: it maintains constant current regardless, so it generates more heat per unit of charge delivered at low SoC than is necessary.
Second, the CV phase is where the longest charging time is often spent in protocols with high CC current. The CV phase at 4.2V with a decaying current profile is slow by design: it is waiting for the concentration gradients in the electrode bulk to equilibrate. The time spent in CV can be reduced if the CC phase is managed to avoid building those gradients in the first place. This is the insight behind multi-step current protocols: by stepping current down before the voltage reaches the upper limit, you can reduce the gradient buildup and shorten or eliminate the CV phase without compromising cell safety.
Where the conservatism comes from
The 1C standard CC current was not derived from thermal first principles for any specific cell. It emerged from years of experience as a rate that is safe for most cells in most conditions, conservative enough that the range of uncertainty about thermal state, cell age, and ambient temperature could be absorbed without causing problems.
The specific concerns that drove the conservatism include: lithium plating risk at high SoC and high current, particularly in cold or partially-degraded cells; accelerated SEI growth at elevated temperature from high current; and, in less controlled applications, the possibility that the cell is being used in conditions (high ambient temperature, poor ventilation) where the nominal safe rate becomes problematic.
All of those concerns are real. None of them necessarily applies with full force to a pack engineer who has full access to the cell characterization data, is designing for a specific ambient temperature distribution, and is building a BMS that will monitor and manage the charge in real time. The conservatism of CCCV is priced for a worst case that is different from the known case.
The optimization window in CCCV
If you want to improve on CCCV, the first step is to characterize where the thermal headroom actually exists for your specific cell. This requires a thermal model and an understanding of the cell's internal resistance as a function of SoC, temperature, and cycle count.
What you typically find when you run this analysis on a well-characterized NMC/graphite cylindrical cell is that there are two distinct windows where the standard CCCV protocol is more conservative than necessary:
The first is the mid-SoC window, roughly 30% to 65% SoC, where the cell's internal resistance is at its lowest, the thermal generation per unit current is minimized, and the lithium plating risk is at its lowest because the anode is well intercalated. In this window, current can often be increased above the nominal 1C rate without approaching thermal limits, particularly when ambient temperature is below 35 degrees Celsius and the pack has reasonable passive cooling.
The second is the transition from CC to CV. Standard CCCV transitions at the upper cutoff voltage (4.2V for NMC), at which point the cell is at roughly 80 to 85% of its nominal capacity. The CV phase then delivers the remaining 15 to 20%. A multi-step protocol that begins stepping current down before the voltage cutoff, staying in a stepped CC mode rather than switching to CV, can deliver the same final SoC with less total CV time and in many cases lower peak temperature, because the current reduction happens at a point where the cell still has electrochemical absorption capacity available.
What actually constrains the optimization
The practical limits on CCCV optimization are not the physics of the mid-SoC window. They are the boundary conditions: cold temperature operation, cell age, and high-SoC plating risk.
In cold temperature operation (below 15 degrees Celsius), the electrolyte ionic conductivity drops significantly and lithium plating risk rises sharply. Any protocol that accelerates above 0.5C in this range needs to be validated specifically for cold-temperature cell behavior. The temperature-adaptive protocol needs a lower bound, not just a nominal-ambient case.
Cell age matters because as SEI grows and cathode degradation accumulates, the internal resistance increases and the safe charge rate at high SoC decreases. A protocol parameterized for a new cell will be aggressive for the same cell at 500 cycles. The protocol either needs to be re-parameterized at intervals (which requires ongoing characterization) or needs to include a conservative aging margin that ensures safety at end-of-life.
High-SoC plating risk (above 80% SoC) sets a hard ceiling on current at the top end of charge regardless of thermal state. This ceiling is electrochemical rather than thermal, and it is not moved by thermal optimization. The gains from thermal optimization live in the lower and middle SoC window; the top window remains constrained for independent reasons.
Making the optimization practical
The engineering pathway from standard CCCV to a thermally-informed multi-step protocol is not exotic. The inputs required are: cell thermal resistance (measured or calculated from geometry and material properties), internal resistance as a function of SoC and temperature (measured with a simple EIS rig), and a reference temperature envelope for the operating environment.
With those inputs, a simple model can predict what multi-step current profile will stay within the thermal limit across the operating temperature range, with explicit margins for cell age and worst-case ambient. The resulting protocol table can be implemented in BMS firmware without requiring a continuous real-time model running on the processor. The model is used offline to generate the protocol; the BMS executes it.
The gain is measurable: in controlled simulation of a 2.2 Ah 18650 cell at 28 degrees Celsius ambient, the thermally-optimized multi-step protocol reduces charge time from 20% to 80% SoC by 18 to 24 percent compared to a 1C CCCV baseline, while keeping estimated core temperature below the threshold that the 1C baseline was implicitly set to protect. The conservatism in the 1C protocol was not protecting the cell at that ambient; it was protecting against a hotter condition that was not present.
That is the CCCV critique in a sentence: it is a good protocol for the worst case, and it applies the worst-case margin even when the actual case is better. For a product designed for a specific market with a known operating environment, you can do better than that by starting with a thermal model of your cell.