There are many scientific studies done on pulse charging of Lithium-Ion batteries. However, I have found nearly none on pulse-discharging those.
Here is probably one of the few-ones.

This study employs a well know effect of Lithium-Ion batteries to regenerate the voltage after the load was interrupted, thus dramatically increasing the specific energy (Wh/kg).

"Figure shows a Comparison between Constant Current and Pulse Discharge at 600 A/m. Constant Current discharge ends at about 2 s while pulsing results in a total on time of 16 s increasing the discharge capacity by 8 times." Pulse discharge vs constant discharge

It, obviously, won't give any benefit where the constant load is applied to a single battery, however, by distributing the constant load in pulse form within the battery pack should prolong its capacity as well as its lifetime greatly.

Besides the active control needed, I do not see any extra drawback of this approach.

Why didn't it become de-facto standard within the battery-powered device market?

I would really appreciate any reference documentation on this topic.

  • \$\begingroup\$ Beides that i See no evidence in your graph to support that claim, it would need extra circuitry which costs money and space \$\endgroup\$
    – PlasmaHH
    Commented Dec 8, 2015 at 12:45
  • \$\begingroup\$ I've added a description for the image from the study URL above. \$\endgroup\$ Commented Dec 8, 2015 at 13:00
  • \$\begingroup\$ This seems to be a very contrived case. What is the discharge rate of that particular battery and its nominal capacity? 100C? 200C? It currently sounds like "if you do this bad thing to your battery, it is bad. If you do the bad thing that way, it is slightly less bad" \$\endgroup\$
    – PlasmaHH
    Commented Dec 8, 2015 at 13:02
  • \$\begingroup\$ Higher currents = higher I^2R losses, so I seriously doubt "pulses" of any kind innately increase efficiency. Make sure you are measuring total charge in and out, not just peak current (at a low duty cycle.) Yes, the duration of the pulses is much longer, but how much actual power was delivered? \$\endgroup\$
    – rdtsc
    Commented Dec 8, 2015 at 13:06
  • \$\begingroup\$ Ok, the same url I've given the link-to claims: "During high rate discharges, steep gradients in concentrations develop resulting in discharge capacities much lower than the theoretical discharge capacity. It could be as low as 2% for a 30C (600 A/m 2) constant current discharge." 600 A/m, 30C is a standart LiPo battery used in UAVs. \$\endgroup\$ Commented Dec 8, 2015 at 13:13

1 Answer 1


The article makes it fairly clear that they aren't measuring the working capacity of the cell, but its capacity under abuse (30C in a cell not designed for high discharge) and it's down to 2% of its nominal capacity.

In that context, "pulse discharge" simply reduces the mean discharge rate to (discharge rate * duty cycle) and that "improves" capacity 8-fold, to 16% or about 1/6 of its rated capacity. As an analysis of the chemistry of the cell it's interesting, but still not very impressive.

Duty cycle appears to be 10%, inferred from the time axis (160s) and the quoted discharge time (16s).

So if you attempted to distribute pulses across multiple units to achieve continuous power, you'd need (ignoring the complexities of synchronising and switching between them) 10 of these batteries.

In that context, the interesting question is, how would this solution compare with the lifetime of the same 10 cells operated in parallel, each discharged at a 3C rate and thus subjected to much less abuse? And the article doesn't ask that question.

I don't know, but strongly suspect you can get higher capacity from 10 parallel cells discharged at 3C - provided you pay attention to cell balancing during both charge and discharge.

Still, does this have any merit in applications needing pulse power? Only if it offers benefits over a battery continually discharged at 3C to recharge a capacitor capable of sustaining the pulses.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.