I'm hoping that someone on the EE StackExchange could give me some insight into what is going on with some lithium polymer batteries that I have under test.

The question I am trying to answer is: what conditions/causes lead to the type of failure mode described below?

The Setup:

I have a fixture that automatically charges and discharges a group of candidate batteries while logging battery voltage so that I can gauge battery operation and usable life. The fixture uses an MCP73837-FCI/UN Lithium Ion/ Lithium Poly charger IC to charge my batteries at a rate of ~65mA. Once they are all at full charge (4.2V or so - I go by the status flags of the MCP73837), the fixture disconnects the charge IC from the power source via a relay, connects into a load circuit, and begins discharging against a load that varies between 30mA and 80mA duty cycled at 80% at a rate of 74Hz (see below). The load is independent of battery voltage.

Battery Discharge Waveform

Once the batteries all reach a voltage of 3.4V, the batteries are disconnected from the load circuit, connected into the charging circuit, recharged and the cycle repeats.

The Batteries:

I am testing 3 different lithium polymer batteries in my setup:

  • A 120mAh battery that has a production history measured in years and is deployed in the field in thousands of products.

  • A candidate 250mAh battery that so far exhibits no issues with my test regimen.

  • A candidate 200mAh battery that has exhibited failures across 3 samples.

The Problem:

After about 10 or so charge/discharge cycles, 3 samples of the 200mAh battery began exhibiting poor battery life. Physically, the affected batteries increased in size (thickness) by 20-30% from their pre-evaluation measurements and have noticeable ruptures / rips in their pouch (picture of affected battery included below):

Picture of ruptured/torn battery pouch

The batteries seem to have rigidly increased in volume. This does not seem to be just a battery outgassing event.

Other useful information:

  • None of the other batteries under test exhibit this failure mode.

  • The affected batteries have an integrated battery protection PCB at the leads that ought to protect against accidental shorts, overcharge, overdischarge etc. (The protection circuitry consists of the DW-01 IC + The 8205 Dual MOSFET)

  • These batteries are not being compressed / under mechanical stress during testing.

  • The batteries had read about 3.9/4V prior to being subjected to the charge/discharge protocol described.

  • I have another fixture with a nearly identical charge/discharge setup where I have a 3.2V depth-of-discharge threshold instead of a 3.4V depth-of-discharge threshold and none of my 200mAh candidate batteries exhibit this failure mode even after 40+ charge/discharge cycles.

Some insight into this phenomena from someone with some general lithium polymer / battery testing / battery chemistry expertise would be very much appreciated.

  • \$\begingroup\$ Are you sure you selected the proper charge voltage for your battery chemistry? Your charger has the option of - 4.20V, 4.35V, 4.4V or 4.5V and selecting a voltage higher than your battery design target will cause the kind of damage you're seeing. \$\endgroup\$
    – John D
    Commented Dec 23, 2019 at 17:04
  • \$\begingroup\$ John - Thanks for your response. I am fairly sure I selected the correct voltage. I'm using the FCI/UN version of the chip which puts the target output of the MCP73837 at 4.2V with a range of 4.179V to 4.221V, The affected battery has a rated charge voltage of 4.2V+/-0.05V. Furthermore, the onboard protection circuit has a overvoltage threshold of nominally 4.3V+/-0.05V. I'll be evaluating the integrated protection circuit on the bench more today and should be able to determine if it is malfunctioning. \$\endgroup\$
    – RobGah
    Commented Dec 23, 2019 at 17:20
  • \$\begingroup\$ You should be looking at changing the charge profile to CVmax=4.1 and Icutoff =10%. Are you rejecting temp rise >10'C? Dendrite failures are worse under some conditions. \$\endgroup\$ Commented Dec 23, 2019 at 17:38
  • \$\begingroup\$ What are the maximum charge and discharge current ratings of the failing battery? Did you notice any 'puffing' or temperature increase during either charge or discharge before it ripped open? Did you record the charge times and completion voltages? \$\endgroup\$ Commented Dec 23, 2019 at 21:25
  • \$\begingroup\$ Have you monitored the actual cycle voltages/currents over time. The change in Vlow affecting survival suggests something strange is happening. || How well known are the candidate battery makers? \$\endgroup\$
    – Russell McMahon
    Commented Dec 24, 2019 at 1:46

2 Answers 2


There might be some interesting oscillation going on between the MCP73837 and the DW01;

If the MCP73837 triggers (Vbat > VOCP) the DW01 in either "FAST CHARGE MODE" or "CONSTANT VOLTAGE MODE", the battery will 'disconnect' and the parallel capacitor will fully charge and put the MCP73837 in "CHARGE COMPLETE MODE".

The MCP73837 goes back into "FAST CHARGE MODE" if the voltage drops below some threshold (Vbat < Vrth).

That will happen eventually because the capacitor will lose charge... however before it reaches that point it is possible that the DW01 resets (Vbat < VOCR) and the battery is 'connected' again.

Should the DW01 reset around the same time that MCP73837 goes into "FAST CHARGE MODE" it is conceivable that the battery might be stuck in a loop that repeatedly charges the battery in "FAST CHARGE MODE" -> overcharge protection -> "CHARGE COMPLETE MODE" -> etc.

You could try removing the DW01 from a couple of fresh batteries and see if the problem persists with charging. If it stops you know the two ICs aren't playing along nicely.


I think you are simply exceeding the C rating of the one battery type that fails - small LiPo cells can have an intended C of 0.1 or less, and you are intermittent loading them to 0.4C.

Note: 1C is equivalent to one full battery capacity of charge, so 200mA output would be 1C.


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.